1 Wednesday, 22 February 2023 2 (10.00 am) 3 MR BEER: Good morning, sir. 4 SIR WYN WILLIAMS: Good morning. 5 MR BEER: May I call Susan Harding, please. 6 SIR WYN WILLIAMS: Yes, of course. 7 SUSAN HARDING (affirmed) 8 Questioned by MR BEER 9 MR BEER: Ms Harding, my name is Jason Beer and, as 10 you know, I'm Counsel to the Inquiry. Can you 11 give the Inquiry your full name, please. 12 A. Susan Mary Harding. 13 Q. Many thanks for the witness statement you have 14 provided to the Inquiry and for coming along 15 today to assist us with our investigation. Can 16 you look in the bundle in front of you. There 17 should be a copy of a witness statement dated 18 23 September of last year, which is 8 pages in 19 length. 20 A. Yes, I have found that. 21 Q. Can you go to the eighth page, please. 22 A. Yes. 23 Q. Is that your signature? 24 A. It is. 25 Q. Are the contents of that witness statement true 1 1 to the best of your knowledge and belief? 2 A. They are. 3 Q. A copy of that witness statement will be 4 uploaded to the Inquiry's website and therefore 5 become public. I'm not going to ask you 6 questions about every part of it, you 7 understand. 8 A. Okay. 9 Q. Can we start, please, with your background and 10 experience. I think you were employed by the 11 Post Office for 40 years or so, is that right, 12 between 1976 and 2017? 13 A. I was. 14 Q. Do you have any professional qualifications that 15 are relevant to the issues that you know we're 16 going to discuss and are considering today? 17 A. Yes. I qualified through the Post Office as 18 an accountant so I'm a CIMA. 19 Q. For the benefit of those who don't know, what is 20 a CIMA? 21 A. Chartered Institute of Management Accountant, so 22 it's more about management accounting than 23 financial accounting. 24 Q. When did you become chartered? 25 A. Oh, heavens, I don't know. Probably about 2 1 20 years into my service but I can't remember 2 the date exactly. 3 Q. Was it before Horizon was born? 4 A. Yes. 5 Q. When was your first involvement in the Horizon 6 project? 7 A. It would have been in the initial Horizon 8 project -- would that have been around 1998 -- 9 something like that. I was working in a back 10 office, transaction processing, which was the 11 place that received the cash accounts from the 12 offices and I was in the development team, so 13 I was responsible for interacting with any 14 projects that would affect that unit in the 15 business. 16 Q. What was your job title there? 17 A. Head of -- I was probably head of development or 18 something like that, yeah. 19 Q. Head of development? 20 A. Yes. 21 Q. Whereabouts was that office? 22 A. It was in Chetwynd House in Chesterfield. 23 Q. Did you manage a team? 24 A. I did. 25 Q. How many were in the team? 3 1 A. I can't remember now but it wouldn't have been 2 a vast team. Maybe somewhere between 10 and 20. 3 Q. And -- 4 A. It depended on what projects we'd got as to how 5 many people we'd have in the team. 6 Q. When you first became involved in the Horizon 7 project, what did you do? What was you role? 8 A. I was involved in looking at the impact of that 9 on the business unit that I was working in, and 10 signing off requirements and getting involved in 11 testing and acceptance. 12 Q. You were involved in the testing of Horizon? 13 A. Well, I was not actually doing the testing. 14 Some of my team might have been there testing 15 certain things but it was more about the outputs 16 into the back office and making sure that the 17 cash account, which was the output at the time, 18 was still valid and still would represent what 19 would have been normally produced through the 20 manual process or through the other systems that 21 were in place at the time. 22 Q. So was this in 1998, 1999? 23 A. It would be around that time, yes. 24 Q. In that period, when you were involved, in the 25 way that you've described, in the testing of 4 1 Horizon, were you aware -- were you told -- that 2 there were errors, bugs and defects with it that 3 affected the reliability of the data that it 4 produced? 5 A. No, I don't recall that. I mean, we were 6 obviously -- in the testing phase you're looking 7 at outputs and we did get involved in Acceptance 8 Incidents, particularly any that would have 9 affected the integrity of the accounts. 10 Q. So moving on to that phase, then, the acceptance 11 phase -- 12 A. Yes. 13 Q. -- if you were involved in Acceptance Incidents, 14 AIs, I think it does follow that you must have 15 been aware of errors, bugs and defects with 16 Horizon that affected the reliability of the 17 data that it produced? 18 A. Well, yes, yes. Obviously, if there's 19 an Acceptance Incident that's what tells you 20 that there is a problem. 21 Q. Who did you report to? 22 A. Um -- 23 Q. Dealing with the acceptance phase. 24 A. Well, probably I was linked into Ruth Holleran, 25 but my direct report was Andy Radka at the time. 5 1 Q. Why were you probably linked in to Ruth 2 Holleran? 3 A. Because she was sort of in charge of managing 4 the acceptance process and she worked for Dave 5 Smith at the time. 6 Q. What was your job title at this time, when you 7 were involved in the acceptance of Horizon? 8 A. I was just the development team manager in 9 transaction processing. 10 Q. You mention a couple of job titles in your 11 statement: head of the operating process and 12 business process architect. 13 A. Yes. 14 Q. Did they come later? 15 A. They did. 16 Q. What did you do in order to perform your role, 17 accepting the system from an accounting 18 perspective, which is how you put it in your 19 statement? 20 A. Yeah, well, we would see the outputs of testing 21 and we would be -- we would -- if you're doing 22 that sort of thing, you'd normally have expected 23 results and then you see what the actual results 24 are for the system and you compare them. 25 Remember that I wasn't doing the detail here; 6 1 I was managing a team that was doing that. But 2 that's a normal acceptance process that you say 3 "What should the system produce and has it 4 produced that?" Yeah. 5 And that would be initially in the test 6 systems. It would not be in the live 7 environment but, obviously, later -- in later 8 phases there would be some acceptance due to 9 some live trials. 10 Q. Looking at the matter at a high level of 11 generality, would you agree with the suggestion 12 that the acceptance process for Horizon was 13 extremely problematic? 14 A. I don't know. You'd need to say what you mean 15 by "problematic". I think it was challenging. 16 It was a big system, it was complex. But 17 I don't think -- yes, there were Acceptance 18 Incidents but there were processes to make sure 19 that those Acceptance Incidents were closed down 20 to the Post Office's criteria. 21 Q. So by the end of acceptance, everything was fine 22 with Horizon, was it? 23 A. As far as I'm aware. I mean, I'm not, you know, 24 I wasn't as high up in the chain. I wasn't -- 25 I was only responsible for accepting certain -- 7 1 or getting involved in certain Acceptance 2 Incidents. I wasn't in the sort of roles that 3 other people were in signing it off. 4 Q. So within that limitation that you've just 5 described, did you work on the basis that, by 6 the time that Horizon was rolled out, it was 7 producing accurate, reliable and robust data? 8 A. Yes, I would say that. I wouldn't have signed 9 it off if I didn't think that was the case. 10 Q. Data that could be used in order to investigate, 11 take proceedings against and criminally 12 prosecute an individual? 13 A. Yes, I would have said that, although I -- of 14 course, the Network would have been involved in 15 having a say on that. So it wasn't our role in 16 transaction processing to do those prosecutions 17 or to do that investigation. So we would 18 simply -- I suppose my role was simply to say 19 "Are these statements that -- or -- that are 20 produced" -- because it produced a paper -- it 21 still produced a paper cash account at that time 22 -- "Are the expected results correct?" 23 Q. You worked on the basis that they were correct, 24 reliable and the system was robust and 25 infallible? 8 1 A. In terms of producing those outputs, our 2 evidence was that it was robust. We wouldn't 3 have signed -- I wouldn't have signed off 4 something that I had any evidence of that it was 5 producing numbers that were not correct. 6 Q. That's slightly self-fulfilling: "I wouldn't" -- 7 "It must have been robust because otherwise 8 I wouldn't have signed it off, if it hadn't have 9 been robust". 10 A. Yeah. 11 Q. On what evidence were you working? Who was 12 telling you that the system was robust, that -- 13 A. My team, who were -- 14 Q. Hold on. If we take it in turns, it'll work 15 better. 16 A. Yeah. 17 Q. On the basis of what evidence did you come to 18 the conclusion that the system was robust and 19 was producing data that was accurate? 20 A. Because we would, on a testing basis, have 21 expected results based on the inputs that we'd 22 reviewed that were going to go in. So if we 23 knew what data was going to be put in the test 24 system we would have expected results for what 25 the output would be and our testing would have 9 1 told us that those outputs were as we expected. 2 And you have to remember, this would have been 3 test systems to start with and then there would 4 have been stuff during -- when it went into 5 pilot, et cetera, that we would have been 6 involved in, as well, before ultimately 7 accepting it. 8 Q. When did your role in acceptance end? 9 A. I would say when the system went live. When it 10 started to -- when it was -- well, probably when 11 we finished pilot. We wouldn't have got 12 involved in anything after that. So we would 13 have been involved in before the first office 14 went live and then, as far as I recall, there 15 was a pilot and we would be reviewing results 16 through that pilot. 17 Acceptance would have been for it to go into 18 pilot to start with but then there would have 19 been another acceptance phase to actually then 20 roll it out. 21 Q. What did you learn about the operation of 22 Horizon in practice after it was rolled out? 23 A. Nothing that -- I was not aware of anything that 24 I heard or was told about afterwards that would 25 say it was any different from what we'd seen. 10 1 Q. You weren't told about one or two people calling 2 up helplines and saying, "We're having problems 3 balancing"? 4 A. No. 5 Q. No? 6 A. No, that wouldn't -- 7 Q. Not even a single call? 8 A. That wasn't really my role that would have been 9 other people in the project team. 10 Q. So you worked on the basis that, up until 11 rollout, the system had been got to a place 12 where it was working perfectly well and then you 13 didn't hear anything adverse about it 14 afterwards? 15 A. No. 16 Q. When did you become involved in the IMPACT 17 Programme? 18 A. Right, so you've mentioned the other roles 19 I did. So after being in transaction 20 processing, at the head of development, I became 21 head of the operating process and business 22 process architect and, as business process 23 architect, I was asked to redesign other parts 24 of the process. So we still had old, antiquated 25 back office systems. So Horizon was required to 11 1 produce a cash account and what we wanted to do 2 was put a SAP system into the back end that 3 would take an automatic feed from the Horizon 4 System. 5 And I'd come up with that all as part of my 6 role as business process architect, et cetera. 7 So as a result of that work that I was done, 8 I was then asked to lead the project to 9 implement that. 10 Q. So my question was when? 11 A. I don't know. Early 2000s. So it might have 12 been three or four years after Horizon had gone 13 live. 14 Q. Were you doing anything in relation to Horizon 15 in the three or four years between Horizon going 16 live and picking up what I'm calling the IMPACT 17 Programme? 18 A. Well, not really in that way because, as I say, 19 I went into more future looking roles, which was 20 where I ended up running IMPACT because that was 21 the future -- so I was looking into where do we 22 go next because, basically, we'd left the back 23 office as it was. We'd done the hard work on 24 automating the front office but we were left 25 with a back office which was a home developed 12 1 system that was not -- was very clunky, that 2 required paper, you know. 3 So the whole concept that I developed was 4 to -- was to put SAP systems in the back office, 5 which would take a daily feed, ultimately, from 6 Horizon. So we would have the data faster, less 7 effort to do the keying and all those sorts of 8 things. 9 Q. Can I confirm that in that three or four-year 10 period you weren't told of any problems about 11 the reliability and accuracy of the data that 12 Horizon was producing in practice? 13 A. I don't recall that at all -- 14 Q. Everything was perfect? 15 A. -- no. I don't know as I thought everything was 16 perfect, but I assumed that any -- that through 17 the problem management processes that are 18 implemented as part of Horizon going in, that 19 they would be picked up and dealt with. 20 Q. So when you picked up the IMPACT role, did 21 anyone, when you were briefed on the IMPACT 22 role -- what you were going to be doing -- say, 23 "We're looking to extend, essentially, the 24 functionality of the system in this way. There 25 have been problems in the operation on the 13 1 ground with Horizon"? 2 A. Not that I can recall, no. 3 Q. When did your role with the IMPACT Programme 4 come to an end? 5 A. I don't know, maybe 2005, 2006. So when the 6 programme was completed. 7 Q. What did you do after that? 8 A. I can't think, to be honest with you, where 9 I went next. 10 Q. Did it have anything to do with Horizon? 11 A. No. 12 Q. Until 2017, when you retired, did you have 13 anything more to do with Horizon? 14 A. Not that I can recall. 15 Q. So the last involvement with Horizon was at the 16 end of the IMPACT Programme in about 2005, 2006? 17 A. Yes, because next project was Horizon Online, 18 which I wasn't involved in. It was -- that was 19 Mark Burley's project. 20 Q. You tell us in paragraph 4 of your witness 21 statement -- no need to turn it up -- that you 22 became an expert and that you were highly 23 regarded in and for your knowledge of the 24 end-to-end business accounting processes 25 operated by the Post Office. 14 1 A. Yes. 2 Q. That would be for many years, would that be 3 right, before Horizon came onstream, whilst it 4 was being rolled out, and for many years whilst 5 it was in operation? 6 A. Yes. So through my role as, initially -- 7 because I worked in transaction processing, so 8 I knew the back end of it. I was given the role 9 of head of development which meant that I needed 10 to understand that and then, I say, as the 11 process architect and head of the operating 12 processes, I was considered an expert in the -- 13 I suppose more of the accounting side of it 14 than -- I mean, I wasn't an expert in the 15 day-to-day, you know, running of a branch office 16 or a sub office, but I was an expert in the 17 accounting process through the various systems. 18 Q. You tell us in paragraph 5 that you were 19 promoted a number of times and held a number of 20 key roles reflecting that expert knowledge? 21 A. Yes. 22 Q. You've told us already that you're responsible 23 for accepting the Horizon System from 24 an accounting perspective? 25 A. Yes. 15 1 Q. So, at that stage, your involvement in the 2 initial implementation of Horizon was like 3 an expert customer; would you agree with that? 4 A. It was initially, definitely, yes. Because 5 I was an expert customer on behalf of 6 transaction processing. 7 Q. Did you, in the light of those roles and in the 8 light of your long service with the Post Office, 9 become aware that the accounting data would be 10 used -- the accounting data from Horizon would 11 be used not only for business purposes of 12 individual subpostmasters and the Post Office, 13 but also for the purposes of the audit of them, 14 the investigation of them and might result in 15 criminal or civil proceedings against them? 16 A. Yes, of course I was. 17 Q. You say "of course", just explain to us why "of 18 course"? 19 A. Well, because -- because that's what -- well, 20 I was in the Post Office for a long time, so 21 I understood all the processes of the post 22 offices and I understood that many people over 23 many, many, many years had been prosecuted 24 because they had false -- they had, primarily, 25 on the basis of signing a cash account saying it 16 1 was accurate when it was paper, when it actually 2 didn't reflect the cash and stock that was in 3 the office. 4 Q. So you can confirm that you were aware that 5 Horizon and then Horizon plus IMPACT would be 6 producing data that would form the basis of 7 criminal and civil proceedings against 8 subpostmasters pursued by the Post Office 9 itself? 10 A. Yes. 11 Q. You tell us in paragraph 31 of your witness 12 statement -- perhaps if we can turn that up. 13 That's WITN03980100 at page 7. 14 A. Is that meant to be on my screen? 15 Q. It will come up on the screen. It's 16 paragraph 31. 17 A. Okay. 18 Q. It reads: 19 "It was agreed during the design of IMPACT 20 that the suspense account ..." 21 We're going to come back to the suspense 22 account and its removal in a moment. 23 A. I understand that, yes. 24 Q. "... would be removed, as historically it was 25 used by Subpostmasters to 'hide' discrepancies 17 1 in their accounts, rather than resolve them." 2 A. Yes. 3 Q. The suggestion that subpostmasters hid 4 discrepancies in their account, on what basis 5 was that suggestion made? 6 A. This was a historic evidence. So on the basis 7 that when a subpostmaster signed their cash 8 account or pressed the button on Horizon, they 9 were saying that the cash and stock, as well as 10 the other transactions, but primarily that the 11 cash and stock that was in that account was 12 actually physically in the branch, and that was 13 the whole concept of it, that -- and a way to 14 get rid of that was to put an amount in suspense 15 which said that "I only have this cash and stock 16 and there's something over here". 17 So it was decided, given that the Finance 18 Director was my -- was the sponsor of my project 19 at the time and that was something that he 20 wanted to do, because -- 21 Q. Who was the Finance Director? 22 A. Peter Corbett. So we wanted to get rid of what 23 was basically a place to hide discrepancies. 24 Q. Are you -- so "hide" in the sense of dishonestly 25 place; is that what you're suggesting? 18 1 A. Yes, I would say that. So if you look back at 2 the record of people who had been prosecuted 3 successful in the past, prior to this, it was 4 normally that they would inflate their cash and 5 stock balances so that the account balanced or 6 they would be signing that that cash and stock 7 existed but it didn't balance or they would move 8 something into suspense, so the cash and stock 9 was right but there was something else that 10 wasn't -- basically, a suspense account is 11 saying there's an amount here I don't -- 12 I haven't accounted for. 13 Q. You thought the solution for this was removal of 14 the suspense account because it would prevent 15 people from dishonestly hiding money within it? 16 A. Yeah, and I know you're going to ask me about 17 the contract in a bit but the contract said that 18 subpostmasters were required to, basically, make 19 that loss good, not to put it in a suspense 20 account and hide it. 21 Q. Presumably, if the aim or one of the aims of the 22 IMPACT Programme included the removal of the 23 suspense account to prevent dishonest 24 subpostmasters from hiding discrepancies in 25 their account, you carried out an analysis of 19 1 the amount of money that was in suspense 2 accounts before Horizon was rolled out and then 3 after Horizon was rolled out? 4 A. No, I don't think so. This was a concept. We 5 were not -- there'd been nothing to suggest that 6 Horizon was creating discrepancies. I was never 7 informed of that. 8 Q. No, that's a different issue, whether it was 9 creating discrepancies. I'm asking if the aim 10 of the -- one of the aims of the removal of the 11 suspense account was to prevent dishonest 12 subpostmasters from hiding discrepancies, you 13 would want to see whether Horizon had 14 contributed to that, wouldn't you? 15 A. Well, Horizon wouldn't contribute to the 16 suspense account. Horizon didn't post anything 17 to suspense. The subpostmaster posted amounts 18 to suspense when they balanced and I would have 19 assumed that, if there were issues, those issues 20 would be reported to the call centre. But I was 21 not aware, nobody said to me, you know, "There's 22 been an increase in calls saying the accounts 23 don't balance", or anything. 24 That's what -- and in my programme, you 25 know, we had representatives from the support 20 1 centre in Dearne, NBSC. So we had 2 subpostmasters involved in all of those. So 3 this was not a decision of my own. It was taken 4 in conjunction with stakeholders throughout the 5 business, including subpostmasters and branch 6 offices. So ... 7 Q. Presumably now, you have heard that there were 8 extreme problems with balancing on Horizon right 9 from its outset? 10 A. I've heard that through being involved in this 11 Inquiry but there was nothing, absolutely 12 nothing -- you know -- 13 Q. What did you think when you heard that evidence? 14 A. Well, I did wonder why I was never advised of 15 that or why I wasn't aware of that. 16 Q. What did your wonderment lead you to think? 17 A. Well, I suppose it leads me to think that maybe 18 I would have taken different decisions or looked 19 at different things. I would not deliberately 20 have removed it, you know, if I thought the 21 system was causing misbalances. 22 Q. You said that the decision to remove the 23 suspense account was a concept and wasn't based 24 on evidence that Horizon had or had not 25 contributed to any problem in the inflation of 21 1 the suspense account, yes? 2 A. Yes, the suspense account could only be -- 3 matters would only go to suspense if the 4 postmaster put them there, so Horizon didn't 5 contribute for amounts in suspense. Horizon 6 might have contributed for differences between 7 the cash and stock that the system said they 8 should have had and the cash and stock that they 9 had. 10 I wasn't aware that any -- anything had been 11 raised, if you like, by the subpostmasters or 12 even branches -- not just subpostmasters, 13 because if it was going to affect -- if the 14 system was faulty, it would have affected Crown 15 Offices as they were at the time, as well as sub 16 offices, and there was nothing that I was aware 17 of. 18 Q. For example, the Inquiry has heard from John 19 Peberdy of the National Federation of 20 SubPostmasters that by March 2001 there was 21 £10 million held in suspense accounts, whereas 22 18 months before, before Horizon had been rolled 23 out, there was only £2 million held in suspense 24 accounts. 25 A. I wasn't aware of that. 22 1 Q. Did you think that "We need to conduct 2 an analysis here to see whether Horizon is 3 contributing to the problem", not because 4 Horizon puts the money in suspense accounts, but 5 it's creating discrepancies and balances -- 6 A. I -- 7 Q. -- that has meant that subpostmasters are 8 putting more money in suspense accounts? 9 A. I wasn't aware that Horizon had made any change 10 in the suspense accounts and we did have -- you 11 know, in the programme and stuff, we did have 12 the unions involved, we did have subpostmasters 13 involved, and I will -- you know, I've taken 14 an oath. I was not aware that the suspense 15 account had increased as a result of Horizon. 16 Q. Did you look, as opposed to waiting for somebody 17 to tell you? 18 A. No, I don't think so because I don't -- you 19 know, somebody would have told me. The people, 20 you know, involved in this would have told me 21 about that. This was it was a design decision. 22 Q. That data would have been available to you. How 23 much money is held in the suspense accounts at 24 any one time? 25 A. I assume I could have got that, yes. 23 1 Q. So do I take it that neither you nor anyone that 2 you asked to looked at, over time, this is the 3 amount of money that has been held in suspense 4 accounts and, if Mr Peberdy is correct, it's 5 gone up, it's multiplied by five, before and 6 after Horizon was rolled out? 7 A. No, I was not aware of that and, as I say, we 8 did have -- we do have subpostmasters involved 9 in the design, et cetera. So -- but nothing was 10 ever raised with me, and I don't -- I'm assuming 11 there's no evidence that, you know, anybody 12 had -- that that was public knowledge, if you 13 like, or that -- I know you're saying I could 14 have found that but, you know, historically, 15 that would have meant, you know, we looked at -- 16 we would have had to look at suspense accounts 17 over time. 18 But it was a principle and that principle 19 was signed off by -- including representatives 20 of subpostmasters. 21 Q. Sorry, which principle was signed off? 22 A. The principle that we should remove the suspense 23 account. 24 Q. Where did they sign? 25 A. Well, they would have signed off the design 24 1 document. 2 Q. So the -- just explain: the National Federation 3 and the CWU would sign a document, would they? 4 A. Well, they would review it, they would be 5 reviewers. So they would have been engaged in 6 the -- they were engaged in the programme. 7 Q. So we should look at the conceptual design and 8 other similar documents for the names of people 9 from the Federation and the CWU and, seeing 10 those names, should take that as meaning that 11 they had signed off what was being done in the 12 name of their union; is that what you're 13 suggesting? 14 A. That's what I'm saying. Whether that's -- you 15 know, it's a long time ago, whether they 16 actually signed off the conceptual design in 17 total, I don't know. But we did have a lot of 18 representation throughout the design phase. 19 Q. That representation, do you mean presence at 20 meetings? 21 A. Well, document reviews, yes, and stuff like 22 that. Given that, you know, impact was 23 primarily -- primarily -- about implementing new 24 back office systems which would automate the 25 back office systems. So it wasn't about 25 1 re-looking at Horizon, except that obviously 2 there were a lot of Horizon changes because it 3 had to produce the data feed that would go into 4 the back office systems. 5 But really, we were -- IMPACT was what, 6 a £25 million programme and there were lots of 7 different streams in it, you know. But really 8 the implementation of the back office systems 9 was there, but -- and I -- if I'm honest, 10 I can't remember where the idea of removing the 11 suspense came out, but certainly, as I say, 12 Peter Corbett, who was my sponsor, I know was 13 very keen on it. 14 Q. Can we look at paragraph 24 of your witness 15 statement, please. That's on page 5. It's at 16 the foot of the page. You're speaking here 17 about the removal of the suspense account, and 18 you say: 19 "This decision was based on the core 20 principle that branches (specifically agency 21 branches) were accountable for the financial 22 integrity of their accounts. The 'suspense 23 account' had been historically used to 'balance' 24 any discrepancies which covered up losses in 25 their [I think you mean subpostmasters'] 26 1 accounts." 2 Yes? 3 A. Yes. 4 Q. The suspense account had been designed, is this 5 right, to allow subpostmasters to balance, even 6 where there was an imbalance? 7 A. Yes, and that was -- that had been there for 8 years and years when we were talking about paper 9 cash accounts. So it was just a way for them 10 to -- I suppose there were two things. It would 11 be one for branches that just couldn't find what 12 the problem was but it was also used for 13 branches that had been removing money from the 14 system, if you like, and it was to hide it for 15 a length of time. 16 So I mean we had a lot of the sort of people 17 who investigated, POID investigations and a lot 18 of investigations into previous cases where 19 people had jailed before Horizon existed, so 20 this was on a paper cash account, you know, that 21 they used suspense to hide that. They would 22 either leave their cash and stock as numbers 23 that didn't exist in the office or they would 24 simply move it to a suspense account. 25 Q. The language which you use in your witness 27 1 statement and today is of hiding and covering 2 up. Does that accurately reflect your core 3 belief that that's what was going on? 4 A. I have to say, you know, that, yes, multiple -- 5 many, many, many subpostmasters, prior to 6 Horizon, had been prosecuted for either -- for 7 falsely accounting, basically. Either saying 8 the cash and stock was there or hiding it in 9 suspense, and I could tell you lots of cases, 10 you know, of somebody where the auditors -- 11 because the auditors only used to audit sub post 12 offices, I think, once every three or 13 four years. 14 And I can remember there was one, I think it 15 was, Harrow on the Hill was quite infamous, 16 where the auditors turned up at the door, 17 because that's what they did, and he said, 18 "I want to fetch the keys out of my car", and 19 then he went; and that was it. And the Finance 20 Director -- so, you know, I ran this on behalf 21 of my sponsor, the Finance Director, but it was 22 clear that -- and I'm not disrespecting any of 23 the people that have been jailed if Horizon was 24 wrong, but there had been many people -- 25 Q. Sorry, you said "if Horizon was wrong". 28 1 A. Yes. Well, I -- I'm not doubting it was wrong. 2 I'm just saying I haven't -- 3 Q. So can we remove "if" from that sentence? 4 A. Yes, okay. But people had been jailed for many, 5 many years for false accounting because that's 6 what they got jailed for. 7 Q. So is it fair to say that this mindset, might 8 I call it, that you operated under was that 9 suspense was all about crime, it's about 10 cover-ups by dishonest postmasters hiding 11 discrepancies, and that mindset continued in 12 your 40 years in the Post Office? 13 A. Well, I wasn't 40 years in the Post Office. 14 I was in Royal Mail and things before that. 15 So -- but once I was involved in that -- it 16 wasn't me, you know, I had a sponsor for this, 17 who was the Finance Director and we were trying 18 to achieve something that meant that accounts 19 were accurate and that the cash and stock that 20 was in the office or the accounts were accurate 21 and that the cash and stock that was in the 22 offices was in the offices. 23 Q. Would you mind kindly answering my question? 24 A. Could you repeat the question, please? 25 Q. Yes. Did you have a mindset in the entirety of 29 1 your time working for Post Office, that the 2 suspense account was used by dishonest 3 subpostmasters to hide and cover up money that 4 they were taking? 5 A. My mindset was that it was a place where they 6 could do that. 7 Q. And did do that? 8 A. Yes, and did do that, because -- 9 Q. And that it wasn't used for any other purpose? 10 A. Well, it might have been used for any other 11 purpose. It might have been used for other 12 purposes as well, but there were other ways 13 those things could have been dealt with. 14 Q. Did you think it was used for any other purpose, 15 like an innocent purpose? 16 A. I don't know. It could have been. 17 Q. Well, did you? Can you help? 18 A. I don't -- I can't remember. All -- I think it 19 could be used. I'll rephrase it then. It could 20 be used for any purpose but why would you, if 21 your account balanced? So it must have been to 22 do with something that either they knew was 23 causing their account to misbalance or because 24 their account misbalanced and they needed 25 somewhere to put the difference; it didn't have 30 1 to be fraudulent. 2 Q. It could be that their account imbalanced and 3 they didn't know why it imbalanced? 4 A. That's true, but that's what I said. That's -- 5 so that there were discrepancies that they 6 couldn't explain. 7 Q. So the suspense account was there to allow 8 subpostmasters to balance where there was 9 an imbalance? 10 A. Yes. 11 Q. If there was a shortfall, the relevant sum would 12 be put into, moved into, the suspense account, 13 returning the branch to balance, permitting the 14 subpostmaster to roll over and continue trading? 15 A. Yes. 16 Q. The reason for the shortfall, would you agree 17 with this, there could be various reasons: it 18 could be a loss within the branch for a range of 19 reasons? 20 A. It could be but then subpostmasters were 21 responsible for those losses. 22 Q. That's a different issue. 23 A. So they should have declared that a loss. 24 Q. Sorry, they should have? 25 A. Should have declared a loss rather than put 31 1 an amount in suspense. 