1 Tuesday, 28 February 2023 2 (10.00 am) 3 MR BLAKE: Good morning, sir. 4 SIR WYN WILLIAMS: Good morning. 5 MR BLAKE: Sir, before we begin with today's witness 6 I'd just like to say that yesterday the Inquiry 7 received over 700 documents from Post Office 8 Limited that are potentially relevant to issues 9 to be explored in Phase 3. This disclosure 10 includes documents said to be provided to 11 members of NBSC to assist in dealing with calls 12 for assistance from subpostmasters. 13 The Inquiry considers that the vast majority 14 of these documents should have been provided 15 further to a Rule 9 Request sent on 18 February 16 2022. 17 We have considered carefully whether to seek 18 to postpone hearing some of the witnesses this 19 week until such time as those documents have 20 been fully reviewed and disclosed to Core 21 Participants. However, we propose not to do so. 22 This is because the witness can speak to 23 a number of other issues and those issues should 24 be investigated now. Your legal team will 25 continue to review the newly disclosed documents 1 1 and provide them, where they are relevant, to 2 Core Participants in due course. 3 We propose keeping under consideration 4 whether it's necessary to arrange additional 5 sitting days to hear oral evidence on those 6 issues raised in that disclosure. 7 The Inquiry can, of course, recall any 8 witnesses heard this week to speak to those 9 documents should it become necessary. 10 SIR WYN WILLIAMS: Thank you, Mr Blake. I think you 11 probably know that I was informed of this state 12 of affairs, to be precise, at about 7.05 last 13 evening, so I have had the opportunity to 14 reflect upon the course of action which you 15 propose overnight and, in short, I agree with 16 it. I don't think anything would be served by 17 adjourning for a short period and asking 18 witnesses who made themselves ready to give 19 evidence not to do so. I think the course that 20 you suggest is preferable, and so that's what 21 we'll do. 22 MR BLAKE: Thank you very much, sir. In light of 23 that, may I please call Mr Shaun Turner. 24 SHAUN TURNER (affirmed) 25 Questioned by MR BLAKE 2 1 MR BLAKE: Thank you. Can you give your full name, 2 please. 3 A. Yeah, Shaun Turner. 4 Q. Mr Turner, you should have in front of you 5 a witness statement dated 13 January 2023. 6 A. I do. 7 Q. It has the URN WITN04640100. If I can ask you 8 to turn to page 91 of that statement, is that 9 your signature there? 10 A. It is. 11 Q. Thank you. Can you confirm that statement is 12 true to the best of your knowledge and belief? 13 A. It is. 14 Q. Thank you. I know that there are couple of 15 areas that you'd like to expand upon or clarify 16 in due course. We don't need to address those 17 now, we can address them shortly. That 18 statement is going into evidence, so I'm not 19 going to ask you about every single detail 20 contained in that witness statement. The 21 questions I'll ask today will be supplementary, 22 although I am going to start by asking you a bit 23 about your background. 24 That takes up several pages of your witness 25 statement because you've held a lot of positions 3 1 and a lot of those positions have quite long 2 titles, so I'm going to try and go through a few 3 of those to assist the Inquiry with a little bit 4 more detail. 5 But it's clear you were employed by the Post 6 Office since 1996. Did you hold any other job 7 before that or was that your first significant 8 employment? 9 A. Apart from part-time work, it was essentially 10 the first job that I picked up after university. 11 Q. Thank you. You say in your statement that 12 between 1996 and 1998 you worked in the 13 northeast region helpdesk and customer care. 14 Can you tell us -- that was pre-Horizon? 15 A. Yes. 16 Q. What did working on the helpdesk pre-Horizon 17 involve? 18 A. It involved taking calls from branches and, at 19 that time, from also Post Office customers, 20 resolving queries, resolving customer 21 complaints. 22 Q. You were, at that stage, familiar with balancing 23 in the paper method? 24 A. Yes. 25 Q. 1999 you became an NBSC analyst. 4 1 A. Mm-hm. 2 Q. You're the first witness from the NBSC, so can 3 you please tell us what was the NBSC? 4 A. The NBSC was the Network Business Support Centre 5 it was set up in line with the Horizon rollout. 6 So the idea was, as branches migrated onto 7 Horizon, they would stop contacting their 8 regional helpline and start contacting the NBSC 9 and the NBSC would then become the helpline for 10 Horizon branches, fulfilling the same function 11 as the regional helplines did. 12 Q. I think we've heard in previous evidence about 13 first line of support, second line of support. 14 Can you tell us briefly where you fit into the 15 mix of different levels of support? 16 A. Yes. I was aligned with the second line of 17 support which was called tier 2, which was 18 operated by Post Office, so that was the 19 advisers that dealt with the more complicated 20 enquiries. The first tier was actually not 21 managed by Post Office; it was managed by 22 another part of Royal Mail Group called SSL and 23 they had their own sort of management structure, 24 team leaders and they dealt with the sort of 25 more basic enquiries, anything that was more 5 1 complicated or required, so a longer diagnosis 2 would be passed to the tier 2 team. 3 Q. If I were a subpostmaster or an assistant and 4 I wanted to speak to somebody about a problem 5 with Horizon, who would I call in the first 6 instance? 7 A. It depends what the problem was. If it was sort 8 of operational procedures or navigation, it 9 would be the NBSC and, in the first instance, 10 that would be the tier 1 NBSC and, if they 11 couldn't deal with it, it went on to tier 2. If 12 it was system-related issues, your printer 13 wasn't working, something like that, it would be 14 HSD, which was managed by Fujitsu. 15 Q. Is that the Horizon Service Desk? 16 A. Yes. 17 Q. Thank you. How would I know which to call? 18 A. There were communications to branches that made 19 it clear that if it is a system problem, then 20 you ring this helpdesk; if it is a navigation or 21 operational kind of enquiry, then you ring the 22 NBSC. That wasn't always followed. We did get 23 a lot of branches ringing the wrong helpdesk and 24 that resulted in, at times, transfers between 25 the helpdesks where we had to reroute callers. 6 1 Q. Were there policies in place that clearly 2 defined the roles of the NBSC and the Horizon 3 Service Desk, vis à vis one another? 4 A. I don't remember any policy documents as such. 5 I remember communications and, in the materials 6 that went out to branches, that that was made 7 clear. And I think from a contractual 8 perspective, from what I understood, it was 9 quite clear from a contractual perspective what 10 the sort of delineation of those two helpdesks' 11 scope and responsibility was. 12 Q. Let's say if I had a bug with my Horizon 13 terminal, who would I call? 14 A. If you thought that there was a problem with the 15 system then you would call HSD. 16 Q. If I had a problem balancing but I didn't know 17 what was causing it, who would I call? 18 A. NBSC in the first instance. 19 Q. You said there was some confusion between the 20 two by postmasters. Can you describe that 21 a little bit for us? 22 A. Yeah, I think it was just general -- this was 23 new to postmasters, there was a learning curve, 24 they were trying to deal with a lot, trying to 25 learn a lot in the branches and sometimes they 7 1 just rang the wrong helpdesk and needed to be 2 rerouted. But there were other kind of queries 3 that may involve or may require the interaction 4 with both helpdesks. 5 So, for example, if you were trying to 6 produce your trading statement, your cash 7 account, and your printer broke, you would need 8 to report that to the HSD but you may also need 9 to report to the NBSC that your cash account 10 would be delayed, for example. 11 Q. Was there ever a single point of contact or 12 would you, the subpostmaster, have to make 13 contact with both? 14 A. We would generally, the helpdesk that dealt with 15 the call first would deal with their bit of the 16 call and then transfer the call across to the 17 other helpdesk to deal with the final bit. 18 Q. Again, in relation to policies that existed, was 19 there something that set that out, that the 20 first helpline would be the one that dealt with 21 the problem or took leadership or ownership of 22 it? 23 A. Not to my recollection. 24 Q. Can you tell us at 1999 what was your role at 25 the NBSC? 8 1 A. So in 1999 I was -- in late 1999, I became 2 an NBSC incident analyst. So that was a role -- 3 one of two roles that were attached to tier 2 in 4 the NBSC and we had accountability for building 5 reports to monitor calls coming in to the 6 helpdesk and we had a link to the Problem 7 Management Team, if there were issues that we 8 saw kind of repeating. 9 Q. Would you speak directly with branches in that 10 role? 11 A. Not typically, no. I'm not saying that that 12 never happened but, typically, it would be the 13 advisers or sometimes the team leaders and they 14 would sometimes bring issues to us, but not 15 typically during the course of my role as an 16 incident analyst, no. 17 Q. So your role was principally analysing trends 18 and things like that? 19 A. Yeah, I mean, it did perform a sort of ancillary 20 support function for the advisers because their 21 team leader may not always be about. So, you 22 know, any manager could be approached with 23 queries from advisers. So I'm not saying that 24 there was no interaction with postmasters but 25 I was saying it would be rare, compared to 9 1 a team leader's, who would have day-to-day 2 interactions with postmasters. 3 Q. September 2001, you became network performance 4 analyst. What, in simple terms, did that role 5 involve? 