2 Q. Well, they might not know, might they? You seem 3 to be operating on the basis that it must be 4 nefarious and, if that's the basis on which you 5 were operating, then you should say so. 6 A. It was -- yes. Their account should have 7 reflected what they saw -- what was in their 8 office. It should have reflected the receipts 9 and the payments they had made and the cash and 10 the stock that then -- so they'd have done the 11 previous opening cash and stock on their 12 previous cash account, they would have done 13 a number of transactions which would be recorded 14 and the difference should have given them 15 an amount of cash and stock that was in the 16 office. 17 Q. The reasons could include theft by the 18 subpostmaster? 19 A. Yes. 20 Q. They could include poor business practices by 21 the subpostmaster? 22 A. Yes. 23 Q. They could include a negligent mistake by the 24 subpostmaster? 25 A. Yes. 32 1 Q. They could include an innocent mistake? 2 A. Yes. 3 Q. They could include, after Horizon, a bug or 4 error in the system causing an imbalance? 5 A. They could. 6 Q. Then if the sum was placed in suspense, that 7 could be investigated before Horizon at 8 headquarters? 9 A. I think the thing was that they had the 10 Network -- the Network Support Business Unit was 11 set up within Horizon, was there, and that 12 should have been the place -- you know, the 13 contract said they should make good their 14 losses. There was nothing that was a sort of 15 grace period in them. But I would have thought 16 if somebody was having large losses, they would 17 have been raising hell with the support centre. 18 Q. On the phone? 19 A. Yes, or with their Retail Line. 20 Q. Just explain what you mean by their "Retail 21 Line"? 22 A. Well, every post office was linked to Post 23 Office's -- somebody in the Retail Line. So 24 through -- and I'm not an expert in the Retail 25 Line structure but they would have people who 33 1 supported their office from the Network Business 2 Unit. 3 Q. What would you expect them to do if they said, 4 "There's an imbalance here, I don't know why 5 there's an imbalance. I think it may be being 6 caused by the system but I don't know why"? 7 A. Well, I would have expected them to investigate 8 that. 9 Q. How would they investigate it? 10 A. I don't know because I'm -- I've never worked in 11 the Network side so I don't know what the roles 12 and how they played those roles in that. But 13 somebody would be saying, you know, "You have to 14 go through your accounts", and a lot of the 15 lines on the cash account had supporting 16 documents, so is it a loss in a supporting 17 document? But that was it. 18 I mean, clearly, you know, some offices did 19 more transactions than others. Was there 20 a transaction that they keyed wrong? But there 21 must have been a way of investigating that, but 22 that wasn't, you know, my role. I wasn't 23 involved in Network support at all. 24 Q. The difference before and after IMPACT was, 25 before IMPACT, if a sum was placed in suspense, 34 1 it could be investigated by Chesterfield or by 2 the audit and security team, yes? 3 A. Yes. 4 Q. Then after IMPACT that could still happen, there 5 could still be an investigation but, after 6 IMPACT, placing the item in suspense could no 7 longer happen, could it? 8 A. No, but you're saying, exactly the same 9 resolution of what somebody thought, because 10 clearly they would know if they hadn't -- 11 couldn't put it in suspense that there was 12 a difference between the cash and stock on what 13 the cash account said and, you know, their 14 contract did they say had to make it good, they 15 could have been straight on the phone to Network 16 Business Support Unit saying, "This is saying", 17 -- I'm assuming, you know, we're not talking, 18 I expect, £10 or 10 pence here; we're talking 19 large amounts generally. If it's small amounts, 20 making it good was what they would do but if it 21 was a large amount, you would have thought they 22 would have been scouring their transaction logs 23 and trying to understand what had happened 24 there. 25 Q. Scouring their transaction logs. Can you tell 35 1 us how a subpostmaster in a branch scours 2 a transaction log? 3 A. Well, they would print it off and have a look 4 for what they thought might be an erroneous 5 transaction but a lot of the -- 6 Q. Hold on. Just stop there. What would they 7 print off? 8 A. I can't remember. There must have been a report 9 that was a transaction log. 10 Q. Was there a report called a transaction log? 11 A. I don't know. 12 Q. Sorry? 13 A. I don't know. But, you know, every week 14 before -- and before any of the automation, 15 people went through that on a weekly basis, 16 following automation that became a monthly 17 basis, although I imagine that some offices 18 still did it weekly. A lot of lines on the cash 19 account had supporting documents and that 20 enabled them to check the accuracy of those 21 things. So nothing changed as a result of doing 22 this in how you would try and identify 23 a discrepancy. It just said you can't place it 24 in somewhere -- in a suspense account. 25 Q. You've said a number of times this morning it 36 1 was their responsibility to make good any 2 losses. 3 A. Mm-hm. 4 Q. It was their responsibility under the contract, 5 I think you said, a number of times. 6 A. Yes, there is a clause, I know, that's in the 7 contract that's in some of my documents. 8 Q. Yes, if we just look, please, at paragraph 24 of 9 your witness statement, that we were on, and see 10 what you say about the responsibility of 11 subpostmasters for shortfalls under the terms of 12 their contract. You say: 13 "This decision was based on the core 14 principle that branches (specifically agency 15 branches) were accountable for the financial 16 integrity of their accounts." 17 Then reading on: 18 "The processes for raising a dispute 19 regarding a discrepancy were not changed by the 20 Programme. Branches were supported by the 21 helpdesk and/or the Network team and would raise 22 any concerns through these channels." 23 Then you say: 24 "As I have stated, agency branches were 25 responsible for making good losses. This 37 1 principle was not changed by the IMPACT 2 Programme." 3 Then in paragraph 25, if we scroll down, you 4 say: 5 "This decision was based on the core design 6 principle that, contractually, agency branches 7 were accountable for the cash and transactions 8 in their branch and liable to make good any and 9 all shortfalls." 10 I just ask you to highlight, in your mind at 11 the moment, "any and all shortfalls". If we go 12 to paragraph 29, please. You say: 13 "In sub post office [sub post offices] the 14 owners were required under their contract to 15 make good any shortfall." 16 I just ask you again to highlight in your 17 mind the words "any shortfall". 18 Then in paragraph 32, over the page: 19 "In contractual terms, they were liable, 20 contractually, for any shortfalls which had to 21 be made good." 22 Why did you believe that subpostmasters were 23 required, under their contracts, to make good 24 any and all shortfalls. 25 A. Because I think -- I'd seen the contract and, on 38 1 some of the documents in my folder, there is 2 a copy of that paragraph. 3 Q. When you were working back in -- say, between 4 '95 and 2005, was it your belief that 5 subpostmasters had to make good any and all 6 shortfalls? 7 A. Yes, although if you're going to ask me, if the 8 system didn't work right, they were responsible 9 for that, then no. But they were, unless, you 10 know, they could -- there was a bug and it was 11 identified. I'm not suggesting that if the 12 Horizon System didn't work correctly, that they 13 were liable for that but the contract said, if 14 you've done these transactions, you've reviewed 15 your account, when you press the button you're 16 saying that account is a true and accurate 17 statement of the transactions. 18 Q. Where did, between '95 and 2005, your 19 information come from that subpostmasters were 20 required under their contracts to make good any 21 and all shortfalls? 22 A. Because we had -- because I had copies of the 23 contract. 24 Q. You had copies of the contract in your office or 25 in folders? 39 1 A. Yes, or somebody -- we would have had people in 2 the teams or stakeholders who were not in the 3 team that confirmed that. So, you know, I mean 4 this was a -- you know, this was something -- 5 I didn't make these things up. I was -- you 6 know, my stakeholder was the Finance Director, 7 and this was something that we recognised and, 8 certainly as far as I'm aware, we tracked all 9 the contractual terms to say that was true. 10 Q. To summarise, it was based on actual possession 11 of the contracts and conversations with other 12 experts who would be expected to know what the 13 terms of the contract were? 14 A. Yes. 15 Q. Did it mean to you that a shortfall that was not 16 the fault of a subpostmaster was nonetheless 17 their liability to make good? 18 A. No. Well, except that it could have been, you 19 know, a member of staff or something like that. 20 If there was a shortfall because they'd had 21 a robbery, then clearly that didn't apply and if 22 I'd thought in any way that the system created 23 incorrect transactions that they wouldn't be 24 able to question, then no, I wouldn't expect 25 them to be held liable. 40 1 Q. In your statement, you say on four occasions 2 that they had a liability under their contract 3 to make good any or, in another place, any and 4 all shortfalls, without including exceptions. 5 You've just written in -- 6 A. Yes, but I didn't -- I -- you know, anybody 7 would say, if you've got an account and you've 8 signed it off, yeah, you're signing it off to 9 say that that is a true and accurate reflection. 10 And, you know, I know from reading the stuff 11 here that there were issues maybe in Horizon 12 but, in many cases, you know, what I don't 13 understand and I don't know the detail of the 14 prosecutions, there's -- why they hadn't been 15 able to identify that the system had inflated, 16 you know, receipts or said they hadn't -- that 17 sort of thing. 18 Because, you know, the process of balancing 19 required them to go through their accounts and 20 to check things, as I say, in many cases with 21 supporting documents, not always supporting 22 documents. 23 Q. Sorry, just to stop there, you're saying that 24 the process of balancing -- 25 A. Yes. 41 1 Q. -- required the subpostmaster to go through 2 a series of documents, this is after Horizon? 3 A. Yes, well, they'd go through a process that 4 they'd done with the cash account. They would 5 validate the lines that were on the cash 6 account. So, for instance, if -- and obviously 7 things changed as technology changed but if they 8 were having Giro deposits, for instance, they 9 would have Giro deposit slips. So the process 10 of balancing would say, "I've got a line on my 11 system that says I've taken this much in cash 12 deposits", and you would expect that they would 13 go through the supporting documents, which often 14 had to be sent in when we used to have paper 15 cash accounts, because that's what Chesterfield 16 would do when they got it. 17 They would check the lines on the cash 18 account that the subpostmaster had submitted to 19 the supporting documents -- 20 Q. Just to stop you there, that's what IMPACT 21 removed, essentially? 22 A. It didn't remove supporting documents. 23 Q. They removed the checking by Chesterfield? 24 A. Well, no, because there would always -- there 25 was still some products where, in the back 42 1 office systems in Chesterfield, we would have 2 a feed from the client which told us. So we 3 would identify errors but that -- those 4 subpostmasters should have been able to find 5 those errors. If they'd sent the supporting 6 documents off, they -- some of them may have 7 been sent off at the same time as the cash 8 account was completed and, therefore, they would 9 be saying, "I have -- my system says I've done 10 these transactions and these supporting 11 documents say I've done these transactions", and 12 if there was a difference they would look to 13 identify the difference. 14 So it wasn't a case of just pressing 15 a button. It was all related to supporting 16 documents and other information, to sign a cash 17 account off. 18 Q. Getting back to where we were, I think you've 19 said that you didn't work on the basis, between 20 1995 and 2005, that all shortfalls were the 21 responsibility of subpostmasters to make good. 22 There were exceptions to that rule? 23 A. Well, yes, there was a process for them to go 24 through to work out why they had got 25 a shortfall, so -- 43 1 Q. What were the exceptions to the rule? 2 A. Well, if they could -- if they could prove that 3 they got a set of supporting documents and that 4 total on Horizon wasn't there, they should have 5 been -- they would have been ringing the 6 helpdesk and saying, "Why is my system not 7 reflecting what I've keyed in?" 8 Q. What if the helpdesk said, "But you've got 9 a responsibility under your contract to make 10 good any and all shortfalls. Make good the 11 shortfall, otherwise you can't continue 12 trading"? 13 A. No, that contractual term "any shortfalls" is, 14 after all these processes have gone. If they 15 can't resolve that shortfall, that's what the 16 contract said. 17 Q. What was your view of how onerous the 18 contractual provision on any and all shortfalls 19 was at the time? 20 A. Well, that provision was no more onerous under 21 IMPACT than it had been for all the years it had 22 been they're so that contract -- that was -- 23 that had always been the case in subpostmasters 24 contracts, as far as I'm aware. Certainly in my 25 time of being involved, that was what was there. 44 1 Q. Can you recall a difference between sub offices 2 and Crown Offices in this regard? 3 A. Yes, because Crown Offices would report a loss. 4 So you don't hold a Crown branch -- were not 5 under the same contract. They were employed -- 6 they were employees, so you would deal with 7 an employee who -- a branch manager who reported 8 that they had got more money than they really 9 had in their tills, you would treat them 10 differently. That would be a disciplinary 11 offence and may lead to them being prosecuted, 12 I suppose. But it's obviously a different 13 contractual situation. 14 Q. One of the desired outcomes of the IMPACT 15 Programme was to pursue losses with more vigour 16 in order to improve debt recovery, yes? 17 A. Yes. 18 Q. Did you ever think that subpostmasters were 19 therefore going to be pursued and pushed harder 20 for losses? 21 A. I think we were trying to make it easier for 22 those losses to be identified and, yes, I would 23 expect that the processes would be there -- the 24 process should have been there anyway, but it 25 was trying -- it was the visibility. 45 1 Q. But a deliberate design aim was essentially to 2 squeeze the subpostmasters more to get onto the 3 Post Office books the accounts that was 4 otherwise held in suspense accounts? That was 5 improving debt recovery? 6 A. Yes, I don't think that means squeezing them 7 more. That's a concept -- 8 Q. But -- 9 A. We were trying to make it that the Post Office 10 accounts and the sub office accounts reflected 11 reality and that we didn't have, if you like, 12 a black hole that was called suspense. 13 Q. But a black hole that was filled by, in your 14 view, dishonest subpostmasters hiding and 15 concealing money that they had stolen? 16 A. It didn't mean they'd stolen it. It may have 17 been lost. They may have given out the wrong 18 change to somebody but the contractual liability 19 was that they should make good those losses, 20 even if they'd given out the wrong change. 21 I wasn't -- I'm not -- I wasn't in any way 22 reflect interesting on why I believe those 23 losses had occurred. 24 I was simply reflecting what was a desire of 25 the sponsors of my project, remembering I was 46 1 here to implement a project, there was 2 a programme -- project programme manager and 3 I had a number of sponsors, and they had numbers 4 of requirements. 5 Q. But you were aware, you were conscious of the 6 fact that one of the aims of the project was to 7 pursue these losses harder in order to improve 8 debt recovery -- 9 A. Yes. 10 Q. -- and therefore subpostmasters would be pushed 11 harder? 12 A. Yes. 13 Q. That was the very aim of the process? 14 A. Of that part of the process, yes. 15 Q. Can we just look at a couple of versions of the 16 contract please. To start with, look at 17 POL00000747. This is the 1994 edition of the 18 subpostmaster contract and can we turn to 19 page 49, please. Hmm. 20 Yes, sorry, internal pagination. So I think 21 four pages on. 22 Yes, that's it, under the cross heading 23 "Losses" at the top and if we can look at 24 paragraph 12: 25 "The Subpostmaster is responsible for all 47 1 losses caused through his own negligence, 2 carelessness or error, and for all losses of all 3 kinds caused by his Assistants. Deficiencies 4 due to such losses must be made good without 5 delay." 6 A. Yeah. 7 Q. Can you keep that in mind and then if we can 8 look at POL00003874. This is the "Post Office 9 Community Subpostmasters Contract" for 2006. 10 Can we look, please -- there's an error with my 11 referencing. 12 I'm going to have to ask to take the break 13 early because I can't locate the paragraph 14 within the 60 pages/70 pages of the document. 15 Apologies for that. 16 A. I think on -- 17 SIR WYN WILLIAMS: So we're just after 11.00, so 18 11.15? 19 MR BEER: Yes, that would be great. Thank you very 20 much, sir. 21 (11.05 am) 22 (A short break) 23 (11.16 am) 24 MR BEER: Sir, thank you for allowing me the time. 25 The fault was entirely mine. 48 1 SIR WYN WILLIAMS: It's always nice to have that 2 admission. 3 MR BEER: So we were looking at POL00003874, please. 4 Can we turn to page 71 in the document, please, 5 and paragraph 12, please: 6 "The subpostmaster is responsible for all 7 losses caused through his own negligence, 8 carelessness or error, and also for losses of 9 all kinds caused by his Assistants. 10 Deficiencies due to such losses must be made 11 good without delay." 12 I've shown you the contract as it stood in 13 1994 and then as it stood in 2006, to sort of 14 bookend the period that we're looking at. 15 You'll see that they are materially the same. 16 A. Yes. 17 Q. Do you agree that a postmaster's losses due to 18 his own negligence, carelessness or error is not 19 the same thing as a postmaster being liable for 20 any and all shortfalls, irrespective of their 21 cause? 22 A. Yes, because it describes those that they are 23 liable for. 24 Q. Why did you say, in four places in your witness 25 statement, that subpostmasters were liable to 49 1 make good all or any and all shortfalls? 2 A. Because I was meaning that it was like that. 3 Clearly, if there was -- you know, it was a loss 4 that was a system error, that wouldn't have been 5 the same. Because you would have resolved the 6 system error. But, you know, this was not 7 something that I would have thought would have 8 happened but -- 9 Q. What wouldn't you had thought have happened? 10 A. I would have thought that somewhere somebody 11 would have reported things that were happening, 12 would have been able to identify maybe which of 13 the transactions there was a problem with, and 14 then that would have got resolved. 15 And that was a Network -- that would have 16 been an issue for the Network because you would 17 have imagined that if you were getting a lot of 18 losses caused in a lot of offices which they 19 simply couldn't explain, that through that 20 process they had to go, which was before they 21 signed off the cash account, that things would 22 have been clear, particularly if there were 23 large losses. You know, you would have expected 24 you could see something in the accounts, but 25 I wouldn't -- that contract has never -- and 50 1 I just paraphrased the contract, which was the 2 first bit, for all losses. 3 Clearly, we were not ever attempting to 4 make, in my project, losses -- subpostmasters 5 responsible for losses which were caused by 6 a system bug. 7 Q. Isn't it the case that what you told us in four 8 places in your witness statement accurately 9 reflected the view that you held at the time, 10 back in the day, namely subpostmasters are 11 responsible for any and all losses, and that's 12 the basis on which you were working, wasn't it? 13 A. No, I would have had access to this and we -- 14 you know, I wouldn't have -- 15 Q. But did you read it? 16 A. I would never -- you're asking me to go back 17 20 years but I had people from the Network, 18 I had people from commercial involved and that 19 was where those decisions were made. It was 20 never intended to suggest that if there were 21 bugs in the Horizon System caused by Fujitsu, 22 that that wasn't there. I suppose what we would 23 have expected is -- maybe thought, is that those 24 would have been identified by the subpostmasters 25 because, I say, a lot of the lines on the cash 51 1 account, as it was, had supporting documents, 2 which they would be able to say "Well, this is" 3 -- and I don't know what problems those bugs 4 caused, I know they existed now but I didn't 5 know at the time. But I would have thought you 6 would see a pattern of losses occurring because 7 this line of receipts is overstated versus some 8 supporting documents or evidence that I had to 9 support it. 10 Q. Were the attitudes that you have displayed in 11 your witness statement, namely that suspense 12 accounts are used by dishonest subpostmasters to 13 hide and cover up losses for money that they 14 have taken and that, in any event, 15 subpostmasters are liable for any and all 16 losses, a common one in the teams in which you 17 worked? 18 A. No, it was not like that at all, not like that 19 at all. This was about trying to reflect that 20 part of the contract and this was a requirement 21 passed to me from my sponsor, the Finance 22 Director, and that's what we intended to try to 23 implement. 24 Q. Can I turn to the reasons for the removal of the 25 suspense account. You tell us in paragraph 18 52 1 of your witness statement -- and can we look, 2 please. It's on page 5, WITN03980100. We 3 should probably just look at the foot of 17 4 first. It's slightly oddly formatted: 5 "The principle objectives-off IMPACT [says 6 17] were to ..." 7 Then 18: 8 "Deliver an integrated, automate solution 9 utilising industry standard packages where 10 possible." 11 I think the second thing is: 12 "Reduce IT operating costs." 13 Then the third thing is: 14 "Reduce losses and improve debt recovery." 15 Was the objective that you've described 16 there, reducing losses and improve debt 17 recovery, something of an underestimate? This 18 was seen as an issue critical to the survival of 19 the business, wasn't it? 20 A. No. No. This was -- I worked for the 21 Finance -- the Finance Director was my sponsor, 22 you know, we did have a lot -- we had a very 23 large network of offices which were not 24 supervised, which were not audited frequently 25 and we did have, prior to any of the automation, 53 1 numbers of postmasters who were prosecuted for 2 false accounting: not theft but false 3 accounting. 4 So these were agreed. This wasn't my, you 5 know, my programme. It was the Finance 6 Director's desire to do this and that was one of 7 things he placed on us to say, "Do that, that's 8 one of the things that we want you to achieve 9 out of this programme". 10 Q. So it was actually more targeted to try to catch 11 or remove the facility for those dishonest 12 subpostmasters to hide or cover up losses in 13 their suspense account. 14 A. Yes, absolutely. 15 Q. That was an aim? 16 A. Absolutely. 17 Q. Can we look, please, at POL00038870. 18 This is version 3.4 of the conceptual design 19 of the accounting and cash management programme. 20 We'll see in a moment that I think it's dated 21 September 2003 and it details the design of the 22 whole accounting and cash management programme. 23 I think it was reviewed by you, is that right, 24 as the programme manager for the entirety of the 25 programme. 54 1 A. Yes, I would have been a -- not just reviewed, 2 I would have been the final sign-off along with 3 Clive Read who was the -- he was the 4 technical -- so I was the programme manager but 5 he was technical design authority. 6 Q. We can see you described as "Programme Manager" 7 on the first page there. What did the role of 8 programme manager involve? 9 A. Right, so this was a significant programme so 10 originally I was -- because I said I was sort of 11 head of design, or whatever it was at the time, 12 so I was asked to design an improved process 13 that would meet these -- meet some of these 14 objectives. And then, having created that 15 design, which is not this conceptual design, 16 this is later on, I was then asked to move to be 17 programme manager to implement it and, as I say, 18 my sponsor was Peter Corbett who was the Finance 19 Director. 20 So those desires to reduce losses were 21 passed to me as a requirement from the sponsor. 22 Q. We can see, I think, the date on page 5 of the 23 document -- at the foot of the page, 3.4, which 24 is what this version is -- is September 2003, 25 and you can see the date above that in the box, 55 1 yes? 2 A. Yes. 3 Q. Then if we go forwards, please, to page 14. 4 Under paragraph 3.2.2, which is just under the 5 middle of the page, "Key priorities", the 6 document provides that: 7 "2 fundamental changes have made Post Office 8 Limited's funding position a critical business 9 survival issue: 10 "[1] The business is trading at a loss 11 "[2] The migration of benefits to ACT will 12 be accompanied by the loss of pre-funding by 13 government departments of the necessary cash in 14 the network. 15 "The business now has to borrow funds to 16 fund its trading losses and to fund working 17 capital needed in branches. Such borrowing is 18 limited in its availability and its costs add to 19 the trading loss. From April 2003 DTI will 20 provide a loan and will require a robust 21 statement of cash holding as security." 22 Does that neatly summarise the reasons for 23 the IMPACT Programme? 24 A. No, not in totality, because the IMPACT 25 Programme, a lot of it, which will come in the 56 1 business drivers, no doubt, was to -- we'd 2 automated the front office through Horizon, and 3 IMPACT -- a lot of that, was around automating 4 the back office, so we replaced the old paper 5 cash account and put in new systems. So the 6 majority of IMPACT was all about the back office 7 systems, really, and then doing some things in 8 the front office to facilitate that. 9 But it wasn't -- so that was the reason for 10 it, really. 11 Q. The purpose of it, or the reasons why it was 12 being done, are set out there, aren't they? 13 A. Well, yes. So we were trying to reduce the 14 costs. So by replacing the old systems and 15 putting in new -- we'd put in a SAP system in 16 the back end, which was very vanilla. We tried 17 not to change SAP to work to the processes -- we 18 tried to change the Post Office processes to fit 19 with the package. 20 So yes, putting in package solutions in the 21 back end meant significant savings and much 22 better control, et cetera. So that was one of 23 the reasons behind that. 24 Q. Were suspense accounts to be removed because the 25 Post Office desperately needed money that was 57 1 held in suspense accounts? 2 A. No, it wasn't because it was desperately needed, 3 it was just while we were making changes to the 4 front end, and I say, it was a -- I can't really 5 remember the -- where it came from, but it would 6 have been an idea that our losses were 7 significant and were there ways to change those? 8 And some of those would have been losses not -- 9 nothing to do with the counter. 10 They would have been losses because we 11 couldn't agree at the back end how much we owed 12 our clients or they owed us, for example. So it 13 was a very large programme and the removal of 14 the suspense account in the counter was a very, 15 very small part of it. It wasn't a big driver 16 for it at all. We wouldn't have done this 17 programme just to do that. 18 Q. Why was the suspense account removed? 19 A. Because it was seen as a place where people 20 would hide losses. 21 Q. So it wasn't primarily aimed at getting the 22 money that would otherwise be in the suspense 23 account to reduce Post Office debt? 24 A. No, losses ultimately -- it's not about debt. 25 Losses ultimately appear in your profit and loss 58 1 account. So yes, this was, so there were 2 a number of things that said "How do we reduce 3 the cost or the loss of the business?" And 4 I can't remember where that actual thing about 5 the suspense account came up from, because we 6 did number of workshops, you know, to develop 7 obviously the programme and what it was going to 8 look like but it was driven by -- and we had all 9 key stakeholders involved. 10 So we would have had people from the counter 11 who would have said, you know, ultimately, we 12 have a lot of losses in the network. 13 Q. Do I take from that that it wasn't part of the 14 initial plan to remove suspense accounts? That 15 was something that was added as the programme 16 developed? 17 A. Yes, it might have been before we set off on the 18 programme because, as I say, I was -- I had 19 a business architecture some of the time, so 20 I was involved in that role or targeted to look 21 at ways to improve that end-to-end process, and 22 particularly in an accounting sense because, 23 I say, I did have -- this project was sponsored 24 by Finance. 25 Q. Can I turn, then, to look at responsibility for 59 1 the designing out of the suspense account 2 facility and start by addressing what 3 Chesterfield did before the IMPACT Programme 4 designed out the local suspense account and 5 therefore how it was changed by IMPACT. Can we 6 start with paragraph 7 of your witness 7 statement. That's on page 2. 8 Thank you. At the foot of the page, you 9 say: 10 "When Horizon was initially implemented it 11 was built to replicate the accounting processes 12 that were historically in place. This meant 13 that a 'cash account' was still produced by 14 Horizon as this was needed to feed the 'back 15 office' accounting systems and processes in 16 place [at] Chesterfield." 17 Was there, essentially, a reconciliation and 18 checking process undertaken at Chesterfield? 19 A. Yes, so we had a -- there were probably at one 20 time 150 people in what was called transaction 21 processing. The cash accounts were produced by 22 the system. Obviously, pre-Horizon we had paper 23 cash accounts coming in to a keying unit in 24 Chesterfield, so there was a big process to key 25 those cash accounts, which by itself could have 60 1 created errors in the process, and then we 2 had -- the remainder of that unit in transaction 3 processing was a large number of teams of what 4 were called error resolution teams. 5 So their job was to understand the 6 difference between -- to resolve errors. And 7 errors were created by -- often by either us 8 getting a separate feed from the client because 9 that's what the Post Office did, work on behalf 10 of clients, it wasn't -- we weren't selling our 11 own products or services. We were acting on 12 behalf of clients. 13 So for many of the products, we would get 14 a feed from Horizon of what the branches were 15 saying on the cash account and then we would get 16 a feed from the client and a whole mass of error 17 notices would be produced and then that all had 18 to be dealt with by the teams in Chesterfield. 19 Q. So this 100, 150 people at Chesterfield -- 20 A. Yes. 21 Q. -- was essentially manual reconciliation but 22 based on weekly production of the cash account 23 by Horizon? 24 A. Absolutely. So this programme was intended to 25 stop that. It was to put in a SAP system. So 61 1 that's what I said. The changes at the branch 2 were minimal, except where it was needed to be. 3 So it was all about the back office systems. So 4 we were putting in a standard SAP accounting 5 system, not a homemade built and a class system, 6 as it was called in the past. So we would put 7 in a standard SAP accounting system, which would 8 make reconciliation much simpler and, out of 9 that, somewhere in the requirements, came the 10 thing that was about the suspense account. 11 But that -- this was really to get things 12 quicker and more accurate, in -- 13 Q. And to get rid of the 150 people? 14 A. Well, yes. I don't think we got rid of them all 15 because we would still have differences in the 16 client data to the feed from the counters. So 17 we wouldn't get rid of them all but it removed 18 a lot of the resource requirement, and that's 19 what you would expect for a business that's 20 making losses, to try to -- 21 Q. Save money? 