6 A. So this was a -- I moved out of the NBSC at this 7 point. This was a role, again principally 8 a role related to data and analysis and this was 9 looking at sort of conformance related data. So 10 this may be errors that the branch was making, 11 and trying to build call campaigns, outbound 12 call campaigns, to speak to branches about those 13 errors and ultimately reduce them. 14 Q. You moved on to a banking project before 15 becoming network conformance and capacity 16 manager in 2003. 17 A. Yes. 18 Q. I think in that role you looked again at error 19 rates in conformance within branches to 20 prioritise calls; is that right? 21 A. That's right, yeah. 22 Q. Can you tell us briefly what that involved? 23 A. That was essentially a rerun of the earlier 24 role. It was just moved into the NBSC. So 25 there was a separate team -- in 2001 there was 10 1 a separate team that made those outbound calls, 2 which was from Chesterfield but, when I moved 3 into the network conformance capacity manager 4 role, that work was moved back to the NBSC. It 5 was actually undertaken by the tier 2 advisers, 6 one of the teams. It was kind of rota'd round 7 the teams each week. 8 And also that role, we did sort of capacity 9 planning, so we were sort of analysing call 10 arrival patterns and trying to make sure we had 11 the relevant cover on tier 2, at the helpdesk, 12 to meet the call volumes. 13 Q. In 2005, you became network co-ordination 14 adviser. Again, very briefly, can you tell us 15 what that role involved? 16 A. Yeah, so this was a -- it was initially 17 a project role when I first moved into it. 18 There was some reorganisation at that time in 19 the retail structure, and what they were trying 20 to do at that stage was -- traditionally, post 21 offices had a retail network manager that was 22 connected to each branch and this piece of work 23 was to have a much more kind of distinct 24 separation between sales and service. So the 25 sort of on the ground resources would be focused 11 1 on the top sales branches and the service 2 related issues, things like ordering date 3 stamps, that kind of thing, that had 4 traditionally gone to the retail network manager 5 were kind of rerouted centrally through the 6 NBSC, and if they couldn't resolve it, it would 7 go on to the service teams who would fulfil 8 a kind of reactive capability of going out to 9 branches to resolve issues. 10 So it started off with that piece of 11 transformation work, building the processes, 12 developing the processes with NBSC, and then it 13 morphed into a team that was essentially 14 attached to that service function and was 15 responsible for managing relationships with 16 other business stakeholders. 17 So, for example, Product and Branch 18 Accounting in Chesterfield, the Problem 19 Management Team. We would get sort of issues 20 raised to us from the service managers and we 21 would take those up with our contacts in those 22 various teams. 23 Q. After that role, you moved to several different 24 roles. Is it right to say that they were 25 predominantly data-related roles, analysing 12 1 data; is that your particular area of expertise? 2 A. Yeah, I think that's a fair categorisation of 3 the roles that I held from sort of 2010 to late 4 2015. They were data-related roles and they 5 were predominantly concerned with conformance 6 and compliance specifically. 7 Q. That involves drawing trends together from 8 various pieces of data to see how Post Office 9 can perform better? 10 A. Yeah, and this is mainly to do with -- or at the 11 time it was mainly to do with sort of mails 12 compliance specifically. In 2012, we signed the 13 Mails Distribution Agreement with Royal Mail, 14 and that had certain sort of SLAs that the Post 15 Office were expected to hit and things like 16 segregating your mail correctly in a branch. So 17 it was mainly related to that sort of data, and 18 sort of targeting outbound call interventions to 19 branches to kind of coach them to improve 20 performance. 21 Q. I think you're still at the Post Office? 22 A. Correct, yeah. 23 Q. What level within the company have you reached? 24 A. I guess, sort of -- well, lower senior manager, 25 I guess, in more recent years. 13 1 Q. Do you ever attend board meetings or anything 2 along those lines? 3 A. No, that would be several rungs above me. 4 Q. So in all of your roles, so some were speaking 5 directly with branches on occasion or others in 6 analyst roles, did you look at common problems 7 that were cropping up and think that there was 8 a problem with Horizon? 9 A. When I was with the NBSC there was certainly 10 problems that were being faced by branches which 11 were coming through in the trend analysis and we 12 raised some of those to Problem Management in 13 due course. I can recall, for example, raising 14 problems on checklistings, which was an area of 15 confusion for branches. I can recall raising 16 problems on the declaration process, so that was 17 rather clunky and could cause confusion, and 18 I recall raising problems on obsolete stock 19 process as well, so various things that would 20 cause calls to come in to NBSC. 21 Q. When we talk about bugs, errors and defects in 22 Horizon, were there any trends that you thought, 23 "Hang on a minute, there's a problem here"? 24 A. No, I don't think -- there was obviously issues 25 in branches with balancing. That was evident on 14 1 the desk, you know, we were getting a lot of 2 calls on balancing, not just discrepancies but 3 just the process itself, how to work through it. 4 But I didn't, at that time, think that there was 5 any kind of particular issues with Horizon. 6 Q. When you say "at that time", which time was 7 that? 8 A. When I was an incident analyst in 1999 and 2000. 9 Q. So the early days of Horizon? 10 A. Yes. 11 Q. As things progressed throughout the life of 12 Horizon, have you ever thought "From what I'm 13 seeing, there's a problem here"? 14 A. I think, as I kind of moved into the sort of 15 service area in 2005, and perhaps in that 16 intervening period, my thinking did evolve. It 17 wasn't at that stage that I thought that the 18 Horizon System was infallible but what I did 19 think was that bugs were monitored, if there 20 were issues with the Horizon System, and could 21 be identified relatively easily by Fujitsu and 22 HSD; if you referred a branch to them, it should 23 be clear to them. 24 So I don't think I thought that the system 25 was foolproof but I thought there was monitoring 15 1 in place and bugs would be quite easily 2 identifiable. 3 Q. We'll come on to talk about bugs shortly but I'm 4 going to start with training, and that's 5 something that's addressed in your witness 6 statement. 7 A. Yes. 8 Q. Can you tell us what level of training did you 9 have, when the NBSC started, in relation to 10 Horizon? 11 A. I don't recall exactly. There was some 12 training. I think either a week or two weeks of 13 training but I don't recall precisely what 14 I personally received. 15 Q. Were you provided, do you think, with more or 16 less training than subpostmasters were provided? 17 A. Um ... I think postmasters, if I recall 18 correctly, had two days or a day and a half. 19 So, on that basis, I'd have to say that we had 20 more, yes. 21 Q. Would you have expected NBSC advisers to know 22 their way around Horizon as well as or better 23 than subpostmasters? 24 A. Yes. 25 Q. You've described in your statement the training 16 1 of subpostmasters in what you describe, I think, 2 the early to mid-years of your career as having 3 gaps. Can you tell us briefly about those, 4 please? 5 A. Yeah, I think this is just a reflection on my 6 experience on the desk. I mean, we'd moved from 7 a situation with the regional helplines where, 8 you know, calls tended to be quite short, even 9 on balancing after Wednesday afternoons when the 10 branches balanced. There was not really too 11 much of a discernible spike in calls. But when 12 we moved to NBSC and the roll out of Horizon, we 13 did start seeing not just more calls but longer 14 calls, calls that took considerable kind of 15 diagnosis to get to the bottom of it and, 16 therefore, that leads me to conclude that there 17 were some weaknesses with the training. 18 My own personal view at that time was that 19 we had underestimated the sort of step change 20 that branches were going through. This was 21 a considerable -- considerably different kind of 22 automated environment that they were operating 23 in when they'd been used to a paper-based system 24 for years. So that's my view at the time and 25 those were things that were escalated up the 17 1 chain as well. 2 Q. Whose job within the Post Office would it have 3 been to plan for that kind of event? 4 A. During the implementation of Horizon? 5 Q. Yes. So the weaknesses you have identified, 6 whose job was it within POL to make sure that 7 those didn't occur? 8 A. The implementation team. 9 Q. Who in particular? 10 A. I think it was led by Don Grey at that time. 11 Q. You said in the early to mid years of your 12 career. What period are we talking about where 13 you have identified gaps in the training of 14 subpostmasters? 15 A. I think it was clear to me that there were gaps 16 during the roll out of Horizon. But I don't 17 think we've got everything right since then. 18 That's not what I'm claiming. I think there 19 were still gaps and we've made improvements even 20 in more recent years in the last three or four 21 years, in areas that I believe improve the sort 22 of training offer to postmasters. 23 So I think there's been gaps around the sort 24 of -- particularly around balancing the 25 diagnosis of discrepancies and rectification of 18 1 discrepancies for a good few years. 2 Q. Horizon began '99, 2000, so are we talking 3 a decade longer? 4 A. I think things improved after Horizon, in terms 5 of the training offer. But I still think the 6 sort of diagnosis of discrepancies is, you know, 7 that area in particular did need improvement. 8 And, as I say, we have improved things in the 9 past three to four years. So, I would say, I'm 10 happy with where it is now but between the roll 11 out of Horizon and, say, four or five years ago, 12 I would have been less happy. I should make the 13 point here, as well, that during -- after sort 14 of 2005, I wasn't sort of directly connected 15 with the helpdesk or the training area, so these 16 were just kind of my personal perceptions. 17 Q. You've said in your witness statement that there 18 was an additional optional classroom session of 19 investigating discrepancies. Do you know when 20 that came into force or was brought in? 21 A. I believe 2020. 22 Q. So, again, that's one of the measures that 23 you're talking about that's improved the 24 position in more recent years? 25 A. Yeah, because I'd moved back into the training 19 1 area since 2020, so I'm kind of more familiar 2 now with some of the improvements of the last 3 two years. There may have been improvements in 4 the intervening period but, as I say, my 5 perception has been, given that we still had to 6 produce or implement that investigating 7 discrepancies course in 2020, that there were 8 still gaps and still things that needed to be 9 improved on. 10 Q. One thing you've said in your statement is that 11 there's still less classroom training nowadays. 12 Can you expand on that briefly? 13 A. Yeah, so when I joined the Post Office I think 14 I'd three to four weeks' training in classroom. 15 That was alongside postmasters and people going 16 to work in Crown branches, and that has been 17 scaled back over the years. Now, some of that 18 has been replaced with digital learning. So 19 there is a chunk of digital learning that you do 20 nowadays before you go to the classroom and 21 there is the additional sort of investigating 22 discrepancies course that we have, as well, and 23 on site, as well. Six days on site. 24 But, yeah, I think there's no doubt that it 25 has been, you know, scaled back over the years. 20 1 Q. Is that for cost reasons or for some other 2 reason? 3 A. I think it's -- the digital offer has obviously 4 replaced the need to be in classroom for some of 5 that, as to what the sort of rationale behind it 6 was, I wasn't connected to those decisions so 7 I couldn't really say whether cost was a factor 8 or there were other rationale. 9 Q. You raise it in your statement, have you raised 10 it elsewhere within the company? 11 A. Not -- I mean, I only came back into -- I came 12 into the training area in 2020. So it wasn't in 13 my sphere, as it were, in the sort of previous 14 years. 15 Q. Can we look at POL00035756, please. This is 16 a document you mention elsewhere in your witness 17 statement under a different topic but I just 18 want to talk about this in relation to training 19 at the moment. It's a document that has the 20 words "Compliance Training" on the left-hand 21 side. 22 A. Mm-hm. 23 Q. It's dated 25 March 2016. Do you know the 24 background to this document at all? 25 A. Yeah, I believe this was part of the Business 21 1 Technology Transformation Programme, as it says 2 there. So this was looking at various 3 improvements to Horizon, at which compliance 4 training was one. 5 Q. If we look at page 18 of this document, I just 6 want to draw to your attention a couple of 7 entries in a table, it's the top table on 8 page 18. It says there "Project Sparrow 9 Recommendations"; do you know what Project 10 Sparrow was/is? 11 A. I believe that it was a project within Post 12 Office looking at discrepancies and ways to 13 identify them. I think, in terms of this 14 particular document, it was -- the Project 15 Sparrow was providing information, as 16 I understood it, on ways that the system could 17 be adapted to prevent branch mistakes or 18 discrepancies. 19 Q. Have you had any direct involvement with Project 20 Sparrow? 21 A. Other than their input into the documents as 22 reviewers, no. 23 Q. The issue identified there is: 24 "Postmasters do not have access to enough 25 training to feel competent in running their 22 1 branch. Where training has taken place, records 2 have not been kept consistently." 3 The "Rationale for change": 4 "POL need to be able to demonstrate training 5 across the network and maintain robust training 6 records for branches and branch staff." 7 This is 2016. Is this a broader reflection 8 that postmasters don't have access to enough 9 training or is this specifically relating to 10 compliance? It certainly seems to be broader 11 than simply the issue of compliance. 12 A. My take on this is that it is relating to 13 compliance specifically, not least because it is 14 in a document about compliance training but 15 I think "Consideration for Front Office" on the 16 right there does refer to compliance training. 17 Q. Is it your belief that in 2016 there was or 18 wasn't some thinking within the Post Office 19 about the sufficiency of the training that was 20 provided, more broadly? 21 A. I'm sure there was but I wasn't party to those 22 discussions around the broader considerations 23 around training. 24 Q. I'm going to move on now to the issue of advice 25 and assistance. In a couple of the roles that 23 1 you've set out for us, you were advising 2 branches, to some extent, and assisting them 3 directly. Can you tell us what roles they were 4 and the level of your involvement directly with 5 branches? 6 A. Yeah, so in the NBSC, as I've already mentioned, 7 as an incident analyst you would sometimes be 8 called upon by a tier 2 adviser to support the 9 call that they were dealing with, so that was in 10 both my stints with -- within the NBSC, so from 11 1999 to 2001 and from 2003 to 2005 I would have 12 fulfilled that role. 13 Wider than that, in the network 14 co-ordination adviser role from 2005, I would 15 occasionally speak to branches, so this -- as 16 I mentioned earlier, we would deal with sort of 17 escalations from the service line where, for 18 example, transaction corrections needed chasing 19 up. So I'd deal with the branches directly with 20 that. 21 And then sort of from 2010 to 2016 when 22 I was in those data roles, that was with the 23 branch standards team, so we were making 24 outbound calls to branches about those various 25 compliance issues so, occasionally, I would 24 1 interact with branches on those as well. 2 Q. Can you tell us your view of the adequacy of 3 support concerning issues such as balancing, in 4 the early years of Horizon? 5 A. Yeah, I think, speaking about the period when 6 I was on the helpdesk on the NBSC as an incident 7 analyst, I think it was a very difficult and 8 challenging environment for the helpdesk. We 9 were -- I think I mentioned this in my witness 10 statement -- we were resourcing in line with the 11 Horizon rollout and we were trying to resource 12 to a central location, so that meant that, 13 whereas the previous regional helplines would 14 have had -- typically, the route into the 15 previous regional helplines was through the 16 Crown Network, so you kind of worked in branch 17 for a number of years and then you went onto the 18 regional helpline. We were recruiting and 19 resourcing much more of a broad mix of 20 individuals, so there was a steeper learning 21 curve for some individuals, if you'd not had 22 that branch experience. 23 And we were obviously dealing with new 24 contractual relationships with Fujitsu around 25 the HSD and the scope of that helpdesk. So 25 1 I think it was -- we were going on the learning 2 journey to some extent with the branches, as 3 well, you know, it was tough. 4 Q. You've also referred in your witness statement 5 to booklets and guides and you say that they 6 didn't provide branches with adequate tools, if 7 they followed the steps but those steps led to 8 unexpected results; do you recall that? 9 A. Yeah, I mean, there was a couple of different 10 communications and guides that the branches had 11 access to. I think the one I particularly 12 remember was the Horizon System User Guide which 13 had a load of flowcharts in of various kind of 14 processes that the branch was supposed to 15 follow, and the feedback -- at least the 16 branches that I spoke to -- about those guides 17 was that they just preferred to speak to 18 somebody on the helpdesk and they found them 19 very difficult to navigate. 20 I think the Quick Reference Guide -- I don't 21 remember when that came in but there was a kind 22 of abridged version that kind of gave you the 23 basic steps, particularly around balancing -- 24 was much more favourably received by branches 25 but certainly, yeah, I would say that the guides 26 1 were in a lot of ways quite complex to follow 2 for branches and the preference was to pick up 3 the phone and speak to the helpdesk. 4 Q. But one thing you specifically highlighted in 5 your witness statement relates to following the 6 steps that are prescribed in a guide but even 7 following those steps leading to an unexpected 8 result; do you remember that? 9 A. Yes. 10 Q. Was that a common scenario? 11 A. Yeah, I think it is -- it was basically because 12 when you came to a position where you were 13 trying to balance and perhaps you had 14 a discrepancy, a loss or a gain, it was not 15 always easy then to backtrack and find out where 16 that had arisen. And, as I say, the sort of 17 branches at least that I spoke to then preferred 18 to just get on the phone and rather than going 19 through some steps, which seemed to confuse 20 them, they would just get on the phone and speak 21 to the helpdesk instead. 