22 A. -- reduce its operating costs, yeah. 23 Q. So post-Horizon but pre-IMPACT, there was a team 24 of people at Chesterfield undertaking manual 25 reconciliation whose job it was to pick up 62 1 errors, chase them down until they were 2 corrected and resolved, either in the cash 3 centre or in the cash accounting system of 4 subpostmasters? 5 A. Yes, and that hadn't changed from pre-Horizon, 6 if you like. That's -- the systems and the 7 processes in the back end were exactly the same, 8 and so IMPACT was all about really addressing 9 that issue. 10 Q. So the purpose or one of the purposes of the 11 IMPACT programme was to introduce automation to 12 remove that manual reconciliation process to 13 investigate, amongst other things, errors and 14 discrepancies? 15 A. Yes. 16 Q. If errors and discrepancies persisted, even 17 under the automated back-end functions, who was 18 to investigate those, under this new system? 19 A. Well, there was still a team in Chesterfield and 20 it was still their responsibility to do that. 21 Q. How many people were in the team in Chesterfield 22 post-IMPACT? 23 A. I can't remember, sorry. 24 Q. Was it still at the 100, 150 level? 25 A. It would have been less than that but it 63 1 wasn't -- remembering that, really, the 2 reduction in those teams was as much to do with 3 the client processes and what feeds of data we 4 got from them and whether they were automated as 5 well. But the SAP system would have done some 6 of that reconciliation but if it still produced 7 differences, then they still had to be 8 investigated. 9 But I can't remember what size the teams 10 went down to. I wouldn't say -- we certainly 11 didn't get rid of everybody because it was 12 product design, really, that allowed the full 13 automation. 14 Q. In your witness statement -- no need to turn it 15 up -- at paragraph 13, you say Peter Corbett and 16 David Smith sponsored the IMPACT Programme. Who 17 took the decision to remove the suspense account 18 facility? 19 A. Well, all of the requirements would have been 20 signed off by stakeholders and but, ultimately, 21 it would have been the programme sponsors who 22 I was reporting into to say this will deliver 23 benefits. 24 Q. Can we translate that into some language that I, 25 at least, understand. You said ultimately it 64 1 would have been signed off by stakeholders? 2 A. Yes. 3 Q. What does that mean? 4 A. Well, when you're doing a project you produce 5 documents which describe what you're going to 6 do, yes? And those documents will have a set of 7 reviewers and a set of approvers. So any 8 document that affected the branches, would have 9 to go through the Network -- would have Network 10 stakeholders as reviewers and signatories. 11 So documents always -- when you set up 12 a document and some of the ones you pulled up 13 say, you would have to -- right up front, you 14 would have a product description, you would 15 create a product description for that document 16 you were going to produce, product descriptions 17 for other things as well, system design, you 18 know, other things, but you would produce 19 a document, you would agree with the 20 stakeholders, because we had, you know, 21 stakeholder meetings, who wanted to see and sign 22 off that document. 23 And you would have a whole -- and you would 24 see it on the front of documents like you've 25 shown before, that those reviewers would be set 65 1 before the document was created. 2 Q. So the presence of a name amongst somebody -- 3 a list of reviewers, you understand to mean 4 signing off, ie approving? 5 A. No. Reviewers were reviewers. They read the 6 document and fed back comments. That was the 7 initial stage, obviously, to get rid of any 8 inaccuracies and errors and then, ultimately, 9 there were signatories after that. 10 So each document would have a set of 11 reviewers and then the people who were 12 responsible for actually signing it off and they 13 would normally sign it off on the basis that the 14 reviewers, who were the people who worked for 15 them very often, were happy that the review 16 they'd done was adequate and that there were no 17 significant issues that hadn't been resolved. 18 Q. So if we get back up on the screen, please, 19 POL00038870. If we look at the front page on 20 that, does that tell us who signed off, 21 ie approved, the contents of the conceptual 22 design for the accounting and cash management 23 programme? 24 A. Yeah, what I probably can't understand on that 25 is we've obviously got business architecture and 66 1 technical architecture -- oh no, they're the two 2 people there, yes. So yes, that's right. 3 So ultimately they were the signators, they 4 weren't necessarily the reviewers. There would 5 be separate reviewers in advance of that. 6 Q. Who signed off this document? 7 A. Myself and Clive Read. 8 Q. So you -- if I asked the question "Who agreed 9 that the things described in this document are 10 accurate and that proposals in it should be 11 implemented?" the answer would be "Me and 12 Clive"? 13 A. Ultimately, yes, there would be somewhere a set 14 of reviewers, and Clive and I would only sign 15 the documents off if those reviewers were 16 satisfied with the document. If they raised 17 issues, we would either resolve them or do 18 something but there is a whole process before. 19 This is just the signatory bit at the end. So 20 ... 21 Q. So if we look at page 7 of the document, where 22 we see a list of reviewers in the top box, is 23 that list of people people who have looked at 24 the document before you have signed it off? 25 They have not signed it off, you have? 67 1 A. No, because they didn't have the authority to 2 sign it off. So they would be the minimum 3 people -- so at the beginning of the programme 4 you create a product description, and that 5 product description will tell you who are going 6 to be the mandatory reviewers who are going to 7 be the signatories, who are going to be the 8 people that create it. And that would all go 9 through -- those documents at the beginning that 10 tell you who those people are would go through 11 its own review process to make sure that 12 everybody that should be involved in that, in 13 the review, was involved in that. 14 But Clive and I basically would ultimately, 15 in most documents -- not always -- be the final 16 signatories, because we -- and we would do that 17 not on the basis that we understood all the 18 detail in that, but the people that we had 19 agreed upfront in the product description should 20 be involved in the review had been involved. 21 Q. Who took the decision, ie signed off the 22 decision, to remove the suspense account 23 facility? 24 A. Well, if it's in this document -- 25 Q. No, no, it's not. 68 1 A. No. Well, you'd have to look at the -- whatever 2 document that was in. 3 Q. I can't see one. That's why. It seems to be in 4 a collection of email chains. 5 A. No, I can't remember that far back because, 6 obviously, it was a small part of the programme. 7 I'm not saying it wasn't an important part but 8 if you asked me who, we clearly would have 9 involved Network in that, as a primary owner, 10 because they were the primary owner of that, but 11 on the back end, Finance were, if you like, the 12 people who were responsible for the outcome of 13 that design, because it was about what losses 14 they saw in the account. 15 So there would have been -- there should 16 have been something that said that. But I don't 17 know, you know, talking 20 years ago here, 18 it's -- 19 Q. On a decision like that, we wish to remove the 20 suspense account -- 21 A. Yes. 22 Q. -- because we believe that it is used by 23 subpostmasters to hide and cover up losses -- 24 A. Yes. 25 Q. -- and it will have the benefit of saving the 69 1 Post Office millions of pounds a year, would the 2 board be involved -- Post Office Board be 3 involved in such a decision? 4 A. No, they would have remitted that. So, 5 actually, if you look at the bottom, "Business 6 Review", Stephen Hirst was in Finance, quite 7 high up in Finance. Ruth Holleran, who I'm sure 8 you're aware of on this programme, was -- 9 I can't remember exactly her role at the time 10 but she was high up in sort of the IT. Vicky 11 Noble was in Transaction -- was the head of 12 Transaction Processing at the time. Ann 13 Cruttenden was the business change manager for 14 the Network. Ann Clarke was in my team. 15 I can't -- Bob Lammin. I'm sure POID, 16 either Bob Lammin or Neil Salter -- Jack 17 MacKenzie, I don't remember Jack MacKenzie, so 18 it's a cross-representation -- 19 Q. Sorry, just to stop there. You think that 20 Mr Lammin or Mr Salter were from the 21 Investigations Division? 22 A. They could have been. 23 Q. Okay. My question was about the board and you 24 said that you don't think the board would have 25 been involved in decision making on this. 70 1 A. No, well at a high level, so the original 2 programme initiation document would have said 3 that that was one of the things we were going to 4 do. So the project initiation document would 5 have been signed off across the business by key 6 stakeholders. This was just the implementation 7 of something, so this is the conceptual design 8 that talks about that, but that's not the 9 initial "This is what the programme is going to 10 deliver". 11 Q. I'll try and ask it more simply. Was the 12 decision to remove the suspense account 13 a decision, to your knowledge, that was taken by 14 the board? 15 A. I think it was in the original business case, 16 because we would have had a figure for the 17 reduction of losses we felt was associated with 18 that. 19 Q. So the board would have had knowledge of it? 20 A. Absolutely. 21 Q. Can we look at some email threads, please. 22 Starting with FUJ00126036. Can we look, please, 23 at page 4. You'll see the last email in the 24 chain is from Clive Read. If we scroll up to 25 the bottom of page 2, you'll see that Mr Read 71 1 send it to Ruth Holleran and Tony Marsh and it 2 was copied to you; can you see that? 3 A. Yes. 4 Q. What roles at this time, early 2004, did 5 Ms Holleran and Mr Marsh perform in relation to 6 you? 7 A. Well, they would be my stakeholders on the 8 project. So they may or may not have been on 9 the board. I can't remember now who was the 10 constituents on the board. But Tony Marsh, 11 I think, was to do with Investigations? But I'm 12 not ... 13 Q. If you can't remember it's best to say, "I can't 14 remember". 15 A. I can't remember. 16 Q. Okay. Let's look at the email. This is from 17 Mr Read who was the Chief Systems Architect at 18 Post Office. So if we go down rather than up. 19 Thank you. 20 He says: 21 "As you know we are currently in the middle 22 of requirements workshops on the final phase of 23 the IMPACT Programme. Although we have 24 a scheduled Stakeholder meeting early in 25 February, given tight timescales there are some 72 1 emerging concerns which I think I need to flag 2 up." 3 Number 1 is "Suspense Account Threshold": 4 "The current assumed position is that 5 a single threshold of £250 will be applied by 6 Horizon below which variances cannot be placed 7 into Suspense Account ... This is a new system 8 control which does not currently exist." 9 Can you remember anything about that? 10 A. No, I can't. I mean, it was obviously some 11 forerunner to us removing the suspense account. 12 Q. He continued: 13 "There is a requirement (from Operations 14 [representatives]) to introduce a number of 15 different thresholds depending on the Office 16 type (eg Community offices to be at a much lower 17 level). 18 "Although this could be accommodated I have 19 concern it begins to add additional complexity 20 both to the system build and subsequent 21 operation ... is this [a must-have]?" 22 A. Mm-hm. 23 Q. Then under item 2 "Suspense Account 24 Authorisation": 25 "The current assumed position is that 73 1 subject to threshold control above, the 2 requirement to seek telephone authorisation for 3 posting variances to Suspense would cease, on 4 the understanding that improved timeliness and 5 visibility of office liabilities ... would 6 provide sufficient control ... 7 "The Operations and Security view was that 8 removal of this control would declare 'open 9 season' on the use of Suspense postings, leading 10 to loss of financial control, spiralling 11 non-conformity, etc ..." 12 Does that tie in with what you were 13 suggesting as to the nefarious use of the 14 suspense account earlier. 15 A. Yes, "open season". That's exactly what that 16 means. Yes. 17 Q. He continues: 18 "I think this is an important position to 19 take in our approach, to underline our objective 20 to simplify and leverage new capability, but 21 recognise the challenge is therefore to define 22 a 'fit for purpose' control framework which 23 tackles these fears head-on." 24 Then if we go up the page, please, to the 25 reply. You're copied in to Mr Marsh's reply. 74 1 Can you see that? 2 A. Yeah. 3 Q. Top right, "Sue M Harding". 4 A. Yeah. 5 Q. In his second paragraph, Mr Marsh says: 6 "On the suspense account issue, I'm afraid 7 that I share the same beliefs as mine and other 8 Ops reps, if there is no independent control and 9 authorisation process for the use of suspense 10 accounts then postings will rapidly increase to 11 unacceptable levels. Irrespective of our 12 aspirations for a simplified process to support 13 commercially minded agents I believe that many 14 of those of a more historic mindset will exploit 15 the facility ..." 16 So that's again very similar to the views 17 that you expressed earlier? 18 A. Yes, as I say, this -- the decision to remove 19 suspense account wasn't mine and mine alone, you 20 know. This is business people saying, "We have 21 an issue here and we need it to be tackled". 22 Q. In the decision to remove the suspense account, 23 was there ever any discussion that "Horizon may 24 contain errors, bugs and defects that would 25 cause discrepancies when it comes to balancing, 75 1 we need to take that into account"? 2 A. Not that I'm aware of but you can see from this, 3 I was copied in on this, the decisions were 4 around the people who worked for me and the key 5 stakeholders. So I don't know whether any of 6 those people were aware of that. I certainly 7 wasn't aware of that. 8 Q. In the meetings that you attended, in the 9 documents that you exchanged, in any workshops 10 that you were present at, was it ever discussed 11 "Hold on, look, Horizon in its design, its 12 testing and its rollout, has been slightly 13 problematic in terms of balancing. We've had 14 hundreds and thousands of calls from 15 subpostmasters saying they can't balance and 16 they can't explain why they can't balance. We 17 need to take that into account in deciding to 18 remove the suspense account facility"? 19 A. No, I was never aware of that, I was never 20 aware -- I know right at the beginning I was 21 involved in an Acceptance Incident, right at the 22 beginning, about balancing and stuff, but I was 23 never aware that in the live system there were 24 bugs, et cetera. 25 Q. Was, instead, the focus on covering up, hiding 76 1 and exploitation by subpostmasters and removing 2 the facility for them to do that? 3 A. Yes. I mean, there was absolutely no doubt that 4 previously and prior to Horizon even being 5 there, that subpostmasters -- remembering that, 6 you know, they were independent, they were 7 sometimes only audited once every three years, 8 that they did have losses and hid those losses. 9 And, you know, it was part of the principle of 10 not particularly IMPACT but it was one of our 11 things to try to keep those to the minimum. And 12 I can't remember -- the removal of the suspense 13 must have come up in the initial requirements 14 workshops and accepted by all of the 15 stakeholders. 16 Q. Mr Marsh continues: 17 "Given that the overall project should 18 simplify reconciliation and settlement 19 significantly and should therefore mean that 20 errors will be identified more rapidly and will 21 be even more clearly the fault and 22 responsibility of the agent, is there any reason 23 to have a suspense facility at all? This might 24 mean that in extreme cases the agent would need 25 to contact the Retail Line or NBSC and negotiate 77 1 a 'loan' (at some level of interest?) to cover 2 very high values of loss but in most cases the 3 agent should be sufficiently capitalised to 4 cover ordinary variations, particularly if the 5 opportunity were offered to make losses good via 6 credit card, thereby enabling them to tap into 7 56 days of interest free credit (a facility 8 faced by the NFSP despite my early misgivings.)" 9 Were you aware at this time that there had 10 apparently been a suggestion by the NFSP that if 11 large losses were shown, their members could 12 always make it up by credit cards, tapping in, 13 therefore, to 56 days of interest free credit? 14 A. I may have been aware of it because I don't -- 15 obviously, my name has been copied in on some of 16 these things but you have to remember as 17 programme manager I was copied in a lot of 18 things and, you know, I trusted -- we had some 19 very competent people here. Clive was a very 20 competent technical architect. Tony, if 21 I remember, was head of security, was he? 22 Something like that. So my understanding of 23 things or my involvement in things was generally 24 at a high level. I wasn't -- I was copied in on 25 these things but I couldn't have dealt with 78 1 everything, if you like. But I do -- I do 2 remember that the decision was made, and it 3 involved lots of stakeholders in that decision. 4 Q. Was that decision informed by a view of the 5 contract "We can get rid of suspense entirely 6 because subpostmasters have to bear the 7 responsibility for any and all discrepancies 8 anyway"? 9 A. Yes, I think that's true, although I don't think 10 anybody would have necessarily meant that that 11 included bugs and errors in the system. 12 Q. Was there ever any consideration of what happens 13 if there are errors, bugs and defects that are 14 the fault of the system? 15 A. Well, if anybody should have done that, I think 16 that would have been Clive, who was the 17 technical architect, who would understand much 18 more about the technical architecture of the 19 system, et cetera. But one would expect, you 20 know, that the subpostmasters would be raising 21 calls with the helpdesk if they could see that 22 there were suspect transactions. 23 I don't know what happened with the bugs and 24 defects and what the result was in the accounts 25 that were produced in the branch. So -- but one 79 1 would have thought that if somebody was having 2 a very significant loss, they would try and pin 3 it down to something and then take some action, 4 either through the Network Support Business Unit 5 on the helpdesk or through the hierarchy in the 6 Network. 7 Q. What would you say if the helpdesk said to them 8 "But it's your responsibility to make good any 9 and all losses: make good the loss"? 10 A. Well, I don't think -- I think that's different 11 to somebody saying "I think there's a problem 12 with the system". If there's a problem with the 13 system I would have expected somebody would deal 14 with that or, you know, look into it. 15 Q. What if a subpostmaster was calling and saying, 16 "When I'm balancing, it's showing, week on week, 17 increased cash that I simply do not hold"? 18 A. I would have expected the Network Business 19 Support Unit to do something about that. 20 Q. And not say "Under the contract you're liable 21 for all and any losses, make good the loss or 22 stop trading"? 23 A. No, you know, if branches were raising concerns 24 that something was happening that they couldn't 25 understand, then if this was across a number of 80 1 people, even if it was one office that kept 2 saying, you know, it happens every week, I would 3 have expected there would have been some -- 4 that's what the Network Business Support Unit 5 were there for. 6 Q. The Inquiry has hearing a good deal of evidence 7 that there were numerous occasions when Horizon 8 would create imbalances, that the system would 9 fail to identify how that imbalance had been 10 created and, even after investigation, a root 11 cause of the creation of the imbalance could not 12 be found. 13 A. Mm-hm. 14 Q. Was it your view that the system was working 15 robustly and well and, therefore, there was no 16 need to make provision in the IMPACT Programme 17 for such events or was it your view that, 18 regardless of any bugs, errors or defects, the 19 responsibility for any losses was that of the 20 subpostmasters anyway? 21 A. I would absolutely never have said that 22 subpostmasters had to make good losses that were 23 caused by the Fujitsu -- the Horizon System. 24 No. That's not my way of working at all, and 25 I wasn't aware of any of that, so ... 81 1 Q. Were you present at any meetings at which the 2 NFSP offered a view similar to or the same as 3 set out in this email here? 4 A. A meeting with the NFSP, did you mean? 5 Q. Sorry? 6 A. Did you mean was I at a meeting with the NFSP? 7 Q. Yes. 8 A. No, I don't recall that. 9 Q. Did you attend meetings with the NFSP or was 10 that below your level? 11 A. Um ... I'm sure I did, because -- oh no, I'm 12 trying to think. No, it was probably -- I would 13 have had somebody in my team who had the 14 relationship with the NFSP. I would only have 15 gone to a meeting if that member of my team had 16 said they thought I needed to attend. 17 Q. Can we look, please, at FUJ00126038. Thank you. 18 This is an invitation to a meeting by email 19 dated 12 February 2004. The meeting was about 20 branch trading, the treatment of the suspense 21 account to be held on the 18 February. You'll 22 see that you're not amongst the list of those 23 who were invited to it. 24 A. Mm-hm. 25 Q. But I just want to ask you about some text in 82 1 the paragraph at the bottom of the page, under 2 the heading "Suspense Account Manual 3 Authorisation process and universal parameter". 4 It says: 5 "The decision was reached yesterday by key 6 senior stakeholders to remove the Suspense 7 Account altogether. This would force Branches 8 to make good all losses immediately. This needs 9 to be considered in terms of how Branches can 10 adjust figures, hardship cases, how Branch 11 accounts will be corrected with errors ..." 12 A. Mm-hm. 13 Q. Can you recall, would you have been one of the 14 people described as "key senior stakeholders" 15 who took a decision to remove the suspense 16 account? 17 A. I honestly can't remember. If I was in 18 a meeting with those senior stakeholders, it 19 would not have been my decision. I was 20 a programme manager of a project that the 21 requirements came from the business, and so that 22 would have been -- I could have been at 23 a meeting. I can't remember. As I say, 24 a wouldn't see myself -- you might prove me 25 wrong in a minute by another document, I don't 83 1 know. But I would not have seen myself as 2 making that decision. I was a programme 3 manager, I wasn't a business -- I wasn't 4 somebody in the business. 5 Q. So who, if you can help us, would be key senior 6 stakeholders who took a decision on 11 February 7 to remove the suspense account altogether? 8 A. Well, at the end of the day, they may have 9 been -- it may have been the programme board, 10 which would have been people like -- so my 11 programme was sponsored by the Finance Director, 12 Peter Corbett, and then there would be somebody, 13 absolutely, from Network, and that might have 14 included POID. So there would be those type of 15 people who were -- who would be making that sort 16 of decision. I say, I may have facilitated the 17 decision being made but it would not have been 18 my responsibility to be one of the people who 19 voted on the decision, if you like. I was just 20 there to implement things if the business 21 decided that's what they wanted to do. But 22 I don't remember the meeting. 23 Q. I wouldn't expect you to remember the meeting; 24 I'm just trying to -- we haven't got a record of 25 it or any other documents, I think, that refer 84 1 to it. 2 A. No, no. But that's who would imagine. It would 3 be somebody -- it would certainly have been 4 Peter Corbett, I would have thought, and then it 5 would include people from Network, which would 6 include people like Security and people like 7 that. 8 Q. Would you agree with the view expressed there 9 that this, ie the removal of the suspense 10 account altogether, would force branches to make 11 good all losses immediately? That was the aim 12 after all, wasn't it? 13 A. Yes, after all, at the end of the day, that's 14 what contractually I think they were required to 15 do. I think that would have been -- not on the 16 basis of that -- those losses including errors 17 and bugs in the system. I can't imagine any of 18 the key stakeholders saying that that was -- but 19 I wasn't there, I don't think. 20 Q. Was that ever considered by anyone? You made 21 an exception today, "Ah, but not if it was 22 caused by an error, bug or defect by Horizon", 23 but in the documents we've got we see no 24 expression to that. In your witness statement 25 you refer to an obligation to make good any and 85 1 all losses and you said a number of times today 2 yes, ultimately it was the contract to make good 3 any and all losses. 4 A. Yes, and I had no visibility, you know. As 5 I say, I've taken an oath, I had no visibility 6 that there were problems with the system. 7 I don't know whether people in POL knew there 8 were problems with the system. I'm not saying 9 they did or they didn't, but I wasn't aware of 10 that. And, certainly, you know, it would not 11 have been -- I would not, hand on my heart, have 12 ever expected a branch to make good a loss that 13 was system created. 14 Q. Can we look at paragraph 38 of your witness 15 statement, please, which is on page 8. 16 You say in 38: 17 "The 'robustness' of Horizon was not 18 considered to be an issue and there was no 19 evidence to suggest there were any 'bugs' 20 affecting the efficacy of the system." 21 Did you hold that view when you were 22 involved as an expert customer? 23 A. Sorry, what do you mean by "expert customer"? 24 Q. It's a phrase you use in your witness statement 25 when you were brought in to the design and 86 1 testing stage of Horizon. 2 A. Ah, right, okay. 3 Q. Did you hold that view? 4 A. Yes, I do remember, as I say, that I was 5 involved in an Acceptance Incident of the 6 original design and I remember going to Feltham, 7 and I think Tony Oppenheim was the -- was he the 8 main guy? And Tony had to -- I think -- and, oh 9 gosh, you know, it is going back a large number 10 of years -- that they did some changes or 11 created -- I can't remember but they did, on the 12 face of it, in the meeting that we had, resolved 13 what was -- what was the Acceptance Incident. 14 So -- and certainly would not have signed 15 off that Acceptance Incident if I thought there 16 were any. But this was acceptance -- I can't 17 remember whether we did acceptance before it 18 went in any offices or after there'd been 19 a trial. I mean, my recollection -- 20 Q. It was the latter. 21 A. Was it, after the trial? So I was certainly not 22 aware that there was anything that would affect 23 the accuracy of the accounts. Otherwise 24 I wouldn't have signed off that Acceptance 25 Incident. 87 1 Q. It was therefore -- this, what we see in 2 paragraph 38 -- your view at the time that 3 Horizon was rolled out? 4 A. Yes. 5 Q. It was your -- 6 A. Well, and that I'd not heard anything since. 7 Q. Sorry? 8 A. And that I was not aware of anything since then. 9 Q. When did you become aware that Horizon was 10 bedevilled with bugs, errors and defects? 11 A. Well, I didn't know, I'm saying I didn't know -- 12 well, until I'd got involved in this, actually. 13 Q. So even up until 2017 when you left, you still 14 thought Horizon was working perfectly well? 15 A. Yes, I had not been advised of anything. 16 Obviously, I'd moved into different roles by 17 then anyway so I wouldn't have been involved in 18 that sort of thing. I mean, I was working on 19 totally different projects and nothing to do 20 with branches. 21 Q. It follows that you held this view throughout 22 your involvement with the IMPACT Programme? 23 A. Yes. 24 Q. Were you actively told that there were no 25 errors, bugs or defects in Horizon or no issues 88 1 with its robustness? 2 A. No, but I would have expected, given that the 3 people who were on my board -- because remember, 4 you know, I was a programme manager and I had 5 a programme board of people high up that 6 somebody would have said, "Well, Sue, if you do 7 this, you know, what's the effect of these bugs, 8 etc?" so I don't recall that I was ever advised 9 that there were any bugs. 10 MR BEER: Yes, thank you. They are all the 11 questions I am going to ask you. I suspect 12 there are other questions from others. 13 SIR WYN WILLIAMS: Mr Stein? 14 Questioned by MR STEIN 15 MR STEIN: Ms Harding, I represent a large number of 16 subpostmasters and mistresses. One of those, 17 his name is Mark Kelly. You've used an example 18 today in your evidence which is that you 19 wouldn't have expected that someone who had 20 suffered a robbery at their Post Office would be 21 then asked to pay up for those losses. Do you 22 remember saying that in your evidence today? 23 A. Yes. 24 Q. Mark Kelly was robbed, and the robber took 25 £47,000, was prosecuted as a result, and was 89 1 given seven years for conducting that robbery. 2 Mr Kelly was pursued by the Post Office for the 3 £47,000. Are you aware of that? 4 A. Absolutely not. 5 Q. He gave evidence about that and what happened to 6 him, on 1 March 2022 last year. Did you by any 7 chance see his evidence? 8 A. No. 9 Q. Did you watch the evidence of any of those 10 people, the subpostmasters and mistresses that 11 gave evidence during that time? 12 A. No. 13 Q. The date of his robbery was 14 November 2004. 14 Now 2004, help us please understand the 15 significance of that date: was that either just 16 at the beginning or just after the start of the 17 IMPACT Programme being born, essentially being 18 discussed? 19 A. Yeah, in -- I can't remember whether it was, but 20 it was around that time when IMPACT was either 21 in -- it was certainly in progress or -- 22 Q. So his robbery in November 2004 was just at the 23 starting point of discussions about the IMPACT 24 Programme and him being pursued, then, for the 25 £47,000 by the Post Office was whilst the IMPACT 90 1 Programme was being discussed, it seems. 2 A. Possibly, yes. 3 Q. I see. Now your statement, which you've been 4 asked a number of questions about by Mr Beer 5 here, who has gone through different aspects of 6 your statement, you make various references 7 that -- I'll only quote one, paragraph 32: 8 "In contractual terms, they were liable 9 [that's subpostmasters] contractually for any 10 shortfalls which had to be made good." 11 You make a number of references throughout 12 your statement in relation to that same topic, 13 this contractual liability. You repeat it, 14 I think, four times. 15 Help us understand a little bit more about 16 how your statement was created. Did you have 17 the support of lawyers, solicitors, in relation 18 to that? 19 A. No. 20 Q. No. So it's a statement you wrote yourself; is 21 that correct? 22 A. Yes. 23 Q. Right. Did you -- so we can just understand 24 exactly what happened, was this a statement that 25 you checked with your previous employers, the 91 1 Post Office, at any stage? 2 A. No. 3 Q. No. I'm grateful. Forgive me for asking those 4 questions in that way. 5 A. No, I'm -- 6 Q. That's the way your statement reads. It reads 7 as a statement that's made by someone that has 8 done it themselves rather than going to someone 9 like me, a lawyer; do you understand that? 10 A. Yes. 11 Q. Right. Now your statement, in terms of the way 12 it refers to the use of the contractual terms 13 that SPMs, subpostmasters, were liable for any 14 shortfalls, at paragraph 31 of the statement it 15 says this: 16 "It was agreed during the design of IMPACT 17 that the suspense account would be removed as 18 historically it was used by subpostmasters to 19 hide discrepancies in their accounts rather than 20 resolve them." 21 Now, you then have the repeated references 22 to the contractual liability. Mr Beer has asked 23 you a good number of questions about the IMPACT 24 Programme itself and gone through some of the 25 documentation. 92 1 Just help us understand a little bit more. 2 This attitude, which is the contractual 3 liability attitude towards subpostmasters liable 4 for any shortfalls which had to be made good, we 5 don't actually see it phrased that way in the 6 IMPACT documents and points that have been 7 raised and have been discussed with you. It 8 doesn't appear that way to be referred to. 9 A. Well, I mean, I was actually on holiday when 10 I was writing my statement, so I wouldn't 11 necessarily have looked at all the 12 documentation. So it's just my words and my 13 recollection. 14 Q. Right. But you've had number of questions being 15 asked of you today about the IMPACT 16 documentation. The IMPACT programme itself took 17 in terms of development time about three years, 18 yes? 19 A. Yes, it was a very significant programme. 20 Q. Significant programme. It took a lot of 21 discussions. 22 A. It did but you have to remember the IMPACT 23 Programme was mainly about replacement of the 24 back offices, back office systems. 25 Q. Yes. 93 1 A. It wasn't -- there was not a massive focus on 2 the branch, because the -- that data was already 3 coming in to the incumbent systems that were 4 there. 5 Q. All right but, nevertheless, three years of 6 development of the programme and then you get 7 the signing off of it. There must have been 8 a lot of meetings involved in the development of 9 the programme; is that correct? 10 A. Yes. 11 Q. Were they minuted, these meetings? 12 A. I'm absolutely certain they were, yes. 13 Q. Yes. Then there must have been a number of 14 different iterations, number of different early 15 copies of the final documents that were then 16 signed off; yes? 17 A. Yes, so ultimately there it was a business case 18 that went right up to the group board, because 19 of the level of cost associated. 20 Q. Now, all of these discussions, and the ones that 21 were minuted, were they kept by you in 22 a particular file on a laptop or a desktop that 23 you had? 24 A. We would have had a depository of those 25 documents but I wasn't, you know, I was 94 1 programme manager so I had a programme support 2 team that would have dealt with all those 3 documents. 4 Q. The programme support team that dealt with 5 documents in the repository, can you give us 6 an idea of, if you like, who was in charge of 7 that? Who had control of that documentation? 8 A. I can't remember that. 9 Q. The repository, any better idea about a way to 10 identify that? 11 A. No. 12 Q. Was it called anything? Would you say something 13 to a particular repository within the system? 14 A. No, I would rely on my team, if I wanted 15 a document, to get that document for me. 16 Q. Right. Your team -- the person that you would 17 go to within your team would generally be who? 18 A. The -- my programme support team. 19 Q. Name, please, if you have it? Who in the 20 programme support team would you routinely go to 21 to say -- 22 A. I can't honestly -- 23 Q. -- "Would you mind terribly, can you find me the 24 meeting minutes of a year and a half ago"? 25 A. I can't remember, sorry. 95 1 Q. Who was in your team? 2 A. So I had Torstein was my technical design -- 3 Q. Torstein? 4 A. Torstein Godeseth, who later moved to Fujitsu. 5 Torstein was involved with me. I had Ann 6 Clarke, who was -- used to work in Transaction 7 Processing, so -- which was really -- because 8 the whole of IMPACT primarily, the main of 9 IMPACT was to replace the back office systems. 10 You know, the changes that were made to the 11 Horizon System were only to facilitate that. 12 So we were putting a SAP system in to the 13 back office. So that was primarily were 14 there -- Ann Cruttenden represented the Network. 15 Q. But in terms of people that can help us identify 16 the documents, these meetings that you claim 17 were minuted, that might be Ann Clarke and 18 Mr Godeseth? 19 A. They might be able to, yes, but neither of them 20 are at the Post Office anymore so if they 21 were -- if they were somewhere, you know, that 22 neither of them -- or -- are there any -- 23 Q. But we should be able to find, should we, within 24 the Post Office, a repository of documentation 25 that deals with the development over the three 96 1 years of the IMPACT Programme, basically, you 2 would expect? 3 A. I think you might do but I don't know where that 4 would be. 5 Q. Right. Now, help us a little bit more. I've 6 asked a number of questions about the 7 contractual liability side of things and then 8 you've been asked by Mr Beer questions about the 9 ability which was to use the suspense account 10 used by subpostmasters to hide discrepancies. 11 I've asked you about the shortfall side of 12 things not being in the IMPACT documentation. 13 Have you seen in the IMPACT documentation that 14 you've gone through a reference to 15 subpostmasters hiding discrepancies in those 16 accounts in that documentation? 17 A. No, I guess it's my -- the wording is my 18 recollection of what I was asked to do in the 19 programme. Remembering that, you know, I was 20 the programme manager. I wasn't the owner of 21 the requirements, although some of it -- because 22 I'd been business architect, the origins of the 23 programme came out of me being the business 24 architect because I was trying to design more 25 effective system relationships. So the main 97 1 focus of IMPACT was about putting a SAP system 2 in the back office. 3 Q. Yes, so the points that I've raised, which are 4 the -- the potential for subpostmasters to use 5 a suspense account to hide discrepancies, 6 paragraph 31, one of the examples, paragraph 32 7 the contractual liability for shortfalls which 8 had to be made good, those are in your 9 statement; these are all things that you 10 understood had to be dealt with by the IMPACT 11 Programme. 12 A. They would have been requirements that were 13 provided to me by the programme board or people 14 representing members on the programme board. So 15 I say Ann Cruttenden represented the Network. 16 We would have had people representing POID who 17 would be the main people who would have come up 18 with that as a requirement. 19 Q. So twinning these two things together, basically 20 you understood that the IMPACT Programme was at 21 least partly going to be designed to stamp out 22 the use of subpostmasters hiding discrepancies 23 and that they were nevertheless liable to make 24 good any shortfalls. These were two big drivers 25 for that part of the programme; is that correct? 98 1 That's what you understood? 2 A. Yeah, well, yes. Yes. While we were -- 3 I wouldn't say they were two big drivers. 4 I would say they were things that were put into 5 the programme. Because we had to make changes 6 to the Horizon System in order to feed the new 7 back office systems we were putting in, we would 8 have had workshops and, at those workshops, 9 there would have been the idea that we should 10 clamp down, as I would say, on the suspense 11 account. 12 Q. Right and these things -- I'm using your 13 words -- these things that were put into the 14 programme, who explained that these things 15 needed to be put into the programme? Who was it 16 that gave you those instructions? 17 A. It would be the key stakeholders. 18 Q. Right. In particular, who? You mentioned 19 Ms Cruttenden. 20 A. Well, as I say, I would have suspected that 21 there was -- there were three areas -- well, 22 Network, which I would include, you know, the 23 Network line, but also -- and I can't remember 24 whether POID were part of that Network line at 25 the time or whether they were in a separate 99 1 area -- and then it would have been from Finance 2 because, clearly, Finance -- and Peter Corbett 3 was the sponsor who was in Finance, who was 4 having to include those losses and, as we've 5 seen in the documents, we were talking earlier 6 about the scale of the losses that were in the 7 accounts, and that was a big driver for the 8 business at the time to reduce those. 9 So we would have -- but because Peter was my 10 sponsor, in the initial requirements workshop 11 those sort of things would have -- I imagine it 12 would have come from key stakeholders. I can't 13 remember the exact -- 14 Q. Right, and the way you described this earlier, 15 you described Peter Corbett as being a Finance 16 Director and you also referred to the fact that 17 you were told to do certain things. The way 18 I've recorded it was that you were told go "do 19 that" and the "do that" includes these 20 particular points, which relate to 21 subpostmasters, "Get rid of these subpostmasters 22 hiding discrepancies in the suspense account and 23 make sure that they're liable for all 24 shortfalls"? 25 A. Yeah, and some of it, because, as I say, because 100 1 I'd been business process architect and a number 2 of other things, they would be things that 3 I would have -- so I don't know where the 4 original idea came from. It could have come 5 from me, the idea, and then the Finance Director 6 said, "Yes, that's a good idea", or it could 7 have come from the Finance Director. I can't 8 remember the order of how it came about. 9 Q. Now, just I've only got one point before I move 10 on. I won't take long on this. This question, 11 this idea of subpostmasters hiding discrepancies 12 in their accounts. Again, across the IMPACT 13 documentation we've seen, there is no evidence 14 that is set out saying that "We've got 30 15 postmasters and mistresses, these are the names, 16 these are the dates, these are what happened, 17 doing all of this?" There's no reference to 18 a whole bunch of evidence of subpostmasters 19 doing this. Was there a document that you were 20 given that's part of your "go do that" programme 21 that set out evidence that demonstrated that 22 subpostmasters were doing this very thing? 23 A. No, but I'm assuming because there was values of 24 losses within there that those losses figures 25 were given to me by people, and there would have 101 1 been -- because POID were involved, they would 2 have had cases, numbers of cases, and done. So 3 there would have been all that sort of thing. 4 And remembering that, ultimately, the 5 requirement is signed off by the programme 6 board, they might have been created by me and my 7 team but they were owned by the programme board 8 and those areas were all represented on the 9 programme board. 10 Q. Right, okay. So this seems to be the situation. 11 You did not see any actual evidence that 12 demonstrated that subpostmasters were carrying 13 out this very programme of hiding discrepancies 14 in the suspense accounts? There was no document 15 that said, "We've got 60, 50 of these people 16 doing this"? Is that correct; is that correct? 17 A. I can't recall that. But I'm sure that people 18 gave me those -- well, it wasn't my decision to 19 do it anyway. As I say, I was programme 20 manager, so I was doing what the board, my 21 program board, wanted me to do -- 22 Q. Right. 23 A. -- which was documented in those requirements. 24 Q. So this is a two-step thing. First of all, you 25 did not see such a document setting out the 102 1 evidence that lurks behind this idea that 2 subpostmasters were going around hiding? Right. 3 A. I can't recall seeing a document but I'm sure 4 that I was provided with information from people 5 who I would have trusted to provide me with the 6 correct information. 7 Q. So are you assuming, is this your evidence now 8 before this Inquiry, that your programme 9 directors, the people that told you to "go do 10 that" had such evidence? 11 A. I would assume that. Yes. They were all decent 12 people. 13 Q. Okay, so that's your programming in terms of the 14 "get this done". I'll just briefly move on. 15 You've spoken about your knowledge of the system 16 and you've spoken about the system you believed 17 that was capable of rectification of any 18 problems within the system. Let's just take 19 that one step further. You've spoken about the 20 helpdesk, and you knew that the helpdesk was 21 there as a line of support for subpostmasters; 22 is that correct? 23 A. Yes. 24 Q. Were you aware that there were four lines of 25 support theoretically available to 103 1 subpostmasters? First of all, the helpdesk, 2 then beneath that, three other lines of support 3 as and when the technical problem got more and 4 more difficult? Were you aware of that? 5 A. Well, I'm assuming the helpdesk would reach out 6 to other people and experts. The helpdesk 7 themselves were not experts in the system or 8 anything like that. So the helpdesk would reach 9 out to whatever they -- 10 Q. Can we avoid assumptions for the moment. In the 11 time whereby you're starting off the IMPACT 12 Programme, let's find out what you knew at that 13 stage. So you've got a first line support which 14 the helpdesk answering the telephone call, yes? 15 A. Mm-hm. 16 Q. You're aware of that. Were you aware at that 17 time, during this early stage of the development 18 of the IMPACT Programme, aware that there were 19 three other lines of technical support for 20 subpostmasters? 21 A. I can't remember if there had been three other 22 lines. I think there was a Horizon -- was there 23 a Horizon Helpdesk in Fujitsu? I'm not sure. 24 No. So I'm not -- I can't recall but I would 25 have been aware at the time, I'm sure. 104 1 Q. You've been asked a good number of questions by 2 Mr Beer about your knowledge of bugs, 3 difficulties with the system, software problems. 4 So what I'm trying to establish with you is what 5 you knew about any system that was there to deal 6 with such problems. So you knew about 7 a helpdesk, in other words people taking 8 telephone calls? 9 A. Yes. 10 Q. Right. You weren't aware about three other 11 lines of support to deal with more and more 12 difficult problems? 13 A. Well, if you could describe what they were -- 14 Q. Yes, certainly I will. 15 A. -- I might say whether I can remember them or 16 not. 17 Q. By any chance you might have spoken to Gareth 18 Jenkins about this? 19 A. I know Mr Jenkins very well. 20 Q. Yes, we see a reference to him on the 21 documentation. During the period of time 22 whereby the helpdesk was in operation, there 23 were about 20 to 100 software designers and 24 developers working in fourth line support at any 25 one time. All of them were allocated bugs, 105 1 errors and defects, depending on their 2 experience and technical knowledge of the 3 component parts of the Horizon System 4 architecture. Did you know that? 5 A. Yes, so I knew there was a Horizon Helpdesk. 6 I didn't know the constitution of it. 7 Q. Did you know, I repeat, that there were about 20 8 to 100 software designers and developers working 9 in the fourth line support at any one time, all 10 of them were allocated bugs, errors and defects, 11 depending on their own experience with and 12 technical knowledge of the component parts of 13 the Horizon System? 14 A. No, I didn't know what the Fujitsu helpdesk 15 looked like or how it was structured. 16 Q. Now I've read that from the statement of 17 Mr Gareth Jenkins. 18 A. Yes, because Gareth was in Fujitsu. 19 Q. Now if at the time of your development of the 20 IMPACT Programme you'd have been told that we've 21 got four lines of technical support for 22 subpostmasters, including one of them, which is 23 the really when it's got sort of difficult line, 24 fourth line support, dealing with 20 to 100 25 software designers and developers, looking at 106 1 bugs, errors and defects, depending on their own 2 type of qualification, all going on at one time, 3 do you think that might have made you think 4 "Well, hang on perhaps we ought to just give 5 couple of thoughts to whether this system is 6 really robust or not"? 7 A. I wasn't aware of that, so I can't say what 8 I would have done -- 9 Q. If you had been told. 10 A. I don't know how I would have reacted if I'd 11 been told because I need to understand more than 12 what you're telling me, you know. I mean, all 13 systems have support teams behind them. So 14 I don't know. I can't answer that and I was 15 a business -- you know, I was a business owner 16 of the programme. I wasn't responsible for the 17 support desks and et cetera. 18 Q. It's possible it might have put you on notice 19 that you may need to ask a few more questions 20 about the nature of these bugs, defects, 21 software problems; do you agree? 22 A. Yes, if I'd been made aware of them I would 23 probably be -- particularly if they were ones 24 relating to accounting because that was really 25 what -- IMPACT was improved accounting, you 107 1 know, so if I thought that there were bugs, 2 et cetera, around accounting, I may have made 3 different decisions. But I wasn't aware of 4 that. 5 Q. Lastly, just on this, Mr Jenkins didn't tell you 6 about that, he didn't tell you "We've got this 7 whole entire system, including a fourth line 8 support system, with loads of people working on 9 it, dealing with bugs and defects"; he didn't 10 tell you that? 11 A. No, because my -- I would have -- so Torstein, 12 who now works with Fujitsu, was my technical 13 architect so he would be, if anybody, the person 14 talking to Gareth. But I can't say whether -- 15 I'm sure if Torstein had been aware of it then 16 he would have told me. But I can't say. Maybe 17 he wouldn't. 18 MR STEIN: Thank you, Ms Harding. 19 Questioned by MS PAGE 20 MS PAGE: Thank you, I have some questions for you 21 as well. My name is Flora Page and I act for 22 some of the subpostmasters. In your witness 23 statement, you've told us that the processes for 24 raising a dispute regarding a discrepancy were 25 not changed by the programme, the IMPACT 108 1 programme and branches were supported by the 2 helpdesk and/or the Network team, and they would 3 raise any concerns through those channels. Yes? 4 A. That was my understanding, yes. 5 Q. I think I understood from your answers just 6 a moment ago that you haven't watched any of the 7 testimony from the subpostmasters in Phase 1, is 8 that right, or read it? 9 A. No, I haven't. 10 Q. What I'd like to do is take you through some 11 testimony that talks about how the support that 12 you were relying on actually operated. So I'm 13 going to ask, please, for INQ00001035 to be 14 brought up, please. This is the testimony of 15 subpostmistress Ms Janet Skinner. We'll just 16 read through some of that evidence when it comes 17 up. 18 Can we go, please, to page 24. We'll start 19 on what's actually the sort of page 94 of the 20 testimony, so we can sort of zoom in, please, on 21 the top half of the page. I'll just read it out 22 if that's all right, so that it's clear for 23 everyone: 24 "When you experienced an apparent 25 discrepancy or shortfall shown by Horizon, what 109 1 did you do?" 2 So that's a question to Ms Skinner and she 3 responds: 4 "Well, you'd have to go through everything 5 that you had on hand, like giro things and 6 making sure your rems had been done correctly, 7 but the only thing you really had was what you 8 had in front of you, and I think what the 9 Horizon system installation did was it took away 10 all of the paperwork that you had that you could 11 cheque. So the only thing -- information you 12 actually had available was what the Horizon 13 system gave you. 14 "Question: To whom did you look for 15 assistance? 16 "Answer: The helpline. I rung the helpline 17 on numerous occasions. To be honest, when you 18 rang them, it was like they were reading from 19 a script. You could tell what they was reading. 20 It was something -- it was a script written 21 because of the way it was said to you over the 22 phone, and if they couldn't help you, they would 23 just say, 'Well, you just have to make it good 24 yourself'. 25 "Question: According to the judgment of the 110 1 Court of Appeal (Criminal Division), you 2 contacted the helpline 116 times between 1 3 January 2004 and 31 January 2005; is that right? 4 "Answer: Yes. 5 "Question: What advice did you receive? 6 "Answer: Basically the same: if the office 7 was short, it was my responsibility to make 8 good, that I was wrong. 9 "Answer: You've explained in your statement 10 that matters came to a head in 2006? 11 "Question: Yes. 12 "Answer: What happened? 13 "Question: The loss -- my office was 14 running at a loss of £40,000. I had a visit 15 from two retail network managers. There's only 16 one of them I can remember her name and it's 17 Diane Oyles, and when they came I took them in 18 the back, explained to them that my offices was 19 running at a loss. They asked if they could do 20 a cash check while they was there and they said, 21 'Well, what we'll do is we'll keep the office 22 open, we'll just do a cash check'. 23 "So I said, 'Well, to be honest, I'd rather 24 close it, have an audit and sort it out', and so 25 they did the cash check, we closed the office, 111 1 they removed the keys from me and told me I had 2 to meet them there the next morning at 9.00 with 3 the auditors. 4 "Question: How did you feel when you were 5 told to leave your branch? 6 "Answer: Devastated, absolutely devastated, 7 but I think I was more relieved, as well, 8 because I thought now, obviously, they're going 9 to get to the bottom of what's gone wrong. 10 I think you, sort of, you put your trust in them 11 because, as they keep portraying themselves, the 12 Post Office are a trusted brand and their trust 13 wasn't to find a problem that, it was to find 14 a solution, and I was the solution. 15 "Question: You have described employees of 16 the Post Office coming to your branch to carry 17 out an audit. What did they tell you about the 18 outcome of that audit? 19 "Answer: They just said that it was 20 actually running at a loss of 59,000 and I was 21 suspended without pay. I was searched before 22 I left and I was also informed that if I removed 23 anything from the office, it would be classed as 24 theft. So all paperwork, nothing, I couldn't 25 move anything. It all became the property of 112 1 the Post Office. 2 "Question: How did you feel when you 3 returned home later that day? 4 "Answer: I was devastated, absolutely. 5 Everything just crushed. You just feel 6 crushed." 7 I'll pick it up, if I may, by going to 8 page 26 on our system, and picking up on 9 page 101 of the page numbers that come up on the 10 screen: 11 "Were you aware when you attended court on 12 2 February that you might be sent to prison? 13 "Answer: No, I really didn't think I had, 14 because to go to jail, you've got to have 15 committed a crime and, if you can't prove that 16 you've committed a crime, because I couldn't, 17 I hadn't and they couldn't proved I had, but yet 18 I still went to jail. 19 "Question: Where in court were you standing 20 when the judge read out your sentence? 21 "Answer: In the dock. 22 "Question: What did the security guards do 23 when the judge sentenced you to prison? 24 "Answer: You're stood in the dock, and he 25 said that he was giving me a nine-month prison 113 1 sentence because of the amount of money that was 2 involved, because I'd stolen from the Crown, and 3 I heard the gate lock and then he told me that 4 it was a custodial sentence. I'll be honest, I 5 though he was going to say 'suspended' but he 6 didn't. 7 "Question: How did you feel when you were 8 escorted out of the courtroom? 9 "Answer: I was just an emotional wreck, I 10 don't actually remember. The only thing I 11 remember is being in the holding cell and then 12 going to Wakefield Prison. 13 "Question: How were you transported there?" 14 We'll just go over the page for a few more 15 lines: 16 "Answer: In -- well, they took me out in 17 handcuffs, put me in one of those police 18 transport things. 19 "Question: What happened upon your arrival 20 at prison? 21 "Answer: I was photographed, fingerprinted, 22 told to remove my clothes so they could strip 23 search me. So they put -- made me stand behind 24 the curtain and squat and then put a mirror 25 underneath to make sure you're not taking 114 1 anything in with you. 2 "Question: Did you have an opportunity to 3 speak to your children when you arrived into 4 prison? 5 "Answer: I was given a phonecall there in 6 the evening and my daughter wouldn't actually 7 speak to me because she was just -- she was an 8 emotional wreck and I didn't want to speak to 9 her, to be honest. I felt so ashamed. I was 10 supposed to be there to protect them and show 11 them what -- you do the right thing but yet you 12 do the right thing and I went to jail anyway. 13 "Question: Were your children able to visit 14 you in prison? 15 "Answer: I didn't want them visiting me. 16 That was the only thing I could control and it 17 was hard enough for me to have that memory of me 18 being in jail, they weren't having that memory 19 of me being in jail. So I refused to see them." 20 So what I've just read out there is the 21 testimony that explains both the so-called 22 support that was offered by the helpdesk and the 23 Network team and the devastating consequences of 24 that so-called support or lack of on the 25 subpostmasters. 115 1 Now, the processes you say for raising 2 disputes didn't change, but perhaps the point is 3 that they should have changed. So what I want 4 to understand is, given that your programme 5 board was putting in place a system that would 6 remove the suspense account facility so that 7 subpostmasters had to rely on these helpdesk 8 calls and these so-called Network support 9 processes, did anyone on the programme board 10 offer any kind of challenge to the people 11 responsible for those systems, any kind of 12 constructive challenge to find out whether those 13 processes were working? 14 A. No, I don't recall anybody. I mean, I got a set 15 of requirements from the business, you know, 16 through a series of Network -- series of, um, of 17 workshops, and I don't recall, you know, 18 anybody -- in fact, you know, the requirement 19 would have come from the Network to remove the 20 suspense account. 21 Q. Who was that, then? Who was responsible for the 22 Network? 23 A. I can't remember who was represented at the 24 time. 25 Q. Who was on your programme board responsible for 116 1 the Network? 2 A. I can't remember. I'd have to -- 3 Q. Do you want to take a bit of time to think about 4 it because you're saying that this is where it 5 came from; this is who wanted it. 6 A. No, I can't -- I mean, Peter Corbett was my 7 sponsor and it may have come from Peter. He was 8 in Finance, obviously the impact of that. And 9 I know Ann Cruttenden was the representative at 10 the next level. But I -- I'm really, really 11 sorry. I mean, my memory is -- I have issues 12 with my memory and I just can't remember who was 13 the Network representative on the programme 14 board. 15 But, you know, it would have been signed 16 off -- there was some documents right at the 17 beginning, the business case would certainly 18 have been signed off at -- well, it was signed 19 off at the group board level. So it would have 20 gone through all of that process. 21 Q. Did Peter Corbett offer any sort of challenge to 22 find out whether the Network support was as it 23 should have been? 24 A. Not -- I can't remember. 25 MS PAGE: Well, thank you. Those are my questions. 117 1 A. Okay. 2 SIR WYN WILLIAMS: This programme board that you're 3 giving evidence about, I don't expect you to 4 remember precisely, but just help me a little. 5 Was this a body which was three, four, five 6 people or a larger body? 7 A. It would probably have been about five or six 8 people. So Peter Corbett was the sponsor so he 9 would have been on there but then the Network 10 would have been represented, and certainly 11 Peter -- 12 SIR WYN WILLIAMS: Can I just stop you there. So 13 Peter Corbett and we've got something specific 14 about him, he's the Finance Director, so if 15 I imagine a Finance line so to speak, he would 16 be the senior person on the board in Finance? 17 A. Yeah, because remembering that the programme was 18 primarily about implementing new system in the 19 back office, a new SAP system, so the losses bit 20 was not a -- was a -- was a byproduct of the 21 fact that we had to alter the Horizon System. 22 So it wasn't particularly a big network impact 23 project. 24 SIR WYN WILLIAMS: No, in general terms, his main 25 concern would have been to reduce losses, but 118 1 reducing losses was only one part of what it was 2 all about. 3 A. Absolutely, which you saw in the document. 4 SIR WYN WILLIAMS: Right. I've got that. So give 5 me an idea of the equivalence of Peter Corbett 6 in other spheres, if you would, so I can 7 understand the make-up of this board? So you'd 8 have had someone of similar seniority to Peter 9 Corbett from which other departments, if I can 10 put it in that way? 11 A. I honestly can't -- I'm really sorry but 12 I can't -- we would have had somebody from 13 Network. Ann Cruttenden was the next level down 14 because she was the business change manager from 15 Network and -- 16 SIR WYN WILLIAMS: I'm sorry to stop you again, but 17 when you say next level down, so in Network 18 terms, there'd be -- are you saying there'd be 19 someone of Peter Corbett's seniority but she was 20 one below that? Is that what you're trying to 21 convey? 22 A. No, I was -- I'm not saying she was one below 23 the Network Director. What I'm saying is the 24 programme board may or may not -- it may have 25 not had a director from Network because, as 119 1 I say, the change in the Network was nothing 2 like the change in Finance so Peter Corbett ran 3 the programme board. There would have been 4 somebody from Network because there were 5 implications on Network. Ann Cruttenden was the 6 business change manager for the Network so she 7 was the main -- she was the person involved but 8 she wasn't a board member in that sense. 9 I'd have to find a document that had got the 10 names of the board and apologies for not 11 remembering. 12 SIR WYN WILLIAMS: All right. So if I wish to 13 enquire into this more closely, I'm looking for 14 the names of four or five other people apart 15 from Mr Corbett, by the sounds of it, roughly? 16 A. I'd have to find a document. I honestly can't 17 remember back that far. 18 SIR WYN WILLIAMS: I'm sorry to take you back to it 19 again because you've been asked many questions 20 about paragraph 31, one way or another, about -- 21 but I'm still slightly bemused, I have to say, 22 about the use of the word "hide". You've put it 23 in inverted commas. What do you actually mean 24 when you say that, historically, the suspense 25 account was used to hide discrepancies? Explain 120 1 it to me in your own words. What was happening? 2 A. So there's -- there's no doubt that in the past 3 there had been people who stole from post 4 offices, people who ran post offices -- 5 SIR WYN WILLIAMS: Can I just stop you there. 6 I understand that to be, as a matter of fact, 7 correct. But from my perspective at least, the 8 last thing that anybody who had actually stolen 9 money from the Post Office would do was to 10 signal a loss by putting it in the suspense 11 account, which is what's bemusing me a little. 12 A. Okay, so they're either going to inflate their 13 cash and stock balances, but that would 14 trigger -- that would have triggered an earlier 15 investigation, because we had a big focus on 16 cash management. So nobody wanted a branch to 17 own significant cash. So if you've lost -- if 18 you've been taking something, at the end of the 19 day you've got the transactions. The system has 20 recorded, and then you have a declaration of 21 cash and stock. 22 So the only way to cover up the difference, 23 because clearly you won't have the cash and 24 stock in the system as you should have, and 25 you're meant to do a translation, then, of the 121 1 cash and stock, yeah? So the only way to hide 2 a discrepancy is either to inflate your cash and 3 stock figures, and cash is probably -- it would 4 have been an easier one in the system than 5 stock, because stock is all about stock in, 6 stock out, et cetera -- so you would either say 7 "I've got more cash than I really have", or you 8 would report the cash you really have, and put 9 the difference in the suspense account. 10 It's an accounting -- double entry 11 accounting system, so we have to make it balance 12 somehow. So if, when you actually count your 13 cash and stock, it's not what you expected, you 14 either declare a loss or you try and put it 15 somewhere and the suspense was a place that you 16 could put it because you couldn't alter the 17 other transactions because they were system 18 generated from having done the transactions. 19 Does that make sense? 20 SIR WYN WILLIAMS: Well, all right. You also used 21 the word "historically" in that paragraph, which 22 conveyed to me, when I first read it, that that 23 was something that had happened in the past, but 24 wasn't necessarily happening at the time, 25 ie 2003 to 2006. 