22 Q. Moving on to Horizon Online and help with that, 23 I think you've said that it has its own Help 24 system. Can you tell us briefly about that? 25 A. Yeah. So prior to Horizon Online Help, if my 27 1 recollection is correct, there were guides in 2 the branches, counter operations manuals. So 3 these were physical instruction guides, 4 operational guidance. Horizon Online Help sort 5 of digitised those and introduced a help 6 mechanism that was available from the Horizon 7 counter, so you could press a button on Horizon 8 if you were stuck with something, and look up 9 the process. 10 Q. It's something called Online Help, I think, is 11 that the Horizon Online Help? 12 A. Yes. 13 Q. When was that introduced, was that from the 14 beginning of Horizon Online or later? 15 A. I believe it was from the beginning in 2010. 16 Q. You've said in your statement that, although 17 that was a step forward, it hasn't evolved. Can 18 you help us with that, please? 19 A. Yes. So this was mainly in sort of reflecting 20 upon a later project that I was involved in in 21 2016, which was to look at improvements to that 22 very Help system and, although it was a step 23 forward in terms of digitising content, it meant 24 that you didn't have to have people manually 25 updating operation guides in the branch. 28 1 I don't think that's been taken forward in 2 the sense of providing easy and quick access to 3 Help. So the help's there, the content is there 4 on Horizon but it's about the speed at which 5 a postmaster or branch can access it, 6 particularly if they have a customer in front of 7 them, and the fact that you kind of have to step 8 out of the transaction that you're doing to kind 9 of access the Help, rather than there being 10 a kind of overlay of that Help that pops up in 11 the corner. 12 Q. Is that still an issue? 13 A. It is at the moment, yes. 14 Q. Who currently is responsible for that? 15 A. I believe it is being looked at as part of the 16 replacement for Horizon. 17 Q. Who in particular is responsible for that? 18 A. The programme that is looking at the replacement 19 for Horizon. 20 Q. Is there a particular individual in charge of 21 that programme? 22 A. Well, it would sit under Zdravko(?) ** in Post 23 Office terms. 24 Q. Thank you. Can we look at POL00039359, please. 25 It's the second page of that. It seems that in 29 1 2017 there were efforts to simplify balancing. 2 Can you tell us briefly about that, please? 3 A. Yes, so -- 4 Q. It's the second page of that. Thank you. 5 A. Yeah. 6 Q. That's an email from yourself. Who is it to and 7 what's the issue there? 8 A. So there was a programme within Post Office at 9 this time called STRN, which was, I think, 10 Simpler To Run Network, and part of was to look 11 at were there ways that we could simplify the 12 balancing such that it was quicker and easier 13 for postmasters. 14 So I was involved in the sort of very early 15 scoping of this programme, gathering feedback 16 from stakeholders, and I think we investigated 17 three areas, which were suspense account, ATM 18 balancing and also the general sort of balancing 19 steps. And this email is to Alison Clark, who 20 worked, I think at the time, as an NBSC team 21 leader, and I was trying to gather 22 information -- we were trying to gather all the 23 collateral together, essentially, all the 24 documents we had on the balancing process. And 25 this email was to gather that information from 30 1 the NBSC. 2 Q. Is there, within the Post Office around this 3 period, a general recognition that balancing was 4 too complicated and that there were issues with 5 balancing? 6 A. At least in terms of my exposure to it, that was 7 the kind of premise behind looking at balancing 8 simplification. 9 Q. Was there any reflection within the Post Office 10 earlier than 2017 that that was something that 11 should be simplified? 12 A. Not in the roles that I was involved in. 13 Q. Do you perceive that there has been a change in 14 attitude within the Post Office, with regards to 15 balancing and simplifying balancing? 16 A. I think there has been a shift but I don't think 17 it's sort of -- it's not led to material changes 18 in Horizon Online at present. 19 Q. I was going to say, as a result of this email, 20 have you seen any material, significant material 21 changes in simplifying the balancing process? 22 A. Not the balancing process itself, not that 23 I recall. I think some things have been taken 24 forward so, for example, the ATM balancing 25 process I think has had two stages of 31 1 improvement since this document was written, 2 this email was written. 3 Q. If we turn to page 1 of this document and it's 4 the bottom of page 1, there's also reference to 5 a "discrepancy diagnostics document". So it's 6 the bottom of that page. Thank you. Is this 7 something you're aware of? 8 A. Not to my recollection, no. 9 Q. No. What was EUHSP? 10 A. That was the -- is that the Help -- I think if 11 you can maybe just help me with a document. 12 Q. Thank you. You can look in you witness 13 statement, if you like, at the beginning you've 14 provided, it says there "Enhanced User Help and 15 Support Programme"? 16 A. Okay, yeah. 17 Q. Is that something you're familiar with? 18 A. Yeah, so that was part of the BTTP programme 19 which was looking at enhancements to Horizon in 20 sort of 2016. 21 Q. Did that go ahead? 22 A. The Help specifically, the Help and Support? 23 Q. Yes. 24 A. No, it didn't. 25 Q. Why not? 32 1 A. So part of the process for all those initiatives 2 was to present to a senior user forum of what 3 the costs and benefits of that particular change 4 would be, in this case improvements to Help, and 5 it didn't get through that governance forum. It 6 wasn't signed off by that governance forum. 7 Q. Who is in charge of that governance forum, do 8 you know? 9 A. I believe at the time it was Gill Tait. 10 Q. From the look of the emails that I took you to 11 and reference to EUHSP, et cetera, it looks as 12 though there was some thinking, in 2016, 2017, 13 about how to provide more help to 14 subpostmasters. Are you aware of any concrete 15 significant projects that have gone ahead along 16 those lines? 17 A. I don't know whether -- I think it was to 18 provide quicker help to postmasters, not 19 necessarily more. I think, as I said, the 20 content is there. I think it's just about the 21 way that they access it. In terms of things 22 that have changed that help postmasters, there 23 has been some changes in more recent times. For 24 example, branch hub, which is a kind of digital 25 hub that postmasters can access, has access to 33 1 resources, knowledge articles, videos and the 2 like, downloadable guides, which are also 3 available on the learning management system that 4 we maintain. But, in terms of Help itself, the 5 Horizon Online Help, I'm not aware of any 6 significant changes to that content. 7 Q. I want to move on to the Known Error Log. 8 You've addressed the Known Error Log in your 9 witness statement. 10 A. Mm-hm. 11 Q. I think you said that your recollection is hazy. 12 Can you tell us what your recollection is of the 13 sharing of a Known Error Log between Fujitsu and 14 the Post Office during your involvement in those 15 matters? 16 A. Yeah, I don't think I can really go beyond what 17 I've said in the witness statement, which is 18 that there was a lot of activity at that time to 19 try to improve the working relationship, 20 operational relationship between NBSC and HSD to 21 prevent, for example, calls being bounced 22 between those two helpdesks. And my 23 recollection is that, as part of that, there was 24 some information sharing between the two 25 helpdesks of issues that the NBSC were aware of 34 1 and issues that Fujitsu were aware of, in terms 2 of HSD, and the Known Error Log was part of that 3 information share. 4 I believe it was a spreadsheet. As I've 5 said in my witness statement, I don't know 6 whether that was a full set of error logs or 7 whether it was a subset but I do recall seeing 8 a spreadsheet when I was working on NBSC. 9 Q. So this is '99 to 2001 or thereabouts? 10 A. No, I'd say it was later than that. I think in 11 my witness statement -- so it probably -- 12 I mean, this is speculation because I don't 13 remember specifically but it feels like it was 14 in the later period, when I was -- was at 15 Network? So this would have been 2003, I think. 16 Around about that time. 17 Q. So 2003 onwards. How long were you in that 18 role? You were in that role until 2005, so 2003 19 to 2005 or did it extend that, to the best of 20 your recollection? 21 A. I'm not sure, given I left the role, I don't 22 know. And I don't know how frequently it was 23 shared. 24 Q. At the end of that period, there's something 25 called the Callendar Square bug and I'm going to 35 1 move on now to the Callendar Square bug. That 2 first presented in October or September 2005? 3 A. Yeah. 4 Q. Do you recall how you first found out about the 5 Callendar Square bug? 6 A. From the service manager who'd been speaking to 7 the branch. 8 Q. Who was the service manager? 9 A. Sandra MacKay. 10 Q. Can we look at FUJ00083812. When you say Sandra 11 MacKay was the service manager, what do you mean 12 by "service manager"? Can you tell us a little 13 bit about that role, please? 14 A. Yeah, I think I explained earlier that in the 15 early -- sort of 2005 or thereabouts, the Retail 16 Line did split, so into Sales and Service, that 17 the sales managers would deal with generating 18 sales and coaching on sales in the branch. 19 Anything else that wasn't sales was referred to 20 an area intervention office and they had various 21 service managers who worked out of that area 22 intervention office and Sandra MacKay was one of 23 those. 24 So this was dealing with any issues that 25 were not directly related to sales and they 36 1 would contact branches or they would go out and 2 visit branches to investigate and resolve those 3 issues. 4 Q. Were they in charge of particular regions or was 5 it all centralised? 6 A. The area intervention offices themselves were in 7 designated regions but it could be any of the 8 service managers that went out to a particular 9 branch for that area intervention office. 10 Q. Do you remember Sandra MacKay being in charge of 11 any particular region? 12 A. I don't. 13 Q. There's mention there of Callendar Square, 14 7 October 2005. Can you tell us what is this 15 log because we're going to see a few of them? 16 A. This is a standard log that was completed by 17 service managers for any interactions with 18 branches. That might be a visit, it might be 19 a call. 20 Q. What would happen to those logs? 21 A. They would get stored against the branch record 22 on the electronic filing cabinet, I believe it 23 was called. 24 Q. Can we go to the second page. I'm going to read 25 to you a little bit from that entry. It says: 37 1 "Expand on any letter requested ..." 2 So this is the entry on the log. It says: 3 "Telephoned Alan [I think Alan is the 4 subpostmaster] as requested. He is concerned 5 that he has still not heard anything regarding 6 the loss that he is rolling. I told him that 7 I had now involved the C&SM ..." 8 Who was the C&SM? 9 A. I don't remember what the job title -- it may be 10 contracts and service manager. I think that 11 would be Sandra's boss. 12 Q. Thank you: 13 "... who in turn has contacted Andy." 14 Who is Andy? 15 A. I think Andy Bayfield. 16 Q. Thank you. And what was his role? 17 A. I think he was the sort of regional service 18 manager, so this -- he would have been the 19 C&SM's line manager. 20 Q. "I agreed that I would make some enquiries and 21 let him know my findings. I discussed this with 22 Andy who has agreed to send another email 23 relating to the shortfall due to the Horizon 24 failure to Shaun Turner ..." 25 Why are you being mentioned there? 38 1 A. At that time I was in the network co-ordination 2 role. So if there were issues, not necessarily 3 related just to issues with Horizon but any sort 4 of issues that required the input of other 5 stakeholders within the business outside of 6 Service, it would be referred to myself or one 7 of my colleagues to kind of chase up or 8 escalate. 9 Q. Chase up or escalate with who? 10 A. In this particular case, it would typically 11 be -- well, it depends. If we could resolve the 12 issue through Fujitsu taking ownership and 13 fixing the problem, it would be that route. If 14 it was -- if we felt it was a wider or more 15 significant problem, then we would raise it 16 through to the Problem Management Team. 17 Q. To who, sorry? 18 A. To the Problem Management Team. 19 Q. Who was in charge of the Problem Management 20 Team? 21 A. Not sure at this point. I think it was later 22 Dave Hulbert. But I don't know at this 23 particular point in time. 24 Q. After your name is mentioned there, it says: 25 "... meanwhile the office should continue to 39 1 roll the loss." 2 Can you explain to us what "roll the loss" 3 means? 4 A. My take on this is that it means not bringing 5 the loss to account while it is investigated. 6 Q. Was that common advice that the NBSC or others 7 gave to subpostmasters if there was a problem? 8 A. I don't know about the NBSC, but it -- I don't 9 think it would be uncommon for a service manager 10 to advise that while something was being 11 investigated. 12 Q. Typically, how long would somebody roll the loss 13 for? 14 A. I don't think there was some specific period of 15 time. It would depend on how long the 16 investigation takes. 17 Q. If the investigation takes quite a while, might 18 it be that a subpostmaster is asked to roll the 19 loss for quite a significant period of time? 20 A. It could be. 21 Q. Can we look at FUJ00083815, please. Sorry, 22 that's the same document. Can we look at 23 FUJ00083664. 24 Now, this is a significant email chain. I'm 25 going to start at the back so I'm going to start 40 1 at page 6 of that chain. Thank you very much. 2 So we're here on 11 January 2006. It's an email 3 from Sandra MacKay to yourself. What was your 4 position at this time? 5 A. I was still in the network co-ordination role. 6 Q. Thank you. I'm just going to read that email. 7 It says: 8 "You may recall that in September the above 9 office had major problems with their Horizon 10 System relating to transfers between stock 11 units." 12 So September, that's the reference, really, 13 to the document we've seen just now, isn't it? 14 A. Yes, it is. 15 Q. We're now in January: 16 "The [subpostmaster] has reported that he is 17 again experiencing problems with transfers, 18 ([5 January 2006]) which resulted in a loss of 19 around [£43,000] which has subsequently 20 rectified itself." 21 £43,000, would that have struck you as 22 a significant sum or was that typical? 23 A. An alarming sum, I would say, yeah. It would 24 have struck me as significant. 25 Q. "I know that the [subpostmaster] has reported 41 1 this to Horizon Support ..." 2 Horizon Support, who was that? Was that 3 Fujitsu or was that yourselves? 4 A. No, I take that to mean HSD. 5 Q. Thank you: 6 "... who have come back to him stating that 7 they cannot find any problem. 8 "Clearly the [subpostmaster] is concerned as 9 we have just dispute number of months trying to 10 sort out the first instance and he doesn't want 11 a repeat performance. He is convinced that 12 there is something wrong with his Horizon kit. 13 I would be grateful if you could investigate 14 this and give him any support that you can. I'm 15 due to visit the office tomorrow to have a look 16 at his paperwork and discuss the situation with 17 him." 18 Why were you, in particular, being contacted 19 about this problem? 20 A. Firstly, because it had happened before and it 21 appeared to be a system issue and, if it was 22 a system issue, it was important to get that -- 23 give that visibility to the Problem Management 24 Team. 25 Q. When you say a "system issue", do you mean 42 1 a wider problem than an individual branch, or? 2 A. Not necessarily. But a system issue -- an issue 3 with the Horizon System not functioning as it 4 should. 5 Q. Rather than, say, a hardware problem? 6 A. Yes. 7 Q. Could we scroll up and look at the email above. 8 This is an email from Brian Trotter to yourself. 9 Who was Brian Trotter? 10 A. I believe at this point he was Sandra's boss so 11 the C&SM who was referred to earlier. 12 Q. He says: 13 "Further to Sandra's email, I visited the 14 branch with Sandra last week and the 15 [subpostmaster] provided clear documented 16 evidence that something very wrong is occurring 17 with some of the processors when carrying out 18 transfers between stock units. To be absolutely 19 sure from our side can we either carry out 20 a thorough check of the alleged faulty 21 processors or swap them out." 22 Can we turn to the email above that, please, 23 so that's page 5. We have here your response -- 24 sorry, your email to Gary Blackburn. 25 A. Yeah. 43 1 Q. I think you've mentioned Gary Blackburn. Who 2 was he? 3 A. Gary Blackburn was in the Problem Management 4 Team in Post Office. 5 Q. You're forwarding the issue to him and you say 6 there: 7 "Gary, 8 "Need your advice on this branch. There 9 appears to be an ongoing problem at this branch 10 with transfers between SUs causing a receipts 11 and payments mismatch. This first came to my 12 attention some 3 or 4 months ago, when the 13 branch was chasing up an error notice to account 14 for loss [et cetera]." 15 Can we look at the final paragraph there, it 16 says: 17 "Since then it appeared to have happened 18 again, although Fujitsu are saying no issue 19 could be detected. I am concerned that there is 20 a fundamental flaw with the branch's 21 configuration, and would be interested to know 22 how FS ..." 23 That's Fujitsu Services, is it? 24 A. That's correct, yes. 25 Q. "... put the first issue to bed." 44 1 So what did you believe you were doing in 2 that email? Were you escalating the matter, 3 raising it with somebody who would speak to 4 Fujitsu, or what was the purpose? 5 A. I was trying to firstly understand what was 6 happening in the branch and I thought Gary would 7 be able to assist me with that because he had 8 contacts in Fujitsu that I didn't have access 9 to. 10 Secondly, I'm trying to get to a position 11 where we can fix this for the branch, given it 12 appears to have happened in September and then 13 repeated in January. 14 Q. Can we scroll up, please, to page 4. There is 15 an email from Liz Evans-Jones to Gary Blackburn. 16 Now, Liz Evans-Jones seems to be from Fujitsu; 17 is that somebody you were familiar with? 18 A. Only in passing but, yeah, I believe she was 19 a Fujitsu problem manager, so Gary's 20 counterpart. 21 Q. So she would be a direct person to speak to in 22 relation to a bug in Horizon? 23 A. For Gary, yeah. 24 Q. For Gary. She says there: 25 "Hi Gary, 45 1 "I have checked the call and this issue is 2 scheduled to be resolved in S90. 3 "S90 has already been deployed to the 4 Datacentre and counter release is scheduled to 5 start on [4 March 2006] due for completion 6 [14 April 2006]." 7 Now, that response is sent to you. Why were 8 you being kept in the loop in relation to this 9 issue? 10 A. So that I could keep the service manager in the 11 loop and therefore the branch. 12 Q. Did you at this stage think that it might be 13 a wider problem than just this one branch? 14 A. Not at this stage, I had no reason to suspect 15 that. 16 Q. Can we scroll up to page 3, please. So we have 17 there: 18 "Shaun 19 "S90 fix for this problem, in the interim TC 20 correction will have to continue. Let me know 21 if you need any further assistance." 22 Then we have an email from yourself to Gary 23 Blackburn, saying: 24 "Gary, 25 "Thanks for looking into this ... Couple of 46 1 questions occur: 2 "Do we understand why this particular branch 3 has been having problems? Or are there any 4 branches in the network that have been having 5 this problem? 