122 1 A. It didn't mean that. 2 SIR WYN WILLIAMS: Should I take that view or not? 3 A. No, you should take the view that that meant 4 that the evidence would have been historical, 5 but then it was expected that that was still 6 what they would do. 7 SIR WYN WILLIAMS: But that's surprising me, if 8 I can put it in that way. The fact that there 9 may have been illegal activity in the past isn't 10 a necessarily safe foundation for concluding 11 it's subsisting at the time, is it? 12 A. No, I wasn't saying it was subsisting; I was 13 saying historically that had happened and 14 therefore removing a way for that to continue to 15 happen was why we had that requirement. 16 SIR WYN WILLIAMS: Okay. Thank you very much for 17 coming to give evidence, Ms Harding. 18 MR BEER: Sir, thank you very much. Can we say 19 1.50, please, this afternoon? 20 SIR WYN WILLIAMS: Always assuming Mr Burley is 21 ready. I had some information that he wasn't 22 ready until 2.00 but if you know better, then 23 we'll say 1.50. 24 MR BEER: I never know better. But if he is 25 available can we say 1.50? 123 1 SIR WYN WILLIAMS: Yes, so everybody should be 2 prepared to start at 1.50, to use those that 3 famous phrase: not before 1.50. 4 MR BEER: Thank you very much. 5 (12.55 pm) 6 (The Short Adjournment) 7 (2.00 pm) 8 MR BLAKE: Good afternoon, sir. 9 SIR WYN WILLIAMS: Good afternoon. 10 MR BLAKE: Can I call Mr Mark Burley, please. 11 MARK BURLEY (affirmed) 12 Questioned by MR BLAKE 13 MR BLAKE: Thank you very much. Can you give your 14 full name, please. 15 A. Yes, Mark Richard Burley. 16 Q. Mr Burley, I'm going to ask for your statement 17 to be brought up on screen and you should be 18 able to see it, it's WITN03850100. Is that your 19 statement? 20 A. It's very blurred at my end. That's better. 21 Page 1 definitely looks good, yeah. 22 Q. Do you have a copy of it in front of you? 23 A. I do, yes. 24 Q. On 4 January, is it right to say that you signed 25 a statement of truth that was separate to that 124 1 statement but it confirmed the truth of the 2 content of that statement? 3 A. I couldn't confirm the exact date. Is it on my 4 signed copy? No. It's not on my copy either. 5 I can't confirm the exact date but it's around 6 about that time, yes. 7 Q. Yes. Is this statement true to the best of your 8 knowledge and belief? 9 A. It is, yes. 10 Q. Thank you. That statement will go into evidence 11 and the questions I will ask you today will be 12 supplementary to that. 13 Your statement is quite light on detail so 14 I'm going to start by asking you a little bit 15 about your background. When were you first 16 employed by the Post Office? 17 A. 1985. It could have been -- I started on 18 31 December, so I'm just thinking was it '85 or 19 '84? It might have been '84. 20 Q. Okay. Did you have another job before that, or 21 was that ... 22 A. I worked for Sainsbury's before that, just to -- 23 as -- I mean, technically, it was a part-time 24 role but I used to do full time hours for them. 25 Q. Can you give us an indication of the types of 125 1 roles you held at the Post Office before 2 becoming involved in HNG-X or what we know as 3 Horizon Online? 4 A. Yeah, I started off as a post office counter 5 clerk. That was my first role. I then moved on 6 to be an area manager support. I then moved on 7 to be a branch manager at three different 8 branches. I then moved on to an internal audit 9 role, auditing various processing systems across 10 the business. I then moved on to be a retail 11 network manager, essentially accountable for the 12 performance of around about 35 of the largest 13 sub post offices. 14 I then moved into project work, held 15 a number of project manager positions and then 16 in 2003 I set up a PMO function, project 17 management office function, before becoming the 18 HNG-X programme manager in 2005, which I held 19 until 2009 before becoming head of delivery, and 20 then I left on a voluntary redundancy term in 21 early 2011. 22 Q. Was your role as head of delivery until 2011? 23 A. It was, yes. 24 Q. Can you tell us what the role of head of 25 delivery involves, briefly? 126 1 A. Essentially, it was to look across the whole of 2 the technology delivery that Post Office were 3 doing at that time. So HNG-X was one of those 4 projects that came under that, but I was no 5 longer the programme manager for the programme. 6 Q. Thank you. You've said in your statement that 7 you had programme management qualifications. 8 Did you have any other qualifications that are 9 relevant, a computer science background, or 10 anything along those lines? 11 A. I didn't have computer science background, no. 12 I had lots of project manager qualifications, 13 PMO qualifications and, as you say, programme 14 manager, business analyst qualifications and -- 15 but not a computer science background, no. 16 Q. I'm going to go broadly chronologically this 17 afternoon. I want to start with Legacy Horizon. 18 In your witness statement, you say you weren't 19 aware of any faults with Legacy Horizon but 20 there were bugs that were investigated and 21 fixed, to the best of your knowledge. What do 22 you see as the difference between a fault and 23 a bug or a series of bugs? 24 A. Well, I'm not even sure I said that in my 25 statement. Could you tell me where I said that 127 1 in my statement because I wasn't that close to 2 original Horizon. I was only on the periphery 3 of original Horizon. 4 Q. Your statement is quite brief on that. Let's 5 have a look at page 2. So it's WITN03850100. 6 It's the top 2 paragraphs that really address 7 Legacy Horizon. 8 A. Yeah, and point 6 that I wasn't aware of any -- 9 Q. You say: 10 "I was competent in the use of Legacy 11 Horizon and was involved in an assurance 12 rollover its development. 13 "I wasn't aware of any 'faults' with Legacy 14 Horizon but was aware it was essentially 15 a modified EPOS solution and as such errors on 16 data input could be made with consequential 17 impacts on weekly accounts." 18 So that's a suggestion that human errors in 19 input could have a consequential impact: 20 "As with any technology solution, bugs did 21 occur but these were investigated and fixed -- 22 appropriately (to the best of my knowledge)." 23 So you have distinguished there between 24 faults and bugs. 25 A. Yeah, I mean, what I'm trying to distinguish 128 1 there is when original Horizon went live, 2 I wasn't aware of any overarching faults with 3 the system that could cause a real problem. 4 A bug is something that may not be technically 5 perfect but doesn't cause a problem because 6 there's either a workaround or it's not actually 7 having a direct impact. 8 So, for example, I would describe a bug in 9 this sense to be something in Arial font rather 10 than Times New Roman, for example. It doesn't 11 really have any real impact on the user. Are 12 you still there? 13 Q. Yes. So your evidence is that there were some 14 problems but they were ones that didn't really 15 have any real impact on the user? 16 A. Yeah, as I recall it, on the original Horizon 17 System, it went through quite an extensive 18 period of what they called acceptance, and that 19 was validating that what the Post Office had 20 asked for had indeed been delivered, and there 21 were numerous examples and I did support some of 22 these where there was a more serious error or 23 a more serious bug, whichever word you want to 24 use, that we would declare wasn't fit for 25 acceptance. In other words, we weren't prepared 129 1 for the system to go live whilst that bug 2 persisted because it could have a more 3 significant impact on a user. 4 Q. Your evidence from a moment ago is that you had 5 quite a peripheral role in relation to what we 6 know as Legacy Horizon; is that right? 7 A. I'd have supported a few of the examples of 8 working through the acceptance criteria. 9 Q. We'll go through some of the documents, and your 10 name does appear on quite a few documents, 11 concerning Acceptance Incidents in relation to 12 Legacy Horizon. Let's look briefly at a couple 13 by way of early background. We can look at 14 FUJ00058445. This is a document from 15 March 1999. This is a model office testing 16 document. If we go over the page, it has you on 17 the circulation list and you're listed as being 18 part of the TIP project. We know TIP as 19 Transaction Information Processing? 20 A. That's right. 21 Q. Do you remember your role in relation to that 22 project? 23 A. I don't remember the detail of the role but 24 I would have been supporting it from a business 25 assurance perspective, so making sure that the 130 1 system being developed operated according to the 2 required business controls. 3 Q. So when it came to testing, you were part of the 4 TIP project, and your role was in relation to 5 the business requirements for the TIP project; 6 is that fair? 7 A. Yes, that's right. 8 Q. Then we'll look at another document POL00090073. 9 This a document from June 1999. It's headed 10 "Horizon Project -- Acceptance Review Comment 11 Sheet", and it has your name there in the top 12 right-hand corner, as part of a panel in 13 relation to the acceptance review. Do you 14 remember your involvement in that panel? 15 A. Well, this panel, this acceptance sheet -- 16 I don't recollect a panel as such -- but this 17 acceptance sheet would probably relate to one of 18 the acceptance items that I was asked to look 19 at. 20 Q. We've heard a lot of evidence, including expert 21 evidence, of serious problems during acceptance 22 and the rollout of Horizon. Do you accept that 23 there were serious problems during the period of 24 your involvement in the acceptance process and 25 rollout of Horizon? 131 1 A. All I can say on that is that the elements of 2 acceptance I was involved in, we did the right 3 degree of due diligence and made the 4 recommendation accordingly as to whether it 5 would either be accepted into -- you know, as 6 a minor bug that wouldn't cause an impact or 7 needed to be fixed and therefore would be fixed 8 before we proceeded. 9 Q. Let's look at the ones that you were involved in 10 or certainly that the documents show you were 11 involved in, let's look at POL00090478. This is 12 a document the Inquiry is familiar with. Can we 13 look at page 75 it's a list of various different 14 Acceptance Incidents. I'm not going to spend 15 too long on Legacy Horizon, I'll just take you 16 through a few of these documents. 17 So it's page 75 and that addresses 18 Acceptance Incident 410. I'm just going to read 19 what it says about Acceptance Incident 410 and 20 it's got your name down as the owner there. It 21 says: 22 "TIP have detected an instance where 23 transactions received in the daily transaction 24 file are not represented on the electronic cash 25 account at the weekend. 132 1 "The transactions missing from the cash 2 account are associated with a product changing 3 from core to non-core." 4 Do you remember that Acceptance Incident at 5 all? 6 A. I don't remember the Incident, no but that would 7 have been, in my view, you know, a high priority 8 Incident that needed resolving. 9 Q. Do you know why you're listed as the owner of 10 that particular Incident? 11 A. No, I can't tell you why I was listed as the 12 owner. I may have been the one leading the 13 investigation into it but I can't recall what at 14 this time, and that's 24 years ago, nearly. 15 Q. Yes. I mean, you would accept that issues with 16 the cash account are quite fundamental to the 17 workings of the Horizon System? 18 A. Absolutely critical. I totally 100 per cent 19 agree and that's why, as I suggested, if I was 20 investigating this one, it would have a severity 21 of -- using the scale on the right, it would 22 have a severity of high. It would absolutely 23 have to be understood before it was, you know, 24 accepted to proceed, or indeed fixed. 25 You know, until you get into the detail of 133 1 these things you don't know the ins and outs of 2 what's truly happened. It could be that it was 3 represented on a different line than the one it 4 was expected to be on just from, you know, 5 a typo in the coding. But, obviously, that's 6 part of the investigation that would need to 7 take place on all these types of incident. 8 Q. Can we look at POL00030393, please. This is now 9 moving from July 1999 to August 1999. This is 10 an electronic memorandum from Ruth Holleran; do 11 you remember Ruth Holleran? 12 A. Yes. 13 Q. What was her role? 14 A. As I recall at this time, I think Ruth led the 15 acceptance process, I believe. I couldn't be 16 100 per cent certain on that, so it would need 17 checking factually, please. 18 Q. Thank you. You're listed there. Can you see 19 you're copied into that memo, your name -- 20 A. Yeah. 21 Q. -- appears? This is a list of high impact, high 22 severity Acceptance Incidents. If we can turn 23 to page 3, please, in the second column there's 24 reference to 376, that's Acceptance Incident 25 376, that's something the Inquiry is quite 134 1 familiar with now and that relates to data and 2 integrity issues. It's described there as: 3 "TIP derived cash account not equal to 4 electronic cash account received by TIP." 5 As part of the TIP team, do you recall that 6 particular Acceptance Incident at all? 7 A. No, I don't recall the specific Incident, no, 8 but clearly that would be a high severity issue 9 that we would not want to go live with. 10 Q. Having been -- we've seen part of that TIP team, 11 is that something you would likely have been 12 familiar with at the time? 13 A. Possibly. I can't remember the specifics of 14 which ones I was asked to get involved with and 15 which ones I wasn't. As I said, I was involved 16 in some of the Acceptance Incidents but only 17 a subset. 18 Q. If we turn over to page 5, we get to Acceptance 19 Incident 410, which is the one that you 20 mentioned you were listed down as the owner of, 21 and that's the second column. If that paragraph 22 there could be expanded, please. It says: 23 "410 24 "This incident was previously recorded under 25 AI376." 135 1 So it appears they were linked: 2 "However as the effect is different, this 3 now stands alone. It refers to a scenario 4 resulting in the TIP derived cash account, 5 calculated from Pathway's daily transaction 6 files, exceeding the transaction totals in the 7 Pathway cash account. (Refer to description of 8 AI376 for more detail on this process). It is 9 this latter cash account stream that feeds the 10 POCL accounting process via CBDB." 11 Does this assist you at all in recalling 12 this particular Acceptance Incident? 13 A. It doesn't assist me in recalling the incident, 14 no. All I can say is I repeat what I said, that 15 anything that impacted the cash account and data 16 potentially being missing would have to be fully 17 explained before I would recommend proceeding. 18 Now, you know, as I say, I can't remember this 19 specific one, but we'd have made investigations 20 of both the Pathway solution and also the TIP 21 solution. 22 Q. Because that kind of an Incident, that's not 23 just a change from Times Roman Numeral to 24 another font? 25 A. Oh, no -- 136 1 Q. That's quite a significant -- 2 A. This is something that, you know -- you know, as 3 I said, I did not manage or lead the acceptance 4 process but my recommendation would have been 5 not to proceed with an Incident like this being 6 open. 7 Q. You say "would have been". Was it or is that 8 something you don't remember? 9 A. I can't recall. I mean, I can almost certainly 10 say that -- well, I can certainly say my 11 recommendation would have been not to proceed 12 with this being open without a satisfactory 13 explanation. I can certainly say that, you 14 know, as I'm on oath. I can't say anything 15 about the specific, because I don't recall it. 16 But, you know, looking at the detail, I'm, you 17 know, trying to be absolutely honest with you. 18 Q. Thank you. Can we turn to page 6, please. 19 Thank you. If we look at the fourth column 20 there, it summarises the current position and 21 a concern of Post Office Counters limited. It 22 says: 23 "POCL also require evidence that integrity 24 checks in support of R818/08 are sufficient for 25 all business data. In the absence of this 137 1 evidence, the integrity of the data that drives 2 the central accounting process is in question. 3 This incident is therefore adjudged high in line 4 with AI376 and AI411." 5 Is that consistent with what you've just 6 told us about the significance of this kind of 7 an Acceptance Incident? 8 A. That's absolutely exactly what I'm referring to. 9 It would be adjudged high and we would not 10 proceed. 11 Q. Looking at this type of a document, it seems as 12 though from early 1999 you were aware of the 13 importance of accounting data and that issues 14 were being experienced in this regard with the 15 new system; is that fair? 16 A. During its development, yes. 17 Q. Can we look at POL00043681, and can we look at 18 the second page of that, please. This was 19 a management resolution meeting of 12 August 20 1999. You weren't at that -- sorry, if we could 21 zoom out. Thank you very much. 22 You weren't at this particular meeting, but 23 it does address the Acceptance Incident you were 24 listed as the owner of, 410. Can we look at 25 page 3, please. It's the first substantive 138 1 line, it says "POCL advised", if we could look 2 at that or highlight those two lines, please. 3 It says: 4 "POCL advised that they rate 410 and 411 as 5 Low conditional on resolution of the controls 6 proposed on 376. RH would provide further 7 detail on this." 8 So it seems as though 410 becomes a low 9 issue because it was going to be in some way 10 tied up with resolution of the controls proposed 11 for Acceptance Incident 376. 12 A. Could you just zoom back out to the full page 13 because I thought there was something on the 14 page that might explain that. Yeah, if you look 15 lower down the page, under point 3, it says: 16 "JD reported as follows: 17 "Pathway recognise that not all transactions 18 had been harvested and sent to TIP. 19 A provisional fix went in on 2nd August and this 20 has worked satisfactorily so far with the effect 21 that all records had been sent." 22 So my assumption would be at this stage -- 23 as I say, I'm trying to remember back over 24 23 years -- but my assumption would be at this 25 stage so they'd put a fix in, we'd done some 139 1 provisional early testing of that fix and all 2 looked good but naturally, on something so 3 significant, we wouldn't want to close it until 4 we'd done sufficient testing to be confident 5 that it's truly fixed. I'm sure you probably 6 appreciate that there's lots of different 7 scenarios that contribute to a cash account. So 8 we'd want to make sure all those scenarios are 9 suitably tested. 10 Q. Absolutely. If we look at that paragraph, it 11 then goes on to say: 12 "Pathway cannot guarantee however that all 13 problems have been trapped. They will need to 14 see evidence from the fix of 8 known problems 15 and will continue to monitor the problem for 16 3 months to be confident of its resolution." 17 So at that stage they can't guarantee that 18 all problems have been trapped; does that mean 19 fixed or resolved? 20 A. No, as I say, we would want to do sufficient 21 testing to have a degree of confidence that they 22 have been and then, as you say, keep it under 23 monitor. You know, that's standard practice. 24 Q. Do you recall that there was a second 25 supplementary agreement in September 1999 which 140 1 permitted errors of up to 0.6 per cent of all 2 cash accounts? 3 A. I don't. I'm not aware of that at all, no. 4 Q. Perhaps if we go to POL00090428, and it's 5 page 21. This is the Second Supplemental 6 Agreement between the Post Office and ICL. You 7 may not have seen this because this is 8 a contractual document. I don't know if this 9 jogs your memory at all? 10 A. I think it's one of the documents I have been 11 sent but I wasn't party to this and I don't 12 recall it -- I'm just looking -- 13 Q. If we look at (i), there is a period in which: 14 "... the percentage of Cash Accounts 15 received by POCL across the TIP Interface 16 containing Cash Account Discrepancies shall not 17 exceed 0.6 per cent of all such Cash Accounts." 18 Is that something -- 19 A. It's not something I'm aware of and, you know, 20 I can interpret what it may mean but I may be 21 taking it out of consequence. This could refer 22 to the fact that we wouldn't pay Fujitsu -- or 23 sorry, Pathway, as they were called then, unless 24 it was in that parameter but that's not saying 25 that 0.6 per cent was acceptable because I can't 141 1 believe we'd have agreed to anything being 2 acceptable, is my personal view. 3 Q. Was there, do you recall, an acceptance that 4 there would still be some degree of cash account 5 discrepancies moving forward? 6 A. No. No. There definitely was no acceptance of 7 that. There may have been an acceptance that we 8 believed the fix had resolved the issue, but 9 there is inevitably a risk that something else 10 does arise because you cannot test the system to 11 a degree where you can 100 per cent hand on 12 heart say it will never occur again, and that's 13 what this may relate to, that it is, you know, 14 such a rare transaction or rare sequence of 15 events that causes it that it would be 0.6 per 16 cent. But I can't believe we accepted or 17 anybody would have signed up to accepting that 18 there could be discrepancies on a cash account, 19 if that was the fundamental accounting document 20 to which all branches were held. 21 Q. Were you still part of the TIP team at this 22 stage? So this is 24 September 1999. 23 A. I honestly can't recall if I was on TIP or on 24 a project at this stage. I'm guessing I would 25 have still been on TIP because it's only 142 1 a couple of months after the last document you 2 showed me. 3 Q. Would you have been expected to be consulted 4 about a contractual agreement of this kind, 5 given your involvement in TIP and being 6 responsible for business aspect of TIP? 7 A. I think the answer will be, yes, if it referred 8 to an ongoing allowance but no if it refers to 9 an allowance, you know, over the period of time, 10 such that we would pay Pathway as long as they 11 met this criteria but wouldn't accept it as an 12 incident. I personally believed that's what it 13 would be. It would be a contractual payment 14 term, not a business acceptance term. 15 Otherwise, I would expect to be consulted, yes. 16 Q. Do you remember a Third Supplemental Agreement 17 or were you aware of a Third Supplemental 18 Agreement on 19 January 2000? 19 A. I wasn't aware of a second so no, I wasn't. 20 Q. Let's look at that. FUJ00118186. I'll only 21 take you to that document very briefly. It's 22 page 4 of that Third Supplemental Agreement. It 23 addresses enhanced integrity control, which was 24 something that was implemented to address the 25 concerns relating to the cash account that we've 143 1 already gone through. Do you remember something 2 called the Enhanced Integrity Control at all? 3 A. No, nothing that I would have dealt with would 4 have been called Enhanced Integrity Control. It 5 might have been called a fix, it might have 6 been, you know, the process they were coding to 7 ensure that the cash account was robust but 8 I don't remember it being called an Enhanced 9 Integrity Control. 10 Q. If we go over the page, it's paragraph 5.3 of 11 the next page that the Inquiry has gone over 12 some time over previously in Phase 2. This 13 seems to be an agreement, I'll read it to you. 14 It says: 15 "The Contractor shall, from the date of this 16 Agreement until the end of the TIP Integrity 17 Checking Period make available to POCL promptly 18 upon request appropriate experts to explain to 19 POCL the Contractor's analysis of all root 20 causes of Cash Account Discrepancies and the 21 measures which the Contractor shall have 22 implemented in order to prevent the recurrence 23 of any Cash Account Discrepancies which would 24 not have been detected by the Accounting 25 Integrity Control Release." 144 1 Do you recall at all there being 2 a likelihood that there would be cash accounting 3 discrepancies which couldn't be detected by 4 a control that was put in place? 5 A. I don't recall anything, no, because the TIP 6 system was designed to ensure that what was sent 7 in reconciled with the -- you know, the two 8 sides fully reconciled so, no, I don't recall 9 that at all. Not that -- in other words, 10 I don't recall that it wouldn't have been 11 detected so I'm unclear what that refers to. 12 Q. Do you recall an acceptance that, because of the 13 nature of the system, there would always be 14 a degree of unreliability with the cash account? 15 A. Sorry, just phrase that question again? 16 Q. Sure. Were you aware or do you recall, that 17 because of the nature of the system, there was 18 an acceptance that there would always be 19 a degree of unreliability in the cash account, 20 however small? 21 A. No, I don't agree that, I don't accept that. 22 The cash account is only formulated from all the 23 inputs to it and there's no reason that it would 24 ever have to accept or be prone to errors. 25 There's no reason. Effectively, every 145 1 transaction that is done over a post office 2 counter has double entry on the cash account. 3 Q. Can we look at POL00028507, please. We've now 4 reached January 2000 and this is an email, the 5 bottom email, that is sent by you to Min 6 Burdett, regarding something called "Receipts 7 Not Equal to Payments". That's what we know as 8 Acceptance Incident 211, where the receipts and 9 payments didn't equal on the cash account. 10 A. Yes. 11 Q. Again, do you remember being involved in that 12 issue? 13 A. I vaguely remember this one, yes. 14 Q. Because that's obviously quite a significant 15 thing if the receipts and payments don't match. 16 What's your recollection of that? 17 A. Well, as it says on the subject, the receipts 18 weren't equal to payments and the fundamental 19 thing about the cash account is your receipts 20 payment and your payments total have to match, 21 and any then balancing entry between the two is 22 the amount of loss or gain for the week in 23 question. And, typically, every branch would 24 make, you know, a few pence, a few pounds 25 potentially, difference on a weekly basis. You 146 1 know, you accidentally give out, you know, a 20p 2 instead of 10p, so you've got a 10p discrepancy 3 straight away and there were -- 4 If I recall at the time, there were examples 5 where it genuinely happened because -- and 6 I think it was books of stamps. If they bought 7 ten stamps they got a slight discount as 8 a customer but it went in effectively at the 9 price -- so the subpostmaster would make, 10 I don't know, 10 pence, 15 -- I can't remember 11 what the product was, so don't quote me on 12 stamps, please, but there was a product that was 13 sold where, if you bought in bulk, you got a bit 14 of a discount. But it worked in favour that you 15 made actually, a slight profit on those. 16 So, you know, that fundamental part of the 17 cash account, the two columns are totalled, and 18 any difference is your loss or gain on the week. 19 It becomes a balancing entry. 20 Q. We've gone through quite a few of those 21 Acceptance Incidents. Am I right to say that 22 your evidence, certainly your oral evidence 23 today, is that, despite those, as far as you 24 were aware, the cash account was flawless upon 25 acceptance and rollout? 147 1 A. I believe the cash -- you know, my 2 recommendation would be that all cash account 3 errors, as part of the development, have to be 4 either 100 per cent explained, and therefore 5 understandable and wouldn't cause an issue to 6 any integrity, or, ideally, completely fixed 7 ahead of any rollout. 8 Q. In the passage I took you to earlier from your 9 witness statement, you said that the system was 10 essentially a modified EPOS solution. Were you 11 aware of any concerns about the EPOS system and 12 the coding of the EPOS system, for example? 13 A. Um, it depends what you mean by was I aware of 14 any concerns. During the development there 15 were, you know, numerous defects identified, you 16 know, that would have been identified through 17 testing. Some of it identified through 18 Pathway's own testing, some through Post Office 19 testing and they would all go in this part of -- 20 you know, ultimately to try to get fixed and, 21 ultimately, against the acceptance criteria, if 22 they remain open. 23 Q. What point in time did this all get resolved, do 24 you think? 25 A. To the best of my knowledge and from my part in 148 1 it, I believe it was all resolved before, 2 certainly, any extended trial was done. It may 3 have been -- there may have been a few small 4 issues outstanding at the onset of the first 5 trial. I can't 100 per cent confirm they 6 weren't, but sometimes that does happen. But 7 before it went out to any extended number of 8 sites, I would believe they would all be 9 resolved, such that integrity-wise, it was 10 robust. 11 Q. If we've heard evidence to the contrary, would 12 that surprise you? 13 A. It would surprise me as far as my involvement 14 would go. I would just say one thing. Of 15 course the system, once the system is live, it's 16 not completely left alone because we do -- you 17 know, like most systems, you develop new 18 products, so you do make changes to it. So 19 it's -- you know, could something have happened 20 after that? I can't comment for certain, but 21 possibly. 22 But if you're saying to me original 23 integrity issues were allowed to be rolled out, 24 were allowed to be included in the rollout, I'd 25 be 100 per cent surprised and shocked. 149 1 Q. At the beginning of your evidence today you said 2 you had quite a peripheral role in Legacy 3 Horizon. Having looked at those documents and 4 your name on various documents including 5 ownership of an Acceptance Incident, do you 6 think that was underplaying things somewhat? 7 A. No, because I wasn't looking at the development 8 to the Horizon System as a full-time role. You 9 know, I was involved with acceptance but 10 I wasn't leading acceptance. I was potentially, 11 you might describe it as leading a few of the 12 items under acceptance. 13 Q. Once you had finished your work on the 14 acceptance of Legacy Horizon, did you go on to 15 a different role before getting involved in what 16 we know as Horizon Online? 17 A. Yes, yes. 18 Q. I'm going to move on now to the development of 19 HNG-X. We'll see, in documents we're going to 20 come to, references to a medium volume pilot and 21 a high volume pilot. Can you tell us briefly 22 what that means, what the difference is between 23 the two? 24 A. It's just the scale of which we would roll out 25 to the number of sites. So I can't remember the 150 1 numbers but a medium volume pilot might be 2 extending from an initial pilot of ten sites up 3 to, let's say, 150, 200 sites, and the high 4 volume would be where we're migrating from that 5 and perhaps doing 100 sites a night to roll out 6 the entire estate. 7 Q. You weren't a technician. You weren't -- as you 8 have explained, you didn't have a computer 9 science background or anything of that sort. 10 What was your role in relation to the pilots? 11 Was it driving things forward? 12 A. So my role as HNG-X programme manager was to 13 ensure we delivered a quality system. 14 Q. Let's start in January 2010. Can we look at 15 FUJ00092754. This a document from 28 January 16 2010. It's the "Notes of the Horizon Next 17 Generation Joint Progress/Release Board 18 Meeting", and you are listed there as chair of 19 this meeting. Can you tell us what was this 20 board? What was its purpose? 21 A. The purpose of this board, it was a weekly 22 meeting to discuss the progress of the 23 development, you know, and obviously development 24 leading into testing, leading into deployment, 25 on the new Horizon next generation system. So 151 1 it's a weekly meeting we held jointly with 2 Fujitsu to discuss anything to do with ongoing 3 development activities. 4 Q. Can we look at page 3, please. Thank you, it's 5 the first shaded entries. We have there "AD", 6 that's Alan D'Alvarez. Do you remember Alan 7 D'Alvarez? 8 A. I remember Alan, yes. 9 Q. He worked at Fujitsu; is that right? 10 A. He did, yes. 11 Q. He's listed there as programme director, HNG-X, 12 Fujitsu. I'm going to read that third column. 13 It says: 14 "The delay in the commencement of volume 15 testing means that we will not be able to 16 perform a significant amount of testing before 17 commencing the Medium Volume Pilot. Hence we 18 will need a significant amount of data to be 19 collected from the live branches and Data 20 Centre. The data will also require careful and 21 thorough analysis. 22 "AD to confirm how this will be achieved." 23 Do you recall there being less testing than 24 originally planned before going to pilot? 