6 "Can the branch be front ended on the 7 counter release of S90 such that it gets a fix 8 as soon as possible? 9 "The email from Liz suggests that there may 10 be a recurrence following S90. What degree of 11 certainty do we have that it will definitely be 12 fixed?" 13 I mean, you've said that the amount, the 14 £43,000, is an alarming amount of money. 15 A. Mm. 16 Q. Were you concerned at this stage about 17 recurrences? 18 A. I was concerned to make sure that the S90 19 release was going to fix it and I was concerned 20 to make sure, as indicated by my question there, 21 that this was, as I believed it to be, a single 22 branch that was having this problem. 23 Q. Can we look at POL00081928, please. It's page 6 24 that I'd like to begin. Thank you. This was 25 an email from Anne Chambers. Who was Anne 47 1 Chambers? 2 A. I believe that she worked within Fujitsu on the 3 second or third line support. 4 Q. And Mike Stewart? 5 A. I don't know who Mike was. 6 Q. This is 23 February 2006, and this email isn't 7 copied to you originally but it is subsequently 8 forwarded to you; have you seen that? 9 A. I have. 10 Q. I'm going to read to you briefly from Anne 11 Chambers' email. If we could scroll down 12 slightly, it says: 13 "Haven't looked at the recent evidence, but 14 I know in the past this site had hit this 15 Riposte lock problem 2 or 3 times within a few 16 weeks. This problem has been around for years 17 and affects a number of sites most weeks, and 18 finally Escher say they have done something 19 about it. I am interested in whether they 20 really have fixed it which [is] why I left the 21 call open -- to remind me to check over the 22 whole estate once S90 is live -- call me cynical 23 but I do not just accept a 3rd party's word that 24 they have fixed something!" 25 Further down in that email, she says: 48 1 "Please note that KELs tell SMC that they 2 must contact sites and warn them of balancing 3 problems if they notice the event storms caused 4 by the held lock, and advise them to reboot the 5 affected counter before continuing with the 6 balance. Unfortunately in practice it seems to 7 take SMC several hours to notice these storms by 8 which time the damage may have been done." 9 Were you aware of this problem at the time, 10 the significance of this issue? 11 A. I don't recall reading this email. I mean, it 12 was copied to me, so I may have seen it. 13 I certainly didn't pick up on the significance 14 at the time. My main focus was getting the 15 branch or branches that were impacted by this 16 fixed, which was going to be done shortly 17 through the S90 release. 18 Q. Having heard about the Callendar Square incident 19 and the £43,000, having thought that it was 20 likely to have only related to one particular 21 branch, might this not have struck you as quite 22 concerning at the time and something really that 23 should be raised to quite a significant level 24 within the Post Office? 25 A. I regret not being more curious about that, the 49 1 line where it is said that it is affecting 2 several sites a week. But I think at the time, 3 if I'd have read that, I would have assumed that 4 this was being managed through the sort of cross 5 to main problem management space, especially 6 since Gary and Lynne were on the circulation. 7 Q. You had been on the distribution list of this, 8 and you were also copied in or sent emails in 9 the early days. I mean, dating back to 2005, 10 you were involved in the Callendar Square 11 incident. Is there a reason why it didn't 12 strike you at the time as something more 13 significant? 14 A. I think, as I say, my focus was to make sure 15 that the branch issue was fixed. I was told 16 that was being done as part of the S90 release. 17 My accountability, therefore, was to push that 18 forward to the service manager, so that they 19 could keep the branch up to date. In terms of 20 this particular aspect, I would have expected 21 that to be being, you know, the wider sort of 22 branch impact, if there were other branches that 23 were being impacted by this error, to be managed 24 and monitored through the problem management 25 process. 50 1 Q. It says at the bottom there: 2 "Please note that KELs tell SMC that they 3 must contact sites and warn them of balancing 4 problems if they notice the event storms caused 5 by the held lock ..." 6 Do you know how event storms would be 7 noticed? 8 A. I don't. That's the other thing about this 9 email. There's some technical content that was 10 beyond my ken. 11 Q. Do you know what an event storm is? 12 A. I don't. 13 Q. How, at this time, did you believe a bug such as 14 this would be known to the wider Post Office 15 community, whether it's branches or within Post 16 Office itself? 17 A. At this time, my assumption was there were two 18 things that would -- two processes that would 19 surface a bug like this. One would be receipts 20 and payments mismatch messages in the branch and 21 the other would be Fujitsu monitoring. 22 Q. Was your belief that Fujitsu monitoring was 23 infallible? 24 A. Not infallible but I thought they had robust 25 monitoring in place and if a branch contacted 51 1 them, they would seek to replicate that, and if 2 it wasn't included in their existing monitoring 3 processes, that it would be added. 4 Q. If a branch didn't contact them, because they 5 didn't know about the problem or didn't connect 6 the dots, how would they have found out about 7 the problem? 8 A. Yeah, that's a fair point. I don't know. If 9 there's no monitoring in place and the branch 10 doesn't contact them, it seems to me that they 11 wouldn't know about it. 12 Q. If we scroll up to page 5, we have the email to 13 yourself from Gary Blackburn. It's the top 14 email on page 5, please -- thank you -- sent to 15 you on 1 March 2006: 16 "Shaun 17 "It appears that Callendar Square is not 18 alone with its mismatch problem. It also 19 appears that Fujitsu are expecting S90 release 20 to resolve this quirk. 21 "We have opened a cross-domain problem 22 record. Lynne Fallowfield is the contact." 23 Who was Lynne Fallowfield? 24 A. She was a colleague of Gary Blackburn's in 25 Problem Management. 52 1 Q. Thank you. So it's not just a matter of Anne 2 Chambers' email being forwarded to you amongst 3 a large distribution list, or anything along 4 those lines. I mean, it is being sent to you as 5 the sole recipient by Gary Blackburn, telling 6 you here that Callendar Square isn't alone with 7 the mismatch problem. Again, would it not have 8 struck you as significant? 9 A. As I say, the key thing for me was getting the 10 branch or branches that were impacted, the 11 situation fixed, and I was being told that that 12 would be implemented with the S90 release and, 13 as I previously mentioned, my feeling at this 14 time was that Fujitsu had adequate monitoring in 15 place to identify these branches. 16 Q. But weren't you also being told by Fujitsu, that 17 is Anne Chambers' email, that she's a little 18 sceptical of the fix? 19 A. Yes, yeah. I would expect post-S90 monitoring 20 to be in place to prove that the fix had worked. 21 That was normally part of the standard problem 22 management process, in my experience. 23 Q. Was it your experience that somebody from 24 Fujitsu would put in an email that they were 25 sceptical about their own fix? 53 1 A. I'd not seen that before. I don't think I ever 2 saw it again. 3 Q. Can we look at page 4, please. There's an email 4 from you to Sandra MacKay and to Brian Trotter, 5 and you say there: 6 "As you can see from the email below though, 7 there is now recognition that this is a wider 8 issue than just a software 'quirk' at just one 9 branch, which means it is now being actively 10 managed as a cross-domain problem within 11 Fujitsu." 12 Why were you sending the email to them? 13 A. Predominantly to -- so that they could update 14 Callendar Square. 15 Q. So we know that this problem has been going on 16 at Callendar Square since September 2005. We 17 know that the software release S90 is going to 18 take place in April 2006. But we know that 19 there are concerns about it still existing after 20 the software fix. We know also now that it 21 affects more than one branch. There seems to be 22 in this chain a bit of a lack of urgency with 23 regards to this particular problem; would you 24 agree with that? 25 A. I think things could have been done quicker, 54 1 yeah. 2 Q. Were, as far as you're aware, messages being 3 cascaded across the network to branches about 4 a problem that affects balancing? 5 A. Not to my recollection there weren't, no. 6 Q. Who was responsible for passing the information 7 to branches that didn't call the helpline? 8 A. Once something like this was accepted as 9 a problem, the Problem Management Team would 10 consider what communications were required both 11 to internal stakeholders and to branches, so 12 I would see it as their accountability to make 13 that judgement. 14 Q. We're talking about a problem that's gone on for 15 years and that, as you have said -- I mean, in 16 the Callendar Square case, £43,000, significant 17 sum. Who do you think, in Post Office, was 18 responsible for making that message more widely 19 known? 20 A. As I say, I think that would be the Problem 21 Management Team. 22 Q. So that's a team. Who in particular? 23 A. I'm not sure who was heading up that area at 24 this time, but whoever is managing that 25 problem -- in this case Lynne Fallowfield in 55 1 consultation with her seniors -- would need to 2 make that judgement based on the number of 3 impacted branches and the likely fix time. 4 Q. Did you ever have a conversation with her about 5 it? 6 A. Not to my recollection, no. 7 Q. Who was responsible for telling auditors about 8 this problem? 9 A. Again, I would see that coming through the 10 Problem Management Team. They should be 11 considering internal stakeholders and branch 12 communications. 13 Q. Who was responsible for telling investigators 14 and prosecutors about this problem? 15 A. The same. 16 Q. Were there policies and procedures in place for 17 them to have raised the issue with auditors, 18 investigators, prosecutors? 19 A. I was never part of the Problem Management Team 20 but I believe there were processes that they 21 could follow to determine who they should be 22 communicating to. That was my understanding as 23 an outsider to that team. 24 Q. As someone who had spent a considerable part of 25 your career analysing data and analysing trends, 56 1 was the issue that was being raised here -- so 2 more than one branch, number of years -- was 3 that something that you would have considered to 4 be significant? 5 A. Yeah, I think in the moment, yeah, it's 6 significant, which is why I was asking those 7 questions. But I think I would have seen it -- 8 I think from the documentation that we saw 9 earlier, the email was suggesting that the S90 10 release was imminent in March 2006. So that 11 would have been my focus at this time, to get 12 the branch situation fixed. 13 Q. So I think the software fix wouldn't be complete 14 until April 2006. It's been going on, at least 15 from the Callendar Square incident, since 16 September 2005. Did you consider the historic 17 position, looking back, even if it is soon to be 18 fixed, albeit in a couple of months' time, what 19 about all those branches that were affected over 20 the course of that year? 21 A. As I say, I would have expected that to come 22 through Fujitsu monitoring and be raised. If 23 there were incidents earlier than Callendar 24 Square, I would have expected that to have been 25 monitored by Fujitsu and raised as a problem by 57 1 them earlier. I had no reason at that stage to 2 believe that wasn't happening. 3 Q. Raised with who? 4 A. Raised -- so the Fujitsu Problem Management Team 5 should raise it to the POL Problem Management 6 Team, across the main problem. 7 Q. Can we look at POL00070134. We're now on 8 6 December 2006 and there is an email, the 9 subject of which is "Callendar Square URGENT". 10 You are a recipient of that email, from Mandy 11 Talbot, Dispute Resolution. She says there: 12 "In [February] of this year you wrote to 13 Gary Blackburn and he wrote to Shaun Turner and 14 then Sandra MacKay about these branches which 15 had apparently registered complaints about the 16 HORIZON system. Fujitsu have told us that in 17 respect of Callendar Square that there was 18 a problem when stock was transferred from one 19 stock unit to another but this would any apply 20 when there was more than one stock unit, ie more 21 than one position at the counter." 22 So it seems as though, even in December 23 2006, you were still being sent emails relating 24 to the Callendar Square. Do you remember that 25 at all? 58 1 A. Yeah, I don't remember receiving this particular 2 email prior to getting the pack, the bundle. 3 But, yeah, I can see that I was sent the email 4 and I think Lynne responded with some details of 5 those four branches. 6 Q. Can you think of why, in December 2006, you 7 would still be being sent emails related to 8 Callendar Square? 9 A. I think it is because there was previous 10 correspondence between Mandy Talbot and Lesley 11 Joyce where the visit logs were sent, and 12 I presume from the content of this email that 13 Mandy Talbot may have seen the emails that were 14 sent back and forth at that time and that's 15 why -- my name is on those, so that's why she's 16 sort of send out a sort of blanket email. 17 Q. Was the Callendar Square bug still something 18 that was bubbling away in December 2006? 19 A. Not in my world it wasn't, no. 20 Q. Do you believe it was or wasn't? 21 A. I believed it had been fixed by S90. 22 Q. Did you check that it had been fixed by S90? 23 A. I don't recall checking. That's not to say it 24 didn't happen. I would have expected, as 25 a matter of course, where I'd raised a problem 59 1 with the Problem Management Team to be notified 2 that it had been successfully resolved. 3 Q. So you would have expected the Problem 4 Management Team to have kept a tab on whether 5 S90 had fixed the problem or not? 6 A. Yeah, once they'd got it across the main problem 7 record, that's the purpose, is to manage that 8 through and ensure that it's resolved or 9 mitigated. 10 Q. Given the significance, the sums involved, the 11 length of time that it was happening, are you 12 aware and did you keep tabs on whether any 13 significant investigation had taken place into 14 whether S90 fixed it? 15 A. I don't recall doing that but that's not to say 16 that I wasn't speaking to Gary and Lynne about 17 this post the implementation of S90. 18 Q. Having been quite a significant issue at the 19 time, significant in number, significant in 20 effect, is there a reason why you can't remember 21 whether you did or didn't? 22 A. I don't know what to say to that. I can't 23 remember. So ... 24 Q. Was it not significant for you? 25 A. No, it was. It was significant. But I don't 60 1 remember -- I believe subsequently, from the GLO 2 findings, that it was fixed in S90. That's my 3 understanding. But I don't remember speaking to 4 Problem Management -- I would have expected -- 5 that would have been my expected process at that 6 time for something like this, and I can only 7 assume that, if I had spoken to them, I would 8 have been told that it had been fixed in S90. 9 Q. You had previously been responsible as network 10 conformance and capacity manager and looking 11 into things like error rates and conformance. 12 Had you come across similar issues to this one 13 in that role? 14 A. No, not really, because when we're talking about 15 error rates in that role, it's not branch 16 balancing issues; it's more presentational 17 issues or it may be not correctly sizing 18 parcels, so it's not specifically related to the 19 balancing process. 20 Q. How about as network co-ordination adviser? 21 Were you coming across similar issues to the 22 Callendar Square issue? 23 A. No, not typically, no. There were escalations 24 to me about Horizon Issues but, typically, it 25 was more things like engineers not being sent 61 1 out on time, where I needed to chase them up. 2 This was fairly unique. 3 MR BLAKE: Thank you. 4 Sir, it may be an appropriate time now to 5 take our mid-morning break. 6 SIR WYN WILLIAMS: By all means, that's fine. 7 MR BLAKE: Thank you very much, could we take 8 15 minutes, which -- 9 SIR WYN WILLIAMS: Yes, 11.35? 10 MR BLAKE: Thank you very much. 11 SIR WYN WILLIAMS: Thank you. 12 (11.20 am) 13 (A short break) 14 (11.38 am) 15 MR BLAKE: Thank you, Chair. Can we bring up 16 POL00070133, please. Now, at the bottom of this 17 page, there is that email I took you to just 18 before the break, but then that email is 19 forwarded by Mandy Talbot to somebody called 20 Steve Parker, do you know who Steve Parker is, 21 was? 22 A. Other than he worked in Fujitsu, no. 23 Q. Yes. As you say, he worked for Fujitsu, and it 24 says that he's been copied into an email 25 because: 62 1 "... it may be that you might have to do 2 a repeat performance tomorrow once the FAD codes 3 had been identified and the name of the branches 4 revealed", et cetera. 5 The second paragraph is the one I want to 6 draw your attention to. It says: 7 "Stephen and Richard our legal team at the 8 Court will be doing their best to persuade the 9 Court not to allow Castleton ..." 10 That refers to Lee Castleton. Are you aware 11 of a Lee Castleton case? 12 A. Any in passing, yes, I believe it was Marine 13 Drive Post Office, I think. But not at this 14 stage, I wouldn't have been aware. 15 Q. So on 6 December 2006, when you're -- the email 16 that was sent to you is being forwarded, you 17 weren't involved in the Lee Castleton case in 18 any way? 19 A. No. 20 Q. Thank you. I'll just continue that email. It 21 says: 22 "... to call this evidence because it is 23 failed late and does not relate to the problems 24 at his branch office. If they are successful 25 there will be no need to progress any further 63 1 with these investigations but as Castleton is 2 a litigant in person it is common for Judges to 3 be sympathetic and may allow him to rely on his 4 evidence. If so you will have to pull out all 5 the stops to investigate what if anything went 6 wrong at these branches and why we can 7 distinguish them from Mr Castleton at Marine 8 Drive." 9 So it seems as though the Callendar Square 10 case is being used potentially as an example of 11 an issue with Horizon in the Castleton 12 proceedings; is that a fair interpretation or 13 your interpretation of that email? 14 A. That's a fair interpretation, I would say. 15 Q. Yes. Were you asked about the Callendar Square 16 bug in December 2006 by anybody? 17 A. Other than the email we've seen from Mandy 18 Talbot, no. 19 Q. Did you know about the context in which that 20 email had been sent, relating to the Lee 21 Castleton case? 22 A. Other than what is the content of the email, 23 that's all I would have known at that stage. 24 Q. So you wouldn't have had a conversation with 25 Mandy Talbot or anybody else on that bottom 64 1 email about why Callendar Square is being raised 2 now? 3 A. No. 4 Q. What was the relationship like between those who 5 worked in NBSC and those who were involved in 6 court cases, be they investigators or 7 prosecutors or the legal team? 8 A. In my experience, rather distant. I don't 9 recall any specific interactions with 10 investigations. They were a little bit further 11 down the line than where you'd be involved with 12 NBSC in dealing with branches. 13 Q. Would those in the NBSC ever be asked about 14 common problems with Horizon or bugs, errors or 15 defects? 16 A. Not to my recollection. You mean by 17 investigation specifically or technically? 