25 A. No, there wasn't less testing. We may have 152 1 varied how we did the testing but there wasn't 2 less testing. One thing that was very, very 3 clear on Horizon Next Generation was that 4 quality was of the utmost importance. With any 5 project that is delivered, you generally have 6 three criteria: you have the time it takes to 7 deliver it; the cost to deliver it; and the 8 quality criteria that it must meet. I can 9 honestly say, as programme manager, that quality 10 criteria was the utmost of those three. 11 Inevitably there's always a little focus on 12 all three and the amount of forecast varies from 13 project to project, but quality was the most 14 important. And, you know, I know several times 15 throughout the development and testing of the 16 solution, we paused, we took a step back, and 17 because of testing, because of -- we couldn't 18 get the right data or whatever, to absolutely 19 make 100 per cent sure we'd got the right answer 20 before proceeding. 21 Q. Thank you. That's what you've said in your 22 statement but if we look at this particular 23 entry: 24 "The delay in the commencement of Volume 25 testing means that we will not be able to 153 1 perform a significant amount of testing before 2 commencing the Medium Volume Pilot." 3 What do you understand that to mean? 4 A. I took that to mean that we've got a different 5 way of doing it because we've actually got -- if 6 you're looking on the right, Dave John's was 7 leading on it and LF is Lee Farman. Lee Farman 8 was my test lead at the time and, you know, 9 I would have taken, you know, his advice on 10 this, that if he said, you know, he was 11 comfortable that the slight change to the 12 approach on testing gave him sufficient 13 confidence, then that was acceptable, and that's 14 why it would have been closed. Because you'll 15 see data has been pass to Lee Farman, Lee Farman 16 believed this is adequate for now. 17 Q. Tell me if I'm wrong on this but it doesn't seem 18 to suggest that there's a different type of 19 testing. What it seems to suggest is that there 20 won't be a significant amount of testing, and 21 instead there's going to be a collection of 22 data. Am I wrong on that? 23 A. It depends how you want to read it. Lee has 24 looked at the data that has been collected from 25 the existing sites, you know, from looking at 154 1 this detail here, and has satisfied himself that 2 what we needed to be assured, through the 3 testing, has actually been done. Now, whether 4 that's through a test or through the data that's 5 effectively been through the system, it's given 6 us the same result, is what that's telling me in 7 the fourth column. 8 Q. Forgive me for -- I don't quite understand what 9 you're saying there because, I mean, certainly 10 a plain reading of that is there won't be 11 a significant amount of testing and instead 12 there's going to be provision of data. Am 13 I wrong to think that there is a difference 14 between getting hold of data and actual testing 15 or are you saying they're the same thing? 16 A. They're clearly different things but the data 17 can give you the proof that what you'd have done 18 in the testing is actually sufficient. So the 19 data can give you the same proof, is what I'm 20 saying. 21 Q. So is data, collection of data, an adequate and 22 an equivalent to a significant amount of 23 testing, or is it less than a significant amount 24 of testing? 25 A. In this particular scenario it was clearly 155 1 considered to be an equivalent to the test. 2 Q. What gives you that impression? 3 A. Because Lee Farman who was a robust external 4 tester, so he wasn't Post Office, he wasn't 5 Pathway, he was actually bought in from 6 external, gave me that confidence and would have 7 explained it to me at the time. 8 Q. It says there "adequate", it doesn't say 9 "equivalent". Adequate seems to be less than 10 significant testing, doesn't it? 11 A. Well, I think the testing at any point it's 12 adequate for now. You'd want to do more testing 13 before you went into a high volume pilot is what 14 that says. Adequate to get into the medium 15 volume pilot. That's what it's saying. It 16 based on my interpretation of the word 17 "adequate". The way I read that is it's 18 adequate for now. In other words, Lee is happy 19 that it gives us what we need for the medium 20 volume pilot. 21 Q. If somebody were to say that you were, at this 22 stage, desperate to get on with the medium 23 volume pilot; would that be right? 24 A. That would definitely be wrong because nobody 25 ever put me under pressure to get on with 156 1 anything. We were always wanting to progress to 2 timescales but we wanted to progress to 3 timescales in the right way. 4 Q. Was the impression that you got from Post 5 Office -- more broadly rather than just 6 yourself -- one that you were desperate, as 7 a company, as an organisation, to get on with 8 the medium volume pilot? 9 A. That's what I -- I totally refute that 10 suggestion. That is 100 per cent factually not 11 correct. At times, you know, there were times 12 definitely when, you know, people would say to 13 me, you know, "Why are we not ready?" And 14 I would explain why we're not ready. And at no 15 point under my time as HNG programme manager was 16 I under pressure, or as delivery manager, was 17 I under pressure to hit a timescale at the 18 expense of quality. I would be expected to 19 explain it in certain situations, but never 20 under pressure. 21 Q. In your statement, you say at paragraph 18 "With 22 any computer system there's always likely to be 23 bugs and this is the purpose of testing." 24 Less testing might mean more bugs, mightn't 25 it? 157 1 A. In theory, yes, but not necessarily in practice. 2 With testing what you do, you agree an approach 3 to testing of a strategy and an approach for 4 specific phase of testing, if you like, that 5 gives you sufficient confidence. Now, with any 6 system I would suggest it's almost impossible to 7 be 100 per cent guaranteed that no further bugs 8 remain, but you do sufficient testing, depending 9 what it is you are developing, rolling out, 10 deploying, that gives you the confidence it will 11 work satisfactorily against what it's intended 12 to do. 13 So, obviously, something like, you know, 14 I don't know, a space rocket, you would do much 15 more testing on. So you'd vary the testing 16 according to what it is you're developing. It's 17 a standard approach called a risk based approach 18 to testing but you cover off the things that are 19 absolutely going to cause the -- any risk. 20 Q. If you intended to carry out significant 21 testing, and that was replaced with a lesser 22 form of testing or a different or a collection 23 of data, might that mean that there may be more 24 bugs? 25 A. No, because in that scenario, as I've said, Lee 158 1 has affirmed that what he saw in the data gave 2 us the same result as he would have expected 3 from testing to move into that medium volume 4 pilot. That was the key of that bit: it was to 5 move into that medium volume of pilot and it was 6 to look at the performance of the system and 7 performance in that sense would be is it holding 8 up speed-wise to be able to process 9 transactions? You know. It's not functional. 10 Q. Can we just go back to FUJ00092754. Page 3. 11 Does Lee there say that it gives the same 12 result? 13 A. Is that a question for me? 14 Q. Yes. Your evidence was that Lee has said to you 15 that the collection of data, rather than 16 significant testing, gives the same result. But 17 I was just wondering if that's your recollection 18 or if that's something that's stated there? 19 A. Well, yeah, in my opinion. I mean obviously 20 this is written by a PMO who takes the notes and 21 their choice of wording. Now, can I edit the 22 choice of wording? Of course I can. As program 23 manager, anybody can challenge the wording in 24 hindsight but my absolute recollection is Lee is 25 saying this is adequate for now, for what we 159 1 would have tested, to get into the medium volume 2 pilot. That's what I am recalling on this one. 3 We wouldn't have gone into that medium volume 4 pilot without a sufficient confidence. 5 Q. Can we look at POL00000874, please. This is the 6 Fujitsu's own testing and integration strategy. 7 If we have a look at this document it's 8 a Fujitsu document but if we look at the bottom 9 it has external distribution and it's something 10 that has been sent to the Post Office library 11 and Andrew Thompson -- do you know Andrew 12 Thompson? Do you recall who he was? 13 A. Andrew was one of the test leads with Lee. 14 Q. I'm going to -- 15 A. I dealt with Andrew primarily -- if I recall 16 correctly, Andrew primarily looked after 17 functional and Lee looked after non-functional 18 performance. 19 Q. Can we look at page 12 of this strategy and it's 20 paragraph 1.25 on page 12 that I'd like to look 21 at. 22 This is Fujitsu's testing strategy and it 23 says: 24 "As a general philosophy, it is important to 25 accept that no system can ever be confirmed as 160 1 completely error-free. It is not possible to 2 prove it. Tests can prove that an error exists. 3 They can prove that a previous error has been 4 corrected. They cannot prove that no further 5 errors remain. However, by concentrating on the 6 important characteristics of the system 7 operation, tests can be used to demonstrate the 8 progressive removal of errors to the point where 9 these characteristics are seen to conform to 10 expectation. So, testing can be seen as 11 a method (the primary method) of reducing the 12 risk of serious defects remaining in important 13 areas of the system." 14 Is that something you agree with? 15 A. Yeah, that's the risk-based approach to testing 16 that I referenced a couple of minutes ago. 17 Q. Yes. So it says testing there is either the 18 primary method of reducing the risk of serious 19 defect, it doesn't say there that data, 20 gathering data is an adequate replacement for 21 tested, does it? 22 A. Well, it doesn't say it, no. 23 Q. If Fujitsu had the impression, rightly or 24 wrongly, that POL were desperate to get on with 25 the pilot and, as a consequence, reduced the 161 1 testing, what kinds of problems would you 2 envisage? 3 A. Well, Fujitsu couldn't reduce our testing. They 4 could reduce their testing but not our testing. 5 Q. If testing, any testing, was reduced, because of 6 an impression, rightly or wrongly, that was 7 given, that you needed to get on with the pilot, 8 what was the risk in that? 9 A. Well, but it wasn't. Testing wasn't reduced 10 because we were under pressure to get on with 11 the pilot. 12 Q. I'll ask once more. If it was, what would the 13 risk be? Reading that, what is the risk of 14 reducing testing? 15 A. Well, the risk is you're risking more bugs going 16 forward. 17 Q. Thank you. Can we look at FUJ00097159, please. 18 We're still on 28 January and this seems to be 19 a different board meeting. This is called the 20 "Horizon Next Generation Release authorisation 21 AG3 -- Joint Board". How is this board 22 different? Do you remember this board? 23 A. This will have been the board -- I vaguely 24 remember it -- this will be the board that 25 looked at, you know -- very similar to the 162 1 acceptance on original Horizon. This is the 2 board that looks after are we ready to authorise 3 the release of Horizon into the branches. 4 Q. Can you tell us what AG, Acceptance Gateway, 3 5 was? 6 A. I can't remember what AG3 was. I'm assuming it 7 was, looking at the date, probably entry into 8 pilot but I can't be 100 per cent certain. 9 Q. Can we turn over the page and it has there "New 10 Actions and points to note from combined meeting 11 of [21 January 2010] -- ACCEPTANCE". It's that 12 first shaded section I'd like to look at first. 13 It says there: 14 "DC [that's, I think, Dave Cooke the Fujitsu 15 acceptance manager] confirmed that there were no 16 outstanding High Severity Acceptance Incidents 17 and that all other thresholds are within 18 tolerance for Acceptance Gateway 3." 19 I just want to look at the third paragraph. 20 It says: 21 "However it was agreed that the high 22 priority fixes in 'Reset 4' -- to be delivered 23 as part of Maintenance Release 01.08 -- could 24 constitute a High Severity AI if not delivered 25 in time for the High Volume Pilot (over 272 163 1 branches)." 2 Then if we scroll down to the next entry, it 3 seems clear from that that there weren't any 4 outstanding high severity Acceptance Incidents, 5 but, as we've heard, there's the potential for 6 there to be one -- 7 A. Yes, it strikes me from that that there was high 8 severity Acceptance Incidents that had been 9 fixed but the fix hadn't actually been deployed 10 yet, reading those words and therefore was 11 subject to fully being signed off, if you like. 12 Q. Then we have look at that final entry. It says, 13 "Non-High priority items in Reset 04": 14 "MB [I think that's yourself] offered 15 an option to remove items from Reset 04 which 16 are not regarded as High priority -- if they are 17 risk of missing the High Volume Pilot deadline 18 or affecting the delivery of items which are 19 High priority. AD will consider if any items 20 are at risk [that's Alan D'Alvarez], and liaise 21 with Post Office ..." 22 Is this something you remember at all? 23 A. I don't recall it, not as such. 24 Q. It seems as though you're offering there to 25 remove some items from what's called Reset 4 164 1 which weren't regard as high priority. 2 A. Yeah, and that's quite possible, if there were 3 things on there that weren't high priority and 4 therefore weren't considered material risk, then 5 to proceed through this gateway -- and don't 6 forget I think it's important to recall that 7 when we're in a pilot you're getting more 8 evidence from the pilot of any potential 9 problems. So, you know, we would have, you 10 know, looked at anything that wasn't classed as 11 high priority and said "Is that material to 12 proceeding to the next phase?" It's not saying 13 we're actually going to remove them; it's saying 14 we'll look at removing them. 15 Q. Was that because the priority at that stage was 16 to get the high volume pilot going and get on 17 with acceptance, given previous historic delays? 18 A. No. The priority was to deliver a quality 19 system, but with any system you would look at 20 the lower priority items and say "Is this 21 important for moving forward at this stage?" 22 Yes, no. And you'll see there that it is being 23 progressed by the POL and Fujitsu testing and 24 requirement teams. So we wouldn't have put 25 anything, given exception to anything that we 165 1 didn't get approval from. 2 Q. Absolutely, but were you making it easier there 3 for Fujitsu so that they could meet the high 4 volume pilot deadline by removing non-essential 5 items? 6 A. Not easier, no. It's a standard approach to 7 saying we need to look at the items on the list 8 that are outstanding. 9 Q. So again, at this stage, still at 28 January, 10 would it be unfair, I think you've said yes 11 before, to say that you were desperate to get on 12 with the pilots? 13 A. Would it be unfair? Yes, it would be unfair. 14 I think the word you're using there is 15 "desperate to get on with". Keen to get on 16 with, I'd say, is fair. But "desperate", 17 definitely not. 18 Q. Can we look at page 9 of this document. It's 19 the bottom of page 9 I'd like to look at. "New 20 Actions and points to note from the combined 21 meeting of [28 January]". 22 I'm going to read to you the first three 23 entries. So the first one: 24 "Branch Trading Statements at Coton House 25 and Warwick. It was confirmed that onscreen 166 1 error messages had occurred at both offices 2 which meant that their Branch Trading Statements 3 could not be completed. 4 "The root cause position at Coton House was 5 understood, but the Warwick office had 6 additional complexities. 7 "It was confirmed however that in both cases 8 the underlying data integrity and external feeds 9 to POL-MI and POL-MS were not affected." 10 Next entry. 11 "Double settlement at Derby. In addition to 12 the above, there have been 2 instances of 13 a transaction appearing to be settled twice. 14 This was picked up by the standard 15 reconciliation process and corrected via BAU 16 processes. However, the root cause is yet to be 17 identified and investigation is under way". 18 The third entry "Decision -- Postponement of 19 the next 10 branches. Based on the lack of 20 a known root cause for AG3.70 [so AG3.70 is the 21 branch trading statements issue] and AG3.71 22 [that's the double settlement issue], it was 23 agreed that the next 10 branches should be 24 postponed until the impact and way forward is 25 fully understood." 167 1 Do you recall those issues at this time? 2 A. No, I recall the Warwick and Coton issue. 3 I vaguely recall -- I don't recall the Derby 4 specifically but I recall a double settlement, 5 yes, and I think the decision in 372 absolutely 6 reaffirms, you know, this was not a time 7 pressure; this was, you know, getting it right. 8 And I can explain to you the 371 and 370, if you 9 like, if it helps. 10 Q. We can get on to them but your evidence is that, 11 28 January, the postponement of the next ten 12 branches shows that you weren't in a rush to get 13 on with things, is that -- 14 A. It shows we weren't pressured into getting on 15 with things, yes. 16 Q. Can we look at FUJ00092875. It's page 3 of that 17 document. Thank you. This is an email from 18 Alan D'Alvarez of Fujitsu, of 3 February. So 19 very shortly after that, just a few days later. 20 This is, I believe it's an internal Fujitsu 21 email but do correct me information I'm wrong on 22 that. Do you recognise the names there? You're 23 not copied into that? 24 A. I think they're all Fujitsu resources, yes. 25 Q. Thank you. Let's look at point 1. It says: 168 1 "The decision has been taken to deploy HNG-X 2 to a further 10 branches with the migration 3 button being pressed tomorrow for migration to 4 complete Friday." 5 So the decision to postpone due to, for 6 example, the branch trading statement issue that 7 we just saw and the double settlement issue that 8 we just saw seems just a few days later to have 9 been reversed. Do you remember that? 10 A. Yes, I do remember that. 11 Q. Then we look at point 2: 12 "There are two issues that require fixing 13 prior to being able to enter into a medium 14 volume pilot: 15 "Branch Trading Statement: This is where the 16 in day migration process that happens once 17 a branch hits the migration button is not 18 correctly migrating across the summary data. 19 This data is used to produce the branch trading 20 statement. The base data is being migrated 21 correctly and therefore the central accounting 22 is unaffected. However, should a post office 23 manager do a trading statement in branch there 24 is a high chance that the statement produced, 25 which uses the summary data, will not reconcile. 169 1 We are manually 'fixing' the summary data prior 2 to the BTS being run for those branches already 3 migrated." 4 Just pausing there, what do you understand 5 by "manually fixing" the summary data. 6 A. So basically when the branches were migrated on 7 to HNG, it was an automated system that migrated 8 the data from their existing solution on to 9 Horizon Next Generation, and this was an error 10 that occurred in certain circumstances. So what 11 happened, we identified the root cause of the 12 error, and what happened was Fujitsu knew where 13 it was, but we decided to continue with the ten 14 because this migration process -- obviously it 15 only happens when you migrate a branch. So, you 16 know, the best test for this is when you migrate 17 a test but you just need to make sure you've 18 absolutely got a control total, which you can 19 then manually fix should any other data not 20 migrate successfully. That's what that manually 21 fixing meant. 22 Q. How would it be manually fixed? 23 A. Well, by the migration team comparing what 24 should be there with what's actually there and 25 then making the appropriate entry onto the 170 1 Horizon System so that can be remigrated onto 2 Horizon Next Generation. 3 Q. So would the entry that's created by Fujitsu 4 technicians show up on the branch trading 5 statement as a different figure? 6 A. I can't actually remember how it worked in terms 7 of the correction, I'm sorry. 8 Q. Turning to the second one: 9 "Counter Pauses In Live: Yesterday and today 10 a number of branches reported 'screen freezes' 11 while operating HNG-X." 12 So you have two uses there that require 13 fixing before entering into the medium volume 14 pilot. Then paragraph 3 it says: 15 "Currently the team are still investigating 16 both issues above and, as yet, have not 17 determined the root cause for either." 18 Do you remember that at that stage? 19 A. I remember that, yes. 20 Q. Then it's paragraph 4 that I want to spend a bit 21 of time on, and it says: 22 "We had a meeting with Post Office this 23 evening which Mark Burley led from the Post 24 Office side. Post Office are desperate for 25 a date to start planning/rescheduling medium 171 1 volume pilot. They accepted our position that 2 we were not able to give this today. I expect 3 that Mark will be keeping Dave Smith briefed and 4 my reading is that if we are not in a position 5 to give a target date by close of play tomorrow 6 it is likely to result in an escalation to Mike 7 Young." 8 So that's where we get the phrase 9 "desperate" from. Was Alan D'Alvarez wrong when 10 he said you were "desperate"? 11 A. I don't know why he used that word, so I think 12 that's a question for Alan. 13 Q. Was there pressure being put on Fujitsu by the 14 Post Office and by yourself to get on with 15 things? 16 A. Not to the extent that it would be desperate. 17 Clearly, for the migration, we have to plan and 18 schedule the migrations. You know. So we have 19 to give the officers that are due to migrate due 20 notice because, you know, we would do that out 21 of hours, typically. They may need resources in 22 place. They have to, you know, prepare 23 themselves for that migration. So, you know, in 24 terms of, you know, we would want a date that we 25 are targeting. Now whether that date was 172 1 tomorrow or next week or next month, we want 2 a date because we want to start the planning 3 process, not desperate to get it going, 4 necessarily. 5 I think it's -- the -- not a term I would 6 have used. You know, we need a date, such that 7 we can continue with our planning rescheduling 8 exercise. 9 Q. Can we keep that on screen but just put 10 alongside it your witness statement, 11 WITN03850100. It's page 4 of your witness 12 statement, paragraph 16. 13 Thank you. Paragraph 16. You say there: 14 "As Programme Manager, the one person who 15 would have been under pressure to migrate to 16 HNG-X quickly would have been myself. Through 17 demonstrating how we, as a team, were holding 18 Fujitsu to account on quality, there was never 19 any pressure to migrate before we had signed off 20 readiness." 21 Do you see the contrast there between what 22 you've said in your witness statement and what 23 Mr D'Alvarez was saying at the time? 24 A. No, I don't. I'm sorry, but you're reading that 25 different to how I'm reading it. "Post Office 173 1 are desperate for a date to start 2 planning/rescheduling". It doesn't say we're 3 desperate for a date that has to be imminent. 4 At no point -- I repeat, at no point was I put 5 under pressure from a timescale perspective. 6 I think it's just how you've chosen to read the 7 two statements are different. 8 Q. By that stage, how long had HNG-X been planning 9 for? 10 A. It depends what you mean by "planning". It had 11 been running, if this was February, the best 12 part of three and a half years. 13 Q. Was one of the reasons that you were brought in 14 to the project to drive things forward? 15 A. Sorry, was it what, sorry? 16 Q. Was one of the reasons you were brought into the 17 project to drive things forward? 18 A. In the sense that a programme manager drives 19 anything forward, yes. 20 Q. Yes. Do you not think that Fujitsu might have 21 felt some time pressure by this stage? 22 A. I think that they felt time pressure for 23 themselves, yes. 24 Q. But your evidence is that that pressure wasn't 25 coming from the Post Office? 174 1 A. As I said, what is said and how things are 2 interpreted can be two different things. Would 3 I have said, "Alan, come on, I need a date, we 4 need to start, you know, and get these 5 replanned, rescheduled", yes, I probably would 6 say something like that. But, you know, there's 7 a difference between saying that and saying, 8 "Come on, Alan, I need to get these started 9 again tomorrow, we're already behind plan". 10 That is definitely not something I would have 11 said. 12 And you'll notice there they accepted our 13 position that we're not able to give this today 14 and, yes, I would want to keep them, you know, 15 under a little bit attention that we're wanting 16 a date because, as I said, scheduling and 17 planning the branches, it wasn't a dead 18 straightforward task. It takes time. You have 19 to accommodate, you know, things that the office 20 might have planned, you know, and, you know, we 21 have to share appropriate evidence with other 22 parties who were, you know, who, you know, for 23 example with the National Federation of 24 SubPostmasters, who I brought in to also give 25 assurance before we deployed. 175 1 MR BLAKE: Thank you, sir. That might be 2 an appropriate moment to take our 15-minute 3 break. 4 SIR WYN WILLIAMS: All right, 15 minutes. So 5 3.20-ish, I suppose to be imprecise. 6 MR BLAKE: Thank you. 7 SIR WYN WILLIAMS: Have a short break, Mr Burley, 8 all right. We have a break in the afternoon and 9 the morning because the person transcribing the 10 evidence needs a break, all right? 11 A. Okay, do I dial off? 12 SIR WYN WILLIAMS: No, mute yourself and go off 13 screen and go for a walk but don't break the 14 connection, all right? 15 THE WITNESS: Okay. 16 (3.10 pm) 17 (A short break) 18 (3.22 pm) 19 MR BLAKE: Thank you very much, Mr Burley. We're 20 back. I'm going to return to the issue of the 21 BTS, the branch trading statement and can we 22 look at FUJ00094268, please. We're still in 23 February 2010. This is an email chain. It's 24 an email chain that's not ideal from 25 a formatting perspective, so it may take us 176 1 a little bit of time to go through. This is 2 page 10 and you'll see there is an email from 3 Geoff Butts, and if you look down it is to Will 4 Russell, yourself and Barry Evans. Who are the 5 other two? 6 A. Will and Barry led the migration activity. 7 Q. Thank you. If we scroll down and over the page, 8 there's the email from Geoff Butts. He says: 9 "Both, 10 This is to confirm the position on a fix for 11 the BTS [the branch trading statement] issue 12 whereby printing the Trial Balance Report for 13 BTS results in incorrect data being displayed in 14 the Final Balance Report. A fix can be 15 delivered and tested for inclusion within the 16 01.08 Maintenance Release as a counter fix. In 17 the meantime, branches need to use the BTS Trial 18 Balance Report, which is correct, and discard 19 the BTS Final Balance Report, which is 20 incorrect." 21 Do you remember this issue? 22 A. Vaguely, yes. I think it's linked to the one we 23 talked about before. 24 Q. So we saw the one related to Coton House and 25 Warwick. 177 1 A. No, this is the one relating to the migration 2 and the incorrect, when they migrate, this will 3 be the -- because they used to run a BTS, branch 4 trading statement, when they'd migrated the 5 data, to prove that it had migrated correctly. 6 Q. Here the suggestion is that a work around is 7 possible until it can be fixed. 8 A. Yeah. 9 Q. You respond, page 9 is your response. You say 10 there: 11 "Geoff 12 "Thanks but can I ask who you have agreed 13 this workaround with? This is a legal document 14 and there is a difference between a trial 15 balance and a Final Balance? I would also 16 expect CS ..." 17 Is that Customer Service or? 18 A. I can't -- I'm guessing so. 19 Q. "... to have a KEL [Known Error Log] for this -- 20 if agreed -- to be able to explain the position 21 to any subpostmaster who calls in. 22 "Graham to confirm please." 23 Do you remember this response? 24 A. Well, vaguely, yes. I mean, it is a long time 25 ago but vaguely I remember it, yeah. It's back 178 1 to the cash account. It's basically, you know, 2 it's their accounts. 3 Q. Can you tell us what you're saying there? 4 A. Well, basically, I'm saying that why do you rely 5 on the trial balance and not the final balance 6 because the final balance, ultimately, is the 7 one that's important. Now, I don't know what 8 his response was on this but it might be it was 9 just on the printout that it was causing the 10 problem and not on the actual system stored 11 data, you know, but you'd still expect there to 12 be a Known Error Log, even if it is just on the 13 printout. 14 Q. That's so that the message can be communicated 15 to subpostmasters that there is a problem with 16 quite a significant document that, as you've 17 said there, is a legal document or used in legal 18 proceedings? 19 A. Well, I'm -- 100 per cent. And don't forget, 20 I think it's worth pointing out at this stage 21 that we had a subpostmaster group as part of our 22 release authorisation process. You know, 23 I bought in the part of the NFSP to support and 24 be comfortable with what we were doing. So they 25 would also -- you know, I'd expect them to sign 179 1 it off as well. 2 Q. So are you saying they would have known about 3 this specific issue? 4 A. Yes, because, you know, absolutely. I brought 5 the NFSP in because I wanted them to be 6 comfortable with what we were doing and what we 7 were deploying. 8 Q. How would you share this kind of information? 9 A. Well, we had tripartite meetings with them, and, 10 you know, I would just share it open and 11 honestly because, you know, they need to be 12 clear, you know. What I'd typically do with 13 something like this is ask for some sort of 14 demonstration of what's happening and for them 15 to sign up to that it's either okay or, 16 actually, if I didn't think it was okay, I might 17 not even share it because there's no point 18 sharing something that I've said no -- you know, 19 if we agree it's not fit for purpose with 20 Fujitsu and they have to put a full fix in with 21 the final balance, and that's why I'm asking the 22 question I'm asking, you know, who they've 23 agreed it with. 24 Q. So you do recall or don't recall this specific 25 issue being mentioned to the NFSP? 180 1 A. I don't recall whether this specific issue was 2 mentioned to the NFSP or not. 3 Q. Can we look at page 7. It's been escalated to 4 Alan D'Alvarez within Fujitsu or sent to Alan 5 D'Alvarez. It says there -- sorry, it's the 6 page before. Thank you. 7 It says: 8 "Alan 9 "Just rang Geoff but he is on leave -- would 10 you be able to help with this one please. 11 "Would you please send any history regards 12 this issue -- specifically what the differences 13 in the Final report are, and can you advise how 14 Fujitsu have established that the trial report 15 is correct -- and that there is no corruption of 16 the data. 17 "A speedy response would be appreciated as 18 I need to explain this issue to our Finance 19 colleagues in more detail." 20 That's from Phil Norton, who is at the Post 21 Office, that's a colleague of yours? 22 A. Yes, who used to work for me on Horizon Next 23 Generation, yes. 24 Q. So what's happening with these branch trading 25 statements is it's the final report that's wrong 181 1 but there's another report earlier on in the 2 consequences that is right and that they want to 3 place reliance on that earlier report rather 4 than the final report? 5 A. Yes. But what it doesn't tell us here is, if 6 that's the system report or just the printed 7 version. 8 Q. Yes. Can we look at page 4, please, which is 9 an email from Gareth Jenkins of 19 February to 10 Phil Norton, copied to Alan D'Alvarez. 11 As I said, they're laid out slightly 12 strangely, but if we go over the page, that's 13 the substance of that email. It says: 14 "Alan D'Alvarez has asked me to respond to 15 your concerns below. I'll try to explain the 16 issue and what has caused it. 17 "I assume you've seen the attached write-up 18 of the issue which was sent to Barry Evans ... 19 "To expand on this a bit further: 20 "When the BTS is being produced, it is done 21 based on data written to the Branch Database ... 22 whenever a Stock Unit Rolls Over. 23 "BTS Production retrieves this data from 24 BRDB to the counter. 25 "The counter then uses it to generate the 182 1 Trial BTS. 2 "There is a bug in the way that the report 3 is produced such that some of the in-memory copy 4 of the data is overwritten when the Trial BTS is 5 produced. 6 "A consequence of this is that when the 7 final BTS is produced it is using incorrect 8 data. 9 "The problem has been fixed by ensuring that 10 the final BTS uses the original data retrieved 11 by the counter at step 2 rather than the data 12 left over after step 4. 13 "The problem was first reported by Warwick 14 ..." 15 So that's the issue we saw earlier that 16 originally paused the rollout to further 17 branches but that was quickly resumed: 18 "... and they point out that the trial 19 Balance figures were correct and the Final 20 Balance figures were incorrect. 21 "comparing the Trial and Final balance 22 figures with the corresponding Stock Unit 23 Balance Reports shows easily that the Trial 24 Balance is correct and the final Balance is 25 incorrect. 183 1 "None of the incorrect data is stored to 2 BRDB so there is no possibility of ongoing 3 corruption of stock levels. 4 "Does this clarify the situation for you?" 5 So what do you understand the position to be 6 in simple terms? 7 A. I understand it to be some sort of corruption 8 between the trial balance and final balance that 9 isn't actually corrupting the data but it is 10 corrupting the way the data is being reported, 11 such that the final balance hasn't got the 12 correct data on it any more. But it's not 13 actually impacting the balance as carried 14 forward and, if that's the case and it's been 15 validated as the case and that it doesn't 16 invalidate the data going forward, then I guess 17 that may have been acceptable. But, as I say, 18 I don't remember the detail of the, you know, 19 I'm assuming that -- well, that would have been 20 tested and proven to be correct before we'd have 21 made the decision to accept this. 22 Phil would also have checked with his 23 Finance colleagues in what was known as P&BA, 24 pension and branch accounting, I think, 25 something and branch accounting. To validate 184 1 that they were comfortable there was no risk of 2 corruption. 3 Q. We'll see what happens. We can look at 4 an email, another response from Gareth Jenkins. 5 That's page 2. Just pausing therefore briefly, 6 what was your relationship like with Gareth 7 Jenkins? 8 A. I didn't have a real relationship with Gareth, 9 to be honest. 10 Q. What did you know or understand his role to be? 11 A. I couldn't recall what his role was at this 12 time. I am guessing he was some sort to analyst 13 within the Fujitsu team. That's why he's 14 responding to Phil. 15 Q. You raised earlier in the email chain the issue 16 of this being a legal document. Did you have 17 any discussions at any time with him about legal 18 proceedings? 19 A. No. 20 Q. Were you aware of his involvement in legal 21 proceedings at all relating to the Horizon 22 System? 23 A. No. 24 Q. So this is the response from him. Now, this one 25 is slightly difficult to understand and it may 185 1 need a bit of explaining from me. The indented 2 parts of this are Phil Norton's email that 3 Gareth Jenkins is responding to. So where he 4 says, "Phil" and then it goes indented, actually 5 the bit that is indented seems to be Phil 6 Norton's comments and Gareth Jenkins is 7 commenting below. 8 So Gareth Jenkins's words seem to be -- and 9 if we look at the penultimate paragraph and 10 final paragraph on that page, the bottom of that 11 page, if we could highlight those entries, thank 12 you -- he says: 13 "It is not practical to explain exactly 14 which figures are wrong and which are correct on 15 the Final Report. However it is restricted 16 entirely to the Stock levels of Volume Stock 17 (ie the second part of the BTS). From the 18 examples we have seen it is likely to be many of 19 the Stock levels are wrong) and the figures will 20 be too high, so indicating that the Branch's 21 Stock levels should be higher than they really 22 ought to be)." 23 Do you remember being satisfied by that kind 24 of an explanation, that it's not practical to 25 explain exactly which figures are wrong? 186 1 A. Was I put onto the distribution list for this 2 one? 3 Q. Well, you do appear in the chain later on, so 4 you certainly did receive this within the chain. 5 A. I mean, my view is that this would be looked at 6 by Phil and by Will's team, so Will or Barry, to 7 ascertain exactly what's going on and whether 8 that's acceptable and they would determine its 9 acceptability or not by consulting others, as 10 I said, the head of P&BA, for example, to make 11 sure there is no financial risk. 12 Q. Can we just zoom out. You are in fact on this 13 particular email. You're there, you're one of 14 the copy recipients. 15 A. Yeah. 16 Q. So at the bottom there again, we move to Phil 17 Norton's email, which says: 18 "Given the legal status of these reports 19 ..." 20 This is the issue that you had previously 21 raised -- 22 A. And I ask (unclear) on my original question. 23 Q. Exactly: 24 "... (they -- the final -- are often used in 25 court proceedings when we are trying to recover 187 1 monies from dismissed subpostmasters) and the 2 potential 'integrity' challenges that could be 3 levelled against the HNG system as a result of 4 any differences; POL need assurance from Fujitsu 5 that they could/will explain the cause of the 6 problem and prove the system has integrity 7 should we be challenged." 8 Gareth Jenkins' response on that is: 9 "I understand this, though I would have 10 [I think that must have been 'thought'] most 11 legal proceedings would be based on the first 12 part of the report (which covers cash levels) 13 rather than the second part." 14 Then we have, again, Phil Norton's email: 15 "I know this is not your area, however in 16 order to help us get concurrence from our 17 Finance Team we do need the commitment. I know 18 Mark has asked Chris for that assurance ..." 19 Is that "Mark" yourself? 20 A. Probably. I don't know for sure. 21 Q. "... however dependent upon the outcome of the 22 meeting schedule with POL Finance on Monday we 23 may need to request more specific assurances." 24 Gareth Jenkins' response is: 25 "I'll leave others to respond to that." 188 1 There's another point that he's responding 2 to, from Phil Norton, which said: 3 "Would you please advise who within FS 4 [that's Fujitsu] would be in a position to 5 provide the assurances/agree to any such 6 request." 7 Gareth Jenkins again says that he would 8 leave that to others. 9 Can we look at page 1 of that email chain, 10 please. This is followed up by an email from 11 Phil Norton to Alan D'Alvarez -- thank you very 12 much. 13 You're copied in to that, and he says: 14 "Alan 15 "Further to the correspondence detailed 16 below -- I have now met with our finance team to 17 discuss this issue and to establish their 18 position regards any requirements that they have 19 to ensure POL does not incur additional expose 20 [I think that must mean 'exposure'] as a result 21 of this defect. 22 "As previously stated the Final Balance 23 Report is a legal document and as such POL rely 24 on the accuracy and integrity of the report 25 during any legal action we undertake to recover 189 1 monies from the Subpostmasters. Given the Trial 2 report can be planted several times -- usually 3 following stock adjustments made during the 4 balancing process, this has no standing from 5 a legal perspective. 6 "Based on this, and the understanding of the 7 defect, how it arose, and the proposed 8 rectification (as provided by Gareth); our 9 Finance team require certain specific 10 deliverables from Fujitsu in order for them to 11 be able to manage the risk this poses and to 12 give them the confidence to support the 13 continued rollout of HNG. 14 "The deliverables required [include]" 15 They want a list of all products where the 16 volumes on the final balance differ and they 17 want a definitive statement explaining how it 18 has happened, et cetera. 19 They also want a commitment to support POL 20 improving the integrity of the system, that's 21 number 5. Over the page: 22 "POL Legal to 'sign off' the Fujitsu 23 Definitive Statement as fit for purpose. 24 "There are several internal processes that 25 will be implemented to ensure any impact is 190 1 minimised, however I do need Fujitsu's agreement 2 to the above to gain the support of our Finance 3 teams." 4 Was the fact that the final balance that was 5 produced by Horizon Online at this stage, that 6 it was inaccurate, was that a serious issue for 7 the Post Office? 8 A. Absolutely. I think, as you can see, you know, 9 from the investigation that was done of it. 10 Q. The workaround that had been suggested that the 11 final balance was simply discarded, was that 12 appropriate? Was that considered appropriate by 13 the Post Office? 14 A. Well, based on the above comments, no, and 15 that's why I asked with Phil to check with P&BA 16 because they're the owners of the financial 17 integrity. 18 Q. Do you think that Fujitsu were taking the legal 19 implications of this seriously enough? 20 A. I'd like to think so. I mean, you know, in my 21 discussions with them, I never had any reason to 22 doubt that, you know. So I'd like to think they 23 were taking that seriously. But without seeing 24 how they responded to this, you know, it would 25 be interesting to see what their response was. 191 1 But, certainly in my discussions with Alan, he 2 would always, you know, absolutely support what 3 we were trying to do to make sure this was 4 really of good quality. 5 Q. It's quite significant, of course, that 6 a document that might be used in legal 7 proceedings is inaccurate in showing the final 8 balance. How far up within the Post Office do 9 you think this issue was raised? 10 A. I honestly can't -- cannot tell you. It doesn't 11 have to be raised anywhere, unless we were put 12 under pressure from Fujitsu to accept their 13 workaround. You know, from everything I've read 14 so far, we did absolutely the right thing. 15 Asked Phil to look into it with P&BA, as I say, 16 who owned the financial integrity. They've 17 given us a very clear statement that they need 18 to understand and to be able to have absolutely 19 definitive proof that the integrity of the data 20 is not compromised in any way and also ensure 21 our legal team to sign off to that effect. 22 You know, so really important to us and 23 them, you know, that this is, you know, 24 absolutely integrity is at the heart of it. 25 Q. The workaround is not relying on the final 192 1 report and having some sort of side assurance 2 that everything is okay. Given that people were 3 prosecuted by the Post Office, do you think that 4 that was sufficient? 5 A. Well, sorry, I don't believe we went live with 6 this workaround. 7 Q. But do you think that attitude of reliance on 8 a side statement, rather than fixing the problem 9 then and there, was sufficient? 10 A. I think that's an unfair question, to be 11 brutally honest. I think Fujitsu were proposing 12 a potential workaround. At no point are they 13 saying this workaround has to be accepted. They 14 were proposing what they deemed a potential 15 workaround. Clearly, it set alarm bells in my 16 head when I saw it, hence I asked who'd agreed 17 to it, and asked Phil to look into it. 18 And, you know, I don't see anything on here 19 that's, you know -- there's a little bit where, 20 you know, he's trying to explain why he thinks 21 it's okay but, clearly, doesn't have the full 22 understanding of the significance of the final 23 balance statement in any potential legal 24 proceedings. 25 So I don't see any pressure from them, you 193 1 know. Clearly, in hindsight, it wasn't a good 2 proposal and, as such, I believe it was rejected 3 on this basis. 4 Q. Given the significance of Post Office's ability 5 to prosecute people, based on information 6 provided to a court in arising from a printout 7 from the Horizon System, do you think that 8 an approach that looks for workarounds is the 9 right approach or do you think, for example, 10 that should have caused the Post Office to take 11 a deep look at the system? 12 A. As I said to you, it set alarm bells in my head. 13 A deep look at the system, this is one issue 14 that arose that absolutely we were trying to do 15 the right thing to make sure it was resolved. 16 Q. How would this information be cascaded down to 17 individual branches? 18 A. Well, it wouldn't be because, as I said, unless 19 you're going to tell me we went live with it, 20 which I don't believe we did, and I'll be 21 surprised if we did, you know, why would we 22 cascade it, because we'd expect it to be fixed. 23 Q. Your suggestion, though, was to make sure it was 24 on the Known Error Log, so that any 25 subpostmasters who phoned in would -- the 194 1 message somehow would be linked? 2 A. No, no, my -- can you put my words back on the 3 screen, please? 4 Q. Yes, it's the same document, and it is page 9. 5 A. My words don't say that. My words say, first of 6 all, who has it been agreed with? This is 7 a legal document so instantly registering my 8 concern about it and there's a difference 9 between a trial and a final, you know, if we 10 were to go ahead with anything like this, you 11 know, we would expect -- if CS is customer 12 service, I don't know -- but certainly a Known 13 Error Log, if it was agreed. And, as I said, 14 I don't know -- my first question is: who has 15 agreed to it? And then, as you see, I asked 16 Phil to determine if it's acceptable and that's 17 what leads to all the investigation that says 18 it's not acceptable. 19 So you wouldn't have a Known Error Log for 20 something that we're not accepting because, 21 actually, we'd be asking Fujitsu to fix it. 22 Q. It's the attitude that I'd like to drill down 23 on, really, because we're talking about very 24 early days of Horizon Online here, and there is 25 a problem with the balance, the final balance, 195 1 and what you're saying there is that you would 2 expect there to be a Known Error Log, if it was 3 agreed, so that subpostmasters -- any 4 subpostmaster who calls in would -- the two 5 would be linked. But what about all those 6 subpostmasters who don't call in? 7 A. Sorry, you know, it's not about being able to 8 explain it; it's about being absolutely 100 per 9 cent certain it's got integrity and that's what 10 I said at the beginning. It's absolutely 11 critical that the system had integrity and we 12 could support that integrity. And, hence, 13 I asked Phil to look at that. And there's no 14 way we would have proceeded without the system 15 having integrity. 16 Q. But where there were problems in the system, 17 known problems, was the attitude to have a Known 18 Error Log for the problem to be able to explain 19 the position to any subpostmaster who calls in? 20 Was that a common response to problems that, put 21 it on the error log and all will be okay? 22 A. No, I think that is really not a fair reflection 23 at all. A Known Error Log is more for things 24 that are genuine workarounds that have no real 25 impact. It's almost back to my font example. 196 1 You know, it's things that do not impact the 2 subpostmaster, do not impact a counter clerk, do 3 not impact, you know, any of our clients at the 4 time. 5 They're just things that, you know, are, 6 shall we say not quite, you know, as we expect 7 them to be, but aren't causing any financial or 8 other material impact or questioning of 9 integrity, and we wouldn't proceed with it. We 10 wouldn't have something on the Known Error Log 11 for something that is an integrity question. 12 Q. So if the Known Error Log did have things that 13 were more significant than just the font issue, 14 that affected the real lives of subpostmasters, 15 would that have been poor practice and 16 unexpected by yourself? 17 A. I mean -- sorry, font example is a very, very 18 simplistic example. A Known Error Log will 19 contain certain scenarios, you know, and they're 20 generally what are classed as low-level issues 21 that will be fixed over time. But they 22 shouldn't have any impact on the integrity of 23 the system. If they have impact on the 24 integrity of the system, I will be surprised, if 25 they're on the Known Error Log. 197 1 Q. Can we look at POL00032999, please. This is the 2 same period, so we're looking here at 3 23 February 2010. This is the acceptance report 4 for HNG-X. Can we look at page 12 -- just to 5 look, the external distribution, you're on the 6 list there as having received this document. 7 Can we look at page 12 and it's 8 paragraph 1.3 that I'd like to look at. 9 It may be slightly earlier than that. It 10 may be the page before. Thank you very much. 11 "Acceptance Decision", could we highlight the 12 "Acceptance Decision"? I'm just going to read 13 that. It says: 14 "The AG3 Acceptance Board ... was held on 15 21/01/10 and approval to proceed through AG3 was 16 granted. The decision stated that '... it was 17 agreed that the high priority figures in 18 "Reset 4" -- to be delivered as part of the 19 Maintenance Release 01.08 -- could constitute 20 High Severity AI if not delivered in time for 21 High Volume Pilot ...' As such, 'Qualified 22 Acceptance' at AG3 was agreed, subject to the 23 successful delivery of the high priority items 24 in Reset 04. 25 "Whilst not an exact match, this the closest 198 1 to the Acceptance Board Recommendation 3: 2 'Proceed at risk through Acceptance Gateway'. 3 "These associated conditions were discussed 4 at the corresponding release authorisation board 5 and recorded in the minutes of that meeting." 6 Do you recall in terms of the acceptance of 7 Horizon Online that it was effectively "proceed 8 at risk", as in there still were some concerns 9 but the agreement was to proceed despite some 10 concerns? 11 A. But the concerns are not to do with anything 12 that is high severity and therefore not 13 an integrity question. It is clearly called out 14 there but yet a number could combine to form 15 a high severity, if whatever the fix was, didn't 16 actually complete successfully, and therefore we 17 would end up pausing again if that was the 18 situation. 19 Q. Can we look at page 9 of this document. It may 20 be at the bottom of page 8. The bottom of this 21 page says: 22 "It should be noted that there are also 23 defects that are not linked to POL Requirements 24 and which are not the subject of an Acceptance 25 Incidents. A separate assessment of the status 199 1 and significance of these has been undertaken 2 and this will be available for consideration at 3 the Release Authorisation Board." 4 Are you able to assist us with that at all, 5 that there were defects that weren't linked to 6 POL requirements? 7 A. Not specifically. I cannot think what would be 8 a defect that wouldn't be specifically covered, 9 you know, other than going to back to my font 10 example, which probably wasn't a specific 11 requirement. But other than things like that, 12 you know, I'd have to see some examples to 13 recall that. 14 Q. Can we look at FUJ00094265. This document, when 15 it comes up, goes back to the issue we were 16 talking about a moment ago. This is an internal 17 Fujitsu document and it has the recipient there 18 was the general council at Fujitsu, and it says 19 that the issue has been flagged as critical to 20 fix before the start of rollout. This is the 21 branch trading statement issue that we talked 22 about a moment ago. 23 There's a draft email there from Geoff 24 Butts, it's a draft response to the Post Office, 25 so it's essentially responding to that request 200 1 for assurance, and it's a draft of that. If we 2 could scroll down. 3 It's drafted to Phil, and it says -- for 4 example the problem description, it says: 5 "On the final report, the stockholding 6 figures in the second section of the report are 7 incorrect on the final balance." 8 So that's the problem we've already talked, 9 about and it provides the reassurance that is 10 asked for. If we look over the page, it has 11 "Key questions" and answers: 12 "Can Fujitsu provide a complete and 13 comprehensive list of all products where the 14 volumes on the final balance report differ from 15 those on the trial balance report? All figures 16 relating to the volume stock holdings can be 17 corrupted by the bug but not necessarily all the 18 figures will be incorrect." 19 If we go down: 20 "How has the defect arisen?" 21 It's a software issue. 22 "What is Fujitsu doing to resolve the 23 defect? 24 "Fujitsu has developed a software fix that 25 resolves the issue by ensuring that the final 201 1 report uses the original data retrieved by the 2 counter at step 2 described above rather than 3 using the data left after step 4. This is 4 a counter fix that is currently scheduled to be 5 delivered within the next maintenance release 6 for HNG-X -- 01.08. In the short-term, before 7 the fix is deployed into live, Fujitsu has 8 written a KEL which includes instructions to the 9 HSD to give advice to Postmasters. The advice 10 is to use the Trial Report, which is correct, 11 and to discard the Final Report." 12 This is something again, we were talking 13 about earlier and it was your suggestion in your 14 email that it certainly needed, if accepted, to 15 go on to the Known Error Log, and the workaround 16 there seems to be to rely on that earlier 17 statement and not the final report. So you rely 18 on the trial report not the final report. 19 A. Yes. 20 Q. So it certainly seems from this that that was 21 the position and that was agreed at that time. 22 A. Where is it saying it was agreed? I think 23 this -- 24 Q. Well, this is -- certainly, the intention here 25 seems to be to respond to the request that was 202 1 made to Fujitsu. Is your evidence that it 2 wasn't agreed or that you don't recall it was 3 agreed, or something else? 4 A. Well, all I know is the previous email from Phil 5 said it wouldn't be agreed without all these 6 caveats being put in place. I've yet to see all 7 those caveats being answered in this draft 8 email. 9 Q. Well, let's look at the final entry. It says: 10 "Can Fujitsu provide a commitment to support 11 POL in proving the integrity of the system in 12 any subsequent legal action (specifically where 13 the difference in the two reports is used as 14 a means to challenge the integrity of the 15 system? 16 "Yes, Fujitsu is willing to provide 17 commitment to prove the integrity of the system 18 in any subsequent legal action." 19 Is that something that you recall ever 20 happening? 21 A. I don't recall it ever happening. I think the 22 one above as well is interesting as well and 23 important. Please don't lose that one. 24 Q. Absolutely. So it says: 25 "Can Fujitsu provide a statement proving in 203 1 the integrity of the data is not compromised in 2 any way and the only impact is incorrectly 3 reported data? 4 "Fujitsu confirms that this defect relates 5 specifically to the printing of report data and 6 is not related in any way to the underlying 7 branch data held in the operational HNG-X 8 system." 9 A. If you recall at the start of this section 10 I talked about is it just the printout that's 11 reporting it incorrectly or is it the system. 12 Q. Yes. 13 A. By that, it sounds like it's just the printout 14 not the actual system. But, as I said, we still 15 need to go on but I still haven't seen, you 16 know -- yes, they've said they're willing to 17 provide a legal assurance to the -- to the 18 integrity and, you know, there's still more 19 missing for me, you know, because this would go 20 to Phil, Phil would, you know, take it to the 21 Finance team and also to the -- our internal 22 lawyers for advice. 23 Q. If this had been accepted as the means to get 24 around the issue, do you think it would be a bit 25 shoddy or would it be satisfactory to provide 204 1 this kind of workaround where you rely on 2 a trial report not a final report and then you 3 have Fujitsu provide a statement in criminal or 4 civil proceedings saying that everything is 5 fine? 6 A. I think, if it is just the printout that is 7 correct and Fujitsu have proven and demonstrated 8 that, and the integrity of the system has been 9 absolutely proven to be 100 per cent, then 10 I think it is reasonable. It's not ideal, but 11 it is reasonable. But, as I said, it's proving 12 that it is only the printout. It's proving 13 that, you know, the system does maintain a full 14 integrity. 15 Q. Would the attitude of the Post Office be that 16 they had to rely on Fujitsu in those 17 circumstances to guarantee those kinds of 18 assurances that you've just talked about? 19 A. No, I think you'll recall on Phil's email he 20 wanted assurance from our Finance team and from 21 our lawyers. 22 Q. Yes, but who would be able to prove the 23 integrity of the system? Were you reliant on 24 Fujitsu or would that be a joint effort? 25 A. We would run various tests to prove, you know, 205 1 and produce examples whereby the report was 2 incorrect but we'd proved the data -- underlying 3 system data was still -- had the integrity. 4 Q. We've seen certainly at the Fujitsu end that 5 this has gone to the General Counsel. Do you 6 recall how high up this particular issue went 7 within the Post Office? 8 A. I don't recall, no. 9 Q. Can we go to FUJ00094472. This is the "Notes 10 from the Horizon Next Generation Joint Progress 11 Meeting" of 11 March 2010. Can we look at the 12 bottom of page 3 into page 4, please. Thank 13 you, the bottom of that page. "PN" is Phil 14 Norton? 15 A. Yes. 16 Q. It says there: 17 "Trial report final balance issue 18 "PN to check if the proposed workaround is 19 acceptable to the business. Permanent fix 20 targeted for R1.08. However, this is dependent 21 on the acceptability of the workaround. It may 22 need to be a hot fix." 23 What was a "hot fix"? 24 A. So Fujitsu would do releases, I can't remember 25 how often they would do a release, maybe once 206 1 a week, maybe once a fortnight. A "hot fix" is 2 where they'd put a fix in between that because 3 we'd deemed it urgent. 4 Q. "GB will also prepare a statement that confirms 5 the legality/integrity of the trial balance 6 report if that needs to be used." 7 So that's the draft of which we saw earlier 8 or document relating to that we saw earlier. 9 Can we scroll up again, so we can look at the 10 next entry, it says: 11 "25 February. POL have requested that this 12 be a hot fix as it is required before we migrate 13 any further branches. Fujitsu to ensure 14 deliverables listed in PN email dated 15 24 February for current five branches are 16 included. 17 "4 March. GB has received feedback on 18 integrity statement from Fujitsu legal. Info 19 will be forwarded to PN. 20 "11 March, PN has passed statement to P&BA 21 ..." 22 You were talking to about P&BA earlier, 23 weren't you. Can you remind us what they were? 24 A. I can't remember what the "P" stands for, 25 I think it was pensions and business accounting 207 1 or -- I can't remember what the "P" stands for. 2 It's basically the central Finance team who look 3 after the cash accounting data and reconciling 4 with the clients, et cetera. 5 Q. "... who are reviewing with POL legal team. PN 6 will feed back to GB." 7 Do you recall at all how this issue ended? 8 A. I don't recall how the issue ended, no, and if 9 you note, I wasn't at this particular meeting. 10 Q. Had you moved into a different position by then 11 or were you just no longer involved in that 12 particular issue? 13 A. No, at the top I've given my apologies. I was 14 actually on holiday at this point. 15 Q. Thank you. Can we look at FUJ00094471. This is 16 a Post Office highlight report, 17 March 2010. 17 It has your name at the top there. It's the 18 third paragraph down, it says: 19 "Fujitsu are now providing evidence to POL 20 which indicates an improvement in stability and 21 therefore the pilot has been scheduled to 22 restart this week, as agreed by the POL internal 23 Release Authorisation Board on 17 March." 24 Then it goes on to talk about high volume of 25 calls to the helpdesk. It says: 208 1 "The average number of calls per Horizon 2 Online branch remains significantly higher than 3 that of the Horizon branches and the concern 4 remains that the HSD will fail to cope with 5 a significant increase in migrated branches." 6 If we look down, an entry that relates to 7 quality. This is all six days later, and it 8 seems as though the pilot is restarting and 9 these are the issues that are still occurring. 10 Screen freezes and -- it says the rate of screen 11 freezes and recoveries has been greater than 12 expected. "User opinion", it then says, "The 13 number of calls", as I just explained, to the 14 HSD: 15 "... the rate of calls from Horizon Online 16 branches is significantly higher from Horizon 17 branches." 18 You're not aware whether the branch report 19 issue had been resolved by this stage? 20 A. I'd have thought it would have been, if it isn't 21 mentioned explicitly. 22 Q. But there were still at that stage other issues, 23 despite the pilot having resumed? 24 A. Well, and that's what makes me infer that it was 25 resolved. But, again, I wouldn't have been 209 1 around at this time because I know by the dates 2 I was away. 3 Q. Looking back at all the things we've been 4 looking at, do you think that HNG-X, Horizon 5 Online, was rolled out too soon? 6 A. No. I regard we did the appropriate testing. 7 We did validate the integrity fully before we 8 migrated. We stopped in the event of any 9 suggestion there was any issue that impacted the 10 integrity or anything else that couldn't be 11 explained, you know. As I said, I personally 12 brought in the NFSP, National Federation of 13 SubPostmasters, to form part of our -- (a) we 14 bought them in to the testing so they could 15 actual get directly involved in the testing but 16 also part of our release authorisation process, 17 and they were party to every decision to 18 proceed. 19 Q. Can we look at POL00022842, please. This is 20 appendix 2 to the judgment of Mr Justice Fraser 21 and it lists or summarises bugs, errors and 22 defects in Horizon. Can we go to the second 23 page of that and there's the summary of bugs, 24 errors and defects. I'm just going to take you 25 through the bugs, errors or defects that relate 210 1 to Horizon Online. 2 We have there number 1: 3 "Receipts and payments mismatch bug 4 "This is a bug present in Horizon Online. 5 The bug occurred in 2010. The majority of 6 incidents are recorded as occurring between 7 August and October 2010." 8 Number 3: 9 "Suspense account bug 10 "This is a bug present in Horizon Online. 11 Its identified years of effect are 2010 to 2013. 12 Number 4: 13 "Dan Wellington(?) ** Bug/branch outreach 14 issue. This is a bug present in Horizon Online. 15 It's effects were experienced between 2010 and 16 2015. 17 Number 5: 18 "Remming in bug. This is a bug present in 19 Horizon Online. It was present for about five 20 months in 2010 between March and August." 21 Number 7: 22 "Local suspense account issue. Not the same 23 as 3, suspense account bug. This is a bug 24 present in Horizon Online. This was reported in 25 2010 and recorded as fixed in September 2010." 211 1 Number 8: 2 "Recovery issues. These are bugs, errors 3 and defects present in Horizon Online with years 4 of effect from 2010 to 2018." 5 Over the page, number 13: 6 "Withdrawn stock discrepancies. This is 7 a bug present in Horizon Online mentioned in 8 2010 and 2011." 9 Number 14: 10 "Bureau discrepancies. This is a bug 11 present in Horizon Online. It arose in 2017." 12 Down to number 19: 13 "Post & Go/TA discrepancies in POL SAP. 14 This is a bug present in Horizon Online. 15 Occurred in 2012. 16 "Recovery failures. This arises in Horizon 17 Online. This is referable to Horizon issue 4 18 and not Horizon issue 1. Two of the PEAKs show 19 errors and defects, they do not show bugs. 20 Mentioned in 2010, 2012 and 2015. 21 Over the page to number 23, please: 22 "Bureau de Change. This is a bug present in 23 both Legacy Horizon and Horizon Online. Arose 24 in 2005, 2006 and 2010 onwards." 25 25: 212 1 "Lyca top-up bug. This is bug present in 2 Horizon Online. Occurred in August 2010." 3 Number 28: 4 "Drop and go. This is a bug present in 5 Horizon Online, occurred in July 2017. The date 6 of the KEL suggests HNG-A but the KEL is 7 expressly headed HNG-X." 8 You were asked to specifically address 9 a list of bugs in you statement but you didn't 10 address these bugs. Is there a reason why you 11 didn't address these bugs in your statement? 12 A. I didn't know I was expected to address these 13 bugs in my statement but all I would say is 14 there's far too little information on these for 15 me to make a comment on them. It doesn't tell 16 me what the actual bug is, it just says there's 17 a bug present. For example, Lyca top-up, it 18 just says there's a bug present in Horizon 19 Online. It doesn't tell me what the nature of 20 that bug is, and, you know, whether it was, you 21 know, if you recall back, one of the things 22 I said to you was, you know, you don't develop 23 a system, deploy it and that's it. You're 24 constantly making improvements to that system, 25 be that new products, be that changes in how we 213 1 operate products, et cetera. 2 And, therefore, there's always a risk that 3 bugs do creep in and that's why you have to, you 4 know, keep testing it and, you know, keep on top 5 of it. But on none of these have I got 6 sufficient information to be able to comment on 7 them, I'm afraid. 8 Q. Do I take it from that that you didn't follow 9 any of the court proceedings relating to 10 Horizon? 11 A. Well, I followed bits of it but, you know, 12 I didn't follow in detail, you know. I looked 13 on from afar and -- 14 Q. You were given a specific list in your Rule 9 15 Request that identified bugs and asked to 16 comment on them, and can we take you to your 17 response. It's WITN03850100, that's your 18 statement, and take you to paragraph 19, page 5. 19 You said there: 20 "I had confidence in HNG-X throughout, in 21 terms of the development and testing process, 22 but we were deploying a system to over 10,000 23 branches and some bugs had been identified, and 24 whilst they were fixed naturally there was some 25 noise as we were very open and honest with the 214 1 NFSP. Ultimately their support was a vital part 2 of our rollout." 3 I mean, it was a bit more than some noise, 4 wasn't there, in terms of the number of bugs 5 that I've just taken you to, the implications of 6 it all? Do you think, again, you've minimised 7 the issues in your statement? 8 A. Sorry, let me explain to you how I've answered 9 this question. The documents I received were 10 several thousand pages, and as I said to you, 11 you know, none of those -- some of those bugs 12 were well after Horizon HNG-X was rolled out. 13 Some of those, you know -- and therefore 14 presumably introduced by a subsequent release. 15 None of those bugs have enough detail for me to 16 comment on, and I'd be commenting from, you 17 know, a guesswork, and that's totally the wrong 18 thing to do, you know. 19 I've never said there weren't bugs in the 20 system. As I said, of course there was noise. 21 We managed it very closely with the -- as 22 I said, we involved the Federation of 23 subpostmasters in all our decisions. We worked 24 with them. You know, they were very much part 25 of the team that agreed, you know, in terms of 215 1 any known errors, would we go live with them? 2 Absolutely. They were part and parcel. 3 Q. Are you still there? 4 A. Yes, I'm still here. 5 Q. Thank you. Is your evidence that after rollout 6 there weren't significant bugs that were drawn 7 to your attention that addressed issues that 8 affected the cash account and affected the lives 9 of subpostmasters? 10 A. No, I can truthfully say that from the point we 11 rolled out HNG-X/Horizon Online, that I wasn't 12 aware of any bugs that impacted the integrity of 13 the system. I wasn't aware of them. I'm being 14 very specific. 15 Q. They weren't drawn to your attention by anybody? 16 A. Well, no, and in part probably because I'd left 17 the Post Office by then. 18 Q. Well, let's say the rollout of Horizon Online 19 was 2010. After 2010 did anybody draw to your 20 attention significant errors or bugs with 21 Horizon Online that affected the cash accounts 22 of subpostmasters? 23 A. Not to my knowledge, no. I'm not saying it 24 didn't happen, but not to my knowledge, no. 25 Q. Can we look at POL00029618, please. It's 216 1 page 2. 2 This is an email from Ron Warmington of 3 Second Sight to Simon Baker. Do you know who 4 either Ron Warmington or Simon Baker are? 5 A. I've no idea who Ron is. I think Simon -- I 6 don't know what his role was. He joined the 7 Post Office before I left, but I don't know what 8 his role was at this time. 9 Q. Were you aware of an investigation being carried 10 out by Second Sight into issues with Horizon? 11 A. I broadly recall it, but I thought it was, you 12 know, with Legacy Horizon. 13 Q. We have here an email from Ron Warmington that 14 says: 15 "Simon, this is a draft section of the 16 report dealing with two defects." 17 And he asked two questions. He wants to 18 know how there were shortages, or how the 19 shortages or surpluses that arose from certain 20 problems were impacting on subpostmasters, and 21 he also wants to know when and how and who knew 22 about those problems. I'll take you to the 23 extract from the interim report. If you scroll 24 down, he has sent Simon Baker an extract from 25 his report to assist him. It says: 217 1 "POL has recently disclosed to Second Sight 2 that in 2011 and 2012 it had discovered defects 3 in Horizon Online that impacted 77 sub post 4 offices. The first defect, referred to as the 5 'Receipts and Payments Mismatch Problem', 6 impacted 63 branches. It was discovered in 7 September 2010 as a result of Fujitsu's 8 monitoring of the system events (although there 9 were subsequent calls from branches). The 10 aggregate of the discrepancies arising from this 11 system defect was £9,029, the largest shortfall 12 being £777 and the largest surplus £7,044. It 13 is not yet clear whether subpostmasters with the 14 surpluses profited from them, or whether those 15 showing shortages had to make them good. 16 "The second defect, referred to as the 17 'Local Suspense Account Problem', affected 14 18 branches and generated discrepancies aggregating 19 to £4,486, including a shortfall of 9,800 at one 20 branch and a surplus of 3,200 at another. POL 21 was unaware of this defect until a year after 22 its first occurrence in 2011." 23 So it's referring to issues in 2011 and 2012 24 that affect the cash account of subpostmasters. 25 If we go on to the page before, page 1., 218 1 Simon Baker says -- he's sending this on, and he 2 says: 3 "Just got this from Ron." 4 Sorry, it's further down. It's the bottom. 5 "Just got this from Ron. 6 "I can get back to him on most of the 7 questions but need your help on who in Post 8 Office knew about it. I know from the email 9 that Rod sent that Mike Young knew, but don't 10 know if it went any higher." 11 This is to Lesley Sewell. Do you know who 12 Lesley Sewell was? 13 A. Lesley Sewell was the new head of IT that was 14 brought in shortly before I left. 15 Q. Can we go to the top, email from Lesley Sewell. 16 She responds, 25 June 2013: 17 "Simon, I don't know if it went any higher 18 than Mike. Andy Mc ..." 19 Who is "Andy Mc"? 20 A. Andy Mc -- it was the Head of Service 21 Management. I can't remember his surname. 22 Q. "I don't know if this went any higher than Mike. 23 Andy Mc also managed the service at the time, 24 and if I remember correctly, Mark Burley was 25 also involved." 219 1 Now is Lesley Sewell wrong to say that you 2 were involved? 3 A. Involved in what? 4 Q. Well, he is asking who in Post Office knew about 5 these problems. 6 A. So I think the one on receipts not equal 7 payments we've discussed and, you know, I didn't 8 believe that was found when we were actually 9 live. I thought that was found before we were 10 live. But from the below, it might have been 11 when we were live. But, you know, the suspense 12 one doesn't ring any bells with me at all, and 13 certainly I wasn't informed of the 2011 one. 14 Q. Did Lesley Sewell ever raise, following receipt 15 of this email with you, any questions that 16 Second Sight had raised, or -- 17 A. No. 18 Q. -- or any issues concerning Second Sight's 19 report? 20 A. Nobody from the Post Office contacted me with 21 anything to do with the system after I left. 22 Q. So you think the extent of your involvement was 23 the involvement in the local -- was it the 24 receipts and payments mismatch, or the suspense 25 account issue? 220 1 A. Sorry, say the question again? 2 Q. The extent of your involvement in the underlying 3 bugs, errors or defects, you're saying that it 4 was the first, which is the receipts and 5 payments mismatch, or the second, which is the 6 local suspense account problem? 7 A. No, I'm saying I remember the receipts not equal 8 payments issue. And we did discuss that 9 earlier. I don't -- 10 Q. I think we looked at that earlier in relation to 11 Legacy Horizon rather than Horizon Online, 12 although I may be wrong on that. 13 A. Yeah, and that's what I think I'm getting 14 confused now, because I don't recall it. 15 I definitely recall it on original Horizon, but 16 all I can say is, you know, obviously there were 17 defects during the development of HNG-X. If we 18 found defects that had any chance of impacting 19 the integrity, we would make absolutely certain 20 they would be fixed. And if that involved 21 pausing, if that involved delay, that's what we 22 would do. 23 Q. Why do you think Lesley Sewell has mentioned you 24 in connection with these particular defects, 25 what are described at that time as "defects"? 221 1 A. I don't think she's saying that in connection 2 with the defects. I think she's saying because 3 I was programme manager when she first joined. 4 Q. And what would that mean? The question that's 5 been asked here by Second Sight is ... I can 6 tell you, but it's asking who knew what when, 7 effectively. And you're being -- it seems to be 8 that you're being suggested as somebody who was 9 involved. 10 A. Yeah, but as I'm saying to you, I don't recall 11 being involved. As I say, the receipts and 12 payments is an interesting one because I was 13 clearly getting confused with the original 14 Horizon. But, you know, as I say, during my 15 time as programme manager, during my time as 16 Head of Delivery, if any defects had come to my 17 attention that would remotely put at risk the 18 integrity of the system, I would have raised it 19 and insist it was fixed. There was absolutely 20 no question in my head whatsoever about that. 21 Q. The question in Ron Warmington's email is: 22 "Can you please let me know whether, when, 23 and who at board level was informed of the 24 defect, and also the later local suspense 25 account defect, and whether any press release 222 1 was issued, et cetera?" 2 Simon Baker emails Lesley Sewell to say: 3 "I can get back to him on most of the 4 questions but need your help on who in Post 5 Office knew about it." 6 Then Lesley Sewell's response is: 7 "I don't know if it went higher than Mike. 8 Andy Mc also managed the service at the time and 9 if I remember correctly, Mark Burley was also 10 involved." 11 Is that recollection of Lesley Sewell wrong? 12 A. Well, I think one, it's saying Lesley clearly 13 knew about it, reading this. And just saying 14 "If I remember correctly." All I'm saying is my 15 recollection is I wasn't aware of those two 16 specifics. But as I've said, it's my 17 recollection, you know, at this stage, which is 18 12 years since I left. 19 Q. Would you not have remembered such significant 20 impact on the cash account, given how 21 significant you have said, in your evidence 22 today, that it was to the very system that you 23 were managing? 24 A. I don't remember a specific instance, not 25 necessarily. Would I remember if anything was 223 1 raised regarding the risk to the integrity of 2 the cash account, then one hundred per cent 3 I would have insisted it was dealt with. But if 4 you ask me do I remember every single incident 5 that was raised that was connected to the cash 6 account? No, I don't. 7 Q. And do you think that Lesley Sewell was wrong in 8 naming you there? 9 A. I think that's a question you have to ask 10 Lesley, why she named me. I don't know why she 11 named me. 12 Q. But do you think it was wrong for her to name 13 you, given your level of involvement, or might 14 it actually have been right? 15 A. She's saying "If remember correctly." My view 16 is she wasn't remembering correctly, because 17 I don't recall those two specifics, especially, 18 as I said, at that date. But I can't be a 19 hundred per cent certain, as I'm saying. 20 Q. I'm going to move on very briefly, before we 21 finish, to a couple of other matters. Can we 22 look at POL00055100. This question relates to 23 your involvement in matters relating to criminal 24 proceedings. 25 This is an email exchange of July 2010, so 224 1 we're going back in time now. And rollout of 2 Horizon Online hadn't been completed by this 3 stage, I don't think. 4 It concerns a disclosure request that was 5 made by solicitors for somebody who was being 6 prosecuted, Seema Misra. She was a defendant in 7 criminal proceedings that related to an incident 8 in 2008. 9 A. I vaguely remember her name, yes. I do. From 10 the news. Not from this. 11 Q. So Merlin Benjamin, on behalf of Jarnail 12 Singh -- do you remember Jarnail Singh at all? 13 A. No. 14 Q. That's the bottom email. In fact it's Jarnail 15 Singh who is sending the email. You can see 16 over the next page. But if we highlight that 17 bottom entry, it says: 18 "I enclose a copy of an email received from 19 Issy Hogg, the defence solicitors, on 20 22 July 2010, the content of which is 21 self-explanatory. Could you please be kind 22 enough to let me have your urgent instructions 23 as to the access and information she is 24 requesting in respect of the system in the 25 Midlands and the operation at Chesterfield and 225 1 the error logs. I will contact Gareth Jenkins 2 to find out what transpired at the meeting with 3 Charles McLachlan." 4 So this was essentially a disclosure request 5 made in criminal proceedings. If we scroll up 6 we have Andrew Winn; do you know Andrew Winn? 7 A. No. 8 Q. Andrew Winn says there: 9 "John, Rod Ismay, the head of P&BA ..." 10 Do you remember Rod Ismay? 11 A. I remember Rod, yes. 12 Q. "... is not happy at the prospect of an 13 open-ended invite. He has asked the question of 14 what are the legal parameters we are working 15 with. Simplistically, if we refuse or impose 16 conditions do we lose the case? I think we need 17 more guidance on how something like this might 18 reasonably operate." 19 If we go up, who is Jon Longman? 20 A. I don't know who Jon Longman is. 21 Q. Jon Longman is responding to Jarnail Singh, and 22 he's saying: 23 "Jarnail, I've spoken to a few people 24 regarding the three points raised by Izzy Hogg." 25 That's the solicitors for Seema Misra. 226 1 "I'm awaiting comments from Penny/Gareth at 2 Fujitsu but the general consensus of opinion is 3 that these points are too vague and more 4 information is required exactly from the defence 5 as to exactly what they require. Our guidance 6 is required as per the email from Andrew Winn 7 below in relation to ..." 8 Ah, sorry. If we scroll down, sorry, 9 there's a key passage I missed off there which 10 is the email from Andrew Winn to Jon Longman 11 which says: 12 "I think Mark Burley would be the route into 13 IT to identify who might be best placed to deal 14 with this aspect." 15 A. Yes. 16 Q. Why is your name being suggested in relation to 17 this disclosure request? 18 A. Probably because at that time I was the head of 19 delivery for IT in the Post Office. 20 Q. And was it usual for you to have a role in 21 relation to criminal proceedings? 22 A. I never had a role in relation to criminal 23 proceedings. I don't even believe I was 24 contacted in respect of this. 25 Q. So why is your name being mentioned in respect 227 1 of this? 2 A. I'm afraid you'll have to ask Andrew Winn that. 3 Q. Did you ever recall being involved or asked to 4 assist in any way with evidence or matters 5 relating to criminal or civil proceedings? 6 A. I don't believe I was ever asked to deal with 7 that unless it was, you know, without me knowing 8 it was a criminal thing. 9 Q. Would you be asked to produce statements, or to 10 produce -- to carry out investigations, or to -- 11 A. No, I mean the only thing I can recall in 12 anything connected -- and I don't know if this 13 is connected to this, or something completely 14 different, was Rod did do a piece of work to 15 investigate the integrity of Horizon -- I don't 16 know if it was original Horizon or 17 Horizon Online, because the two blend, but Rod 18 did an investigation into the integrity of the 19 system at the time, whichever one it was. And 20 I'm sure there are records of that, because I'm 21 sure it's in one of the packs that I've seen, 22 where he confirmed that, you know, in his view 23 and the view of people he had looking at it, it 24 was operating with integrity. 25 Now I'd been told that was before this time, 228 1 and as I said, it could even be original 2 Horizon. As I said, I don't ever recall being 3 consulted on anything with a legal connection 4 like this. 5 Q. So you don't recall being consulted about the 6 Second Sight -- the matters raised by 7 Second Sight, and in relation to these criminal 8 proceedings, you don't recall being -- 9 A. No. 10 Q. And in both times, people have proposed your 11 name? 12 A. Yeah. 13 Q. Thank you. 14 Sir, I don't have many questions left. I 15 have about ten minutes left. I see it's 4.30. 16 I'm hoping that there will not be any or many 17 further questions. There may be something from 18 Mr Whittam. I may address it very briefly. 19 SIR WYN WILLIAMS: I'm reluctant, for a number of 20 reasons, to go much beyond, say, quarter to 21 five. And we have got the witness remote, not 22 in being, so to speak, so that if necessary we 23 could have ten minutes, quarter of an hour, in 24 the morning. Subject to his availability -- 25 I should add, Mr Burley. 229 1 THE WITNESS: I have no availability tomorrow. I 2 wasn't informed it was going to risk carrying 3 on. 4 SIR WYN WILLIAMS: I appreciate -- anyway, don't 5 let's debate it for the moment. 6 MR BLAKE: Can we see how we get on? 7 SIR WYN WILLIAMS: Let's see where we get with 8 Mr Blake. 9 MR BLAKE: Thank you. 10 You've just mentioned the Ismay report. Can 11 we go to POL00088991, please. That report is 12 2 August 2010, so it's actually very shortly 13 after this email about the Seema Misra case. So 14 Seema Misra was 27 July 2010, the top email. 15 And this report is 2 August 2010. Was this the 16 report you were talking about? The Rod Ismay 17 report? 18 A. It looks like it, yes. 19 Q. And you're copied there? 20 A. Yes. 21 Q. And this is called Horizon Response to 22 Challenges Regarding System Integrity. Do you 23 remember why this report was produced? 24 A. I think, you know, because of the reasons in the 25 first paragraph. 230 1 Q. "Post Office Limited has, over the years, had to 2 dismiss and prosecute a number of subpostmasters 3 and Crown staff following financial losses in 4 branches. A small number of those have made 5 counterclaims that they were not guilty of the 6 charges made but that Horizon System was 7 faulty." 8 So he was asked to produce a report that 9 effectively defended the position, or? 10 A. No. So the first paragraph talks about, you 11 know, a small number of claims were not guilty 12 of the charges but the Horizon System -- so 13 that's the original Horizon System. But it's 14 actually then just validating that the new 15 system has got integrity. Obviously you can't 16 validate that the old system had got integrity 17 because we'd replaced it by then. 18 Q. Can we scroll down to the executive summary, 19 please. It says there: 20 "POL has extensive control spanning systems, 21 processes, training and support. Horizon is 22 robust but like any system depends on the 23 quality of entries by the users." 24 So it seems there to be suggesting that the 25 quality of the entry of users is the issue, or 231 1 the significant issue. 2 "Horizon Online builds on this and brings 3 benefit to running costs and change management. 4 It is not being done because of any doubt about 5 the Horizon of integrity." 6 It then says: 7 "The integrity of Horizon is founded on its 8 tamper-proof logs, it's real time back-ups and 9 the evidence of 'backdoors' so that all data 10 entry or acceptance is at branch level and is 11 tagged against the log-on ID of the user. This 12 means the ownership of the accounting is truly 13 at branch level." 14 Now that wasn't correct, was it, about the 15 absence of backdoors and everything being at 16 branch level? 17 A. How do you mean? 18 Q. Well, was that correct? Did you agree with that 19 proposition? 20 A. To a degree. I might have worded it slightly 21 different. And the absence of unauthorised 22 backdoors. 23 Q. So there were backdoors, but they were -- 24 A. Well, any system has to be maintained, yes. So 25 you do need, you know, a control system such 232 1 that you can go in and make appropriate, you 2 know, new coding, for example, et cetera, that 3 needs to be done on any system. But it's that 4 authorised backdoors that I'm emphasising. 5 So on Horizon Online, it was certainly 6 always built and tested as a two-person entry. 7 So there had to be two persons involved to make 8 sure it had that authorisation. 9 Q. When you read this for the first time, did you 10 say, "That bit's not right"? 11 A. I didn't say it's not right; I just questioned 12 the wording. 13 Q. Who with? 14 A. Myself. I read it -- you said when I read it 15 for the first time, so when I -- 16 Q. You received it 2 August 2010, you're copied in. 17 On reading this, did you say, "Hang on a minute, 18 what it's saying about backdoors is wrong"? 19 A. I don't recall, to be honest, at that point, who 20 I raised it with. I was more interested in what 21 were the important lessons that we had to learn 22 from it. 23 Q. The final paragraph on that page: 24 "There are several improvement opportunities 25 for POL, and these are set out in Appendix 1. 233 1 They do not undermine POL's assertion regarding 2 the integrity of Horizon but they would tackle 3 some of the other noise which complainants feed 4 on." 5 That phrase, "noise", that was something 6 that you -- that was a phrase that came up in 7 your witness statement, as well, that I took you 8 to earlier. Was that a common word that was 9 used at the Post Office to describe the 10 complaints about the Horizon System? 11 A. Yeah, but only in the sense of, you know, noise 12 can be good or bad. So, you know, don't 13 interpret it as being necessarily negative. 14 It's negative in this context, and obviously we 15 had to take note of the noise and make sure we 16 acted upon it. And that's one of the reasons, 17 as I said, we had the Federation in with us -- 18 so they could help us with that noise and help 19 us either (a) understand it, or (b) you know, 20 they would help promote the right communications 21 that would help alleviate the noise. 22 One of the biggest bits of noise that we had 23 was in the early days of deployment, because we 24 made a conscious decision to substantially 25 change the user interface, and because of that, 234 1 we had quite a bit of what I'm going to use the 2 term "noise", because people were not used to 3 it. But actually, very quickly that noise died 4 down from the users, but it almost kept a 5 balance because as new users came on, you got 6 the same noise because they just weren't used to 7 that new user interface. And it was 8 specifically done that way because, had we kept 9 the user interface as it was, that was part of 10 the problem with original Horizon and why some 11 transactions were taking longer than they needed 12 to: because they had to navigate several 13 screens, whereas we made it much quicker and 14 easier to navigate to transactions. 15 Q. Thank you. That's not quite the point I'm 16 making, but I'm going to have to move on because 17 of time. But I just ask, are you really saying 18 that the phrase "noise" there is used in 19 a neutral term, as a neutral term, or is it 20 being critical? 21 A. I think on this one it's critical. It's 22 a negative in this one. 23 Q. Can we look at POL00091384, please. This is 24 document from November 2010. It's an email. 25 It's an email from Lynn to Mike and Rod, so 235 1 including it to Rod Ismay. It's the second -- 2 Lynn Hobbs. If we look over the page, sorry. 3 Who was Lynn Hobbs? 4 A. At the time, Lynn would have been probably 5 a regional manager, I think, at the time. 6 Q. This is November 2010. She says: 7 "I found out this week that Fujitsu can 8 actually put an entry into a branch account 9 remotely. It came up when we were exploring 10 solutions around a problem generated by the 11 system following migration to HNG-X. This issue 12 was quickly identified and a fix put in place, 13 but it impacted around 60 branches and meant 14 a loss/gain incurred in a particular week in 15 effect disappeared from the system. One 16 solution quickly discounted because of 17 implications around integrity was for Fujitsu to 18 remotely enter a value into a branch account to 19 reintroduce of the missing loss/gain. So POL 20 can't do this, but Fujitsu can." 21 Was this the point you qualified earlier, 22 that actually, it is possible for Fujitsu to -- 23 A. Yes. That should have been two people doing 24 that. I don't know, obviously, in this 25 particular case whether it was, but it should 236 1 have been. 2 Q. If we look at page 1, we have, at the bottom of 3 page 1, an email from Lynn Hobbs. She says 4 there in the third sentence: 5 "I haven't seen anything further, but I did 6 have a conversation with Mike and the whole 7 remote access to Horizon issue. This was being 8 looked into by Andy McLean and Mark Burley." 9 Do you remember being asked to look into it 10 at all? 11 A. I don't remember being specifically asked to 12 look into this one, no. 13 Q. Do you think that's likely, or again, was your 14 name being misused? 15 A. I honestly don't recall. At 3 December 2010 it 16 could have been I was asked to, and I would have 17 then passed that on to the then Horizon 18 programme manager, which was Will Russell, 19 because I would have been in a head of projects 20 role at this stage. And, you know, I'd have 21 actually given an honest answer that Fujitsu can 22 access it, as long as there's two people doing 23 the piece to give that control. 24 Q. If we look up to the top of this chain, from 25 John Breeden. Do you know who John Breeden was? 237 1 A. No. 2 Q. He was the national contract manager, North Post 3 Office. He's emailing Angela van den Bogerd. 4 Do you remember Angela van den Bogerd? 5 A. Yes, to a degree, yes. 6 Q. And who was she? 7 A. She was something in the network, but I don't 8 remember what her exact role was. 9 Q. Okay. Do you recall how high up within the Post 10 Office this issue of remote access, and the 11 ability of Fujitsu to remotely access the 12 account, was known? 13 A. No, I don't know. 14 Q. Do you think it would be higher than Angela van 15 den Bogerd? 16 A. I don't know what level Angela van den Bogerd 17 was, but yes, I'd like to think so. 18 Q. Having read the Ismay report in the summer -- 19 A. -- to be unauthorised access. That's the key 20 thing here. If it's just a misunderstanding and 21 it turns out that Fujitsu have legitimately 22 accessed the system, then I wouldn't expect it 23 to go higher. 24 Q. Having read the Ismay report in the summer, and 25 we're now in December of 2010, did you tell 238 1 anyone, "Hang on a minute, the Ismay report, 2 once again this has reminded me, that bit about 3 remote access, that's wrong"? 4 A. I'm not sure why I would, because I'm not on 5 this, am I? 6 Q. Well, no, but you've been mentioned about 7 looking into the problem, looking into the 8 remote access. So you were in 2010, it seems, 9 according to the email from Lynn Hobbs at the 10 bottom of this page, charged with looking into 11 the remote access with Horizon issue. Having 12 looked into it, did you consider raising any 13 flags? 14 A. As I said to you, I was only asked to look into 15 it in December 2010. 16 Q. You can't recall looking into it, did you say? 17 A. I don't recall being asked to look into it. I'm 18 not definitively saying I was or I wasn't, but 19 I don't recall. 20 Q. Thank you. One final, very small issue -- and 21 I will only take five minutes on this -- it's 22 the issue of training. Can we look at 23 FUJ00000559. In your witness statement, 24 paragraph 8, you say you can't comment on 25 training for Legacy Horizon. There is 239 1 a document that I've been asked to bring to your 2 attention. It's 12 May 2000. It's the final -- 3 at the bottom of it -- it's a Change Control 4 Note and it refers to you in that bottom 5 paragraph: 6 "A number of discussions have been held to 7 identify a suitable mechanism for delivering the 8 results requested, and Mark Burley confirmed to 9 John Pope by email on 12th May 2000 that the 10 only change Pathway was now being asked to make 11 was the one detailed below." 12 And we go over the page, the Change Control 13 Note in itself isn't important. The passage 14 that I'm going to draw to your attention is 15 this, it's the second paragraph: 16 "This Change Control Note is based on the 17 assumption that to minimise costs, the training 18 systems will not be updated." 19 Do you recall, during your involvement with 20 Horizon -- Legacy Horizon/Horizon Online -- 21 training systems not being updated because of 22 costs issues? 23 A. I don't recall it, no, and I'm surprised that it 24 wasn't. But looking at the actual detail of the 25 change, this was a change to a non sort of 240 1 financial value cash account entry. It's not 2 a financial entry. My guess is it's not that it 3 won't be updated; it will not be updated at this 4 time. I can't believe it wasn't updated, full 5 stop. 6 Q. Thank you. 7 And the final question also relates to 8 training. You have said in your statement -- 9 you refer to the NFSP in the context of the 10 training and assurance with HNG-X. You say that 11 there was a test before official use of the 12 system was approved, and that this was developed 13 across multiple parties including the NFSP. Can 14 you tell us who else was involved, briefly, and 15 your recollections of what the NFSP involvement 16 in developing that was, vis à vis any others? 17 A. Yes. So one of the things that I put in place 18 was -- we called it a tripartite way of working, 19 and it involved the NFSP, who represented all of 20 the subpostmasters; the CWU, and some members of 21 my programme team. Now, on occasions that 22 created conflict in itself because sometimes the 23 CWU, who represented the directly owned 24 branches, and the NFSP would have different 25 priorities. But they were all involved in 241 1 putting together the testing, which my Business 2 Change Team would have led. And probably -- 3 I can't recall exactly, but probably -- 4 undertook that testing to assure it was fit for 5 purpose, would be my guess, that they had got 6 involved with. 7 Q. Put very simply, what was the difference between 8 the NFSP and the CWU, in very simple terms, to 9 the best of your recollection? 10 A. Well, the difference is a priority of things, 11 because it depends on how their members get 12 paid. So the NFSP might want to prioritise 13 a transaction of, you know, to a higher screen 14 than the CWU because their members get a higher 15 payment for it, whereas for the CWU, you do more 16 of transaction X which wasn't as important to be 17 higher up on the Federation because they didn't 18 get paid as much for it. 19 It was always resolved amicably. We had 20 discussions, we had -- it was just, you know, 21 one of the examples that I recall where there 22 was a conflict. 23 MR BLAKE: Thank you very much. Mr Whittam has 24 a brief comment or question. 25 MR WHITTAM: Sir, given the hour, a document has 242 1 been referred to. We can address the issues we 2 want to address at a later stage. I don't need 3 to ask any questions. 4 SIR WYN WILLIAMS: Yes. I think that we've now run 5 our course, Mr Blake. And if anybody did have 6 any further questions, I was going to say that 7 they would have to submit them to me first so 8 I could decide whether it was necessary to ask 9 them. And if I so decided, we'd have to arrange 10 with Mr Burley when he could be asked them. 11 There's a limit for both him and the transcriber 12 and, I have to say, me, as I get older. After 13 4.30 it's quite difficult. 14 Mr Burley, that's it. Thank you for coming 15 to give evidence. And I suspect that we won't 16 trouble you further, but that's not an absolute 17 guarantee, all right, because I have cornered 18 people into being quiet for the moment. 19 THE WITNESS: Well, any help I can give. 20 SIR WYN WILLIAMS: Right, fine. 21 10.00 tomorrow, Mr Blake? 22 MR BLAKE: Thank you. 23 SIR WYN WILLIAMS: Fine. 24 (4.47 pm) 25 (The hearing adjourned until 10.00 am 243 1 the following day) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 244 I N D E X SUSAN HARDING (affirmed) .............................1 Questioned by MR BEER .........................1 Questioned by MR STEIN .......................89 Questioned by MS PAGE ................108 MARK BURLEY (affirmed) .............................124 Questioned by MR BLAKE ......................124 245