18 Q. We see here, for example, Mandy Talbot and 19 Fujitsu in correspondence about ongoing 20 proceedings and that relates to, or in some way 21 relates to, the Callendar Square incident. Was 22 there ever any discussion, to the best of your 23 recollection, with those kinds of people or 24 anybody else involved in court proceedings and 25 people who worked on the NBSC about bugs, errors 65 1 or defects within Horizon? 2 A. No, not to my recollection, no. 3 Q. I'm going to now ask you about another bug and 4 now we're moving on to Horizon Online. Can we 5 look at POL00034433, please. Before we go on to 6 that particular bug, I just want to ask you, 7 this is a document that concerns Horizon Online 8 migration and it seems to have been written by 9 yourself or it's got your name in the bottom 10 left-hand corner. Is this a document you 11 recollect? 12 A. Only after having seen the bundle. I didn't 13 recollect it before but it does appear that I'm 14 the author of it, yes. 15 Q. In the first paragraph there it says: 16 "The delivery of Horizon Online is a key 17 business strategy in delivering some of the cost 18 savings that underpin bringing the business back 19 into profit by 2011." 20 Do you recall at this stage, so 2010, 2011, 21 there being financial pressures at the Post 22 Office? 23 A. In general terms, yes. But I -- in terms of the 24 business benefits of Horizon Online 25 specifically, I don't recall. 66 1 Q. Do you recall any pressures to get on with 2 Horizon Online, following delays alongside these 3 possible profitability issues? 4 A. Not specifically, no. But then I wasn't 5 connected to the programme directly; I wasn't 6 working on the Horizon Online programme. 7 Q. Can we look at POL00028838, please. This is the 8 issue that I'd like to ask you about. It's a 9 "Receipts/Payments Mismatch issue". We can see 10 that at the top. You're not listed there as 11 an attendee of this meeting, you are 12 mentioned -- I know you're aware of this a bit 13 later on in this document, but let's see where 14 we get to with looking at this page. So it 15 says: 16 "What is the issue? 17 "Discrepancies showing at the Horizon 18 counter disappear when the branch follows 19 certain process steps, but will still show 20 within the back end branch account. This is 21 currently impacting circa 40 Branches since 22 migration on to Horizon Online, with an overall 23 cash value of circa £20,000 loss. This issue 24 will only occur if a branch cancels the 25 completion of the trading period, but within the 67 1 same session continues to roll into a new 2 balance period." 3 Is this something you recall at all at the 4 time? So we're here now in 2010. 5 A. Yeah, this was in the additional documents 6 bundle that I received, and I think I also saw 7 this in some previous evidence -- it may have 8 been by John Simpkins -- and, at that point, 9 I did recall there being some issue that I was 10 advised of in 2010. Specifically, I remember it 11 because it was in the very early days Horizon 12 Online but, beyond that, in terms of the nature 13 and the number of branches and what the specific 14 underlying technical problems were, I don't 15 remember anything. 16 Q. It says below that: 17 "At this time we have not communicated with 18 branches affected and we do not believe they are 19 exploiting this bug intentionally." 20 Now, you're somebody who has spent your 21 whole career at the Post Office. Are you aware 22 of branches not being told about bugs dependent 23 on whether or not they are using it 24 intentionally, exploiting it intentionally. 25 A. Only from my experience of the Callendar Square 68 1 bug. I don't believe there was a wider 2 communication on that to branches. 3 Q. But it seems from the wording here that it's not 4 being communicating because it's not being 5 exploited intentionally and they would only 6 communicate if it was exploiting the bug 7 intentionally. Do you think I've misread that 8 or do you think that is a fair reading of that 9 sentence? 10 A. Yeah, I mean, I think that's a reasonable 11 reading of that sentence, yeah. 12 Q. Do you think that the Post Office was better at 13 communicating with branches, in respect of bugs, 14 errors or defects, if the Post Office felt that 15 such bug, error or defect was being exploited 16 intentionally? 17 A. I think my experience with bugs is -- it's sort 18 of limited that I wouldn't want to make that 19 judgement call more generally. I can only go on 20 where I've been involved. 21 Q. From your experience, was the Post Office better 22 at communicating with branches, where the Post 23 Office had something to lose? 24 A. No, I don't think so. As I say, I've not been 25 involved in communications or these decisions 69 1 and it seems to me you're asking about what 2 underpins certain communication decisions, which 3 I've not been party to. 4 Q. It then goes on to say: 5 "The problem occurs as part of the process 6 when moving discrepancies on the Horizon System 7 into Local Suspense." 8 Can we turn over the page, please. Page 2, 9 it talks about the impact and it has some bullet 10 points at the bottom of that page. The impact 11 is: 12 "The branch has appeared to have balanced, 13 whereas in fact they could have a loss or 14 a gain." 15 So that's one of the fundamental uses that 16 this Inquiry is looking into. It says: 17 "Our accounting systems will be out of sync 18 with what is recorded at the branch. 19 "If widely known could cause a loss of 20 confidence in the Horizon System by branches." 21 Again, is that something you were aware of 22 from your long experience within the Post 23 Office, that there is a concern in the Post 24 Office, certainly at this time, that by 25 spreading the word there would be a loss of 70 1 confidence in the Horizon System. 2 A. That as a general theme, yes, I was aware of, 3 yes, and I would say that is the case in my 4 experience. 5 Q. Can you tell us how you got that impression? 6 A. Largely from the messaging that was coming out 7 from the business, particularly in the sort of 8 post-2009 period, around the robust nature of 9 the Horizon System, which does lead to 10 particular sensitivities around any issues or 11 any perceived issues with the Horizon System. 12 Q. We know in 2009 there was the Computer Weekly 13 article, for example. Was that the driver or 14 was there some other driver? 15 A. From my recollection, I think it was the 16 Computer Weekly and, you know, the early days of 17 the Justice for Subpostmasters, were certainly 18 things that were mentioned in the business and 19 where messaging was coming out to internal staff 20 like myself, around the sort of robust nature of 21 Horizon. 22 Q. Where was that messaging coming from? 23 A. I don't remember specifically but senior 24 leaders. 25 Q. When you, say "senior leaders", do you mean 71 1 Chief Executive level or below that? 2 A. I don't think I can say for certain but, you 3 know, I was a recipient of some of that 4 messaging as it came down the chain and my 5 impression was that that was coming from senior 6 leadership. 7 Q. Can you clarify just so that we know who you 8 mean by senior leadership? 9 A. I guess in my area it would be a couple or three 10 rungs up the ladder from my direct line manager. 11 But I would imagine that messaging was coming 12 from board level down. 13 Q. When you say some way up from your level, can 14 you give a name? 15 A. I don't know, at that stage. I can't remember 16 the structure. 17 Q. Board level, though, again, Chief Executive 18 level? 19 A. That -- this was my impression of where the 20 messaging was coming from. We received it down 21 the management line. 22 Q. But your impression was it was coming from Chief 23 Executive level? 24 A. That was my impression, yeah. It was 25 a significant concern within the business, so my 72 1 impression would be that that would be coming 2 from senior leaders within the business. 3 Q. "Potential impact upon ongoing legal cases where 4 branches are disputing the integrity of Horizon 5 Data." 6 Again, was that something you were aware of? 7 A. Yeah. 8 Q. So you mention the concern arising, for example, 9 from the Computer Weekly article. Did you also 10 get the impression or were you told that there 11 was concern relating to ongoing legal cases? 12 A. Yeah, yeah, certainly. 13 Q. Both of those? 14 A. Yes. 15 Q. "It could provide branches ammunition to blame 16 Horizon for future discrepancies." 17 Again, that's all rolled up with the same 18 issue? 19 A. Yeah. 20 Q. If we go over the page, please: 21 "The Receipts and Payment mismatch will 22 result in an error code being generated which 23 will allow Fujitsu to isolate branches affected, 24 by this problem, although this is not seen by 25 branches, we have asked Fujitsu why it has taken 73 1 so long to react to and escalate an issue which 2 began in May. They will provide feedback in due 3 course." 4 Now, we saw in relation to the earlier bug 5 that there was interaction with Fujitsu. We saw 6 that Anne Chambers email, et cetera. Was that 7 something you experienced with Fujitsu, that 8 they take a long time to react to problems such 9 as this? 10 A. I don't think typically that was my experience 11 but my exposure was very limited. If I did need 12 to have -- or put forward issues for particular 13 branches, it would generally be other people 14 that were having the interaction with Fujitsu 15 rather than myself. 16 Q. So it wasn't something you had direct experience 17 of but were you aware of any complaints from 18 anybody within the Post Office about the time 19 that it took for Fujitsu to react to these kinds 20 of issues? 21 A. I wasn't aware of any particular issues that 22 were flagged to me. 23 Q. If we go down to the "Proposal for affected 24 Branches", we have a number of different 25 solutions and the recommendation is that 74 1 Solution Two should be progressed. Let's just 2 look at those solutions briefly. We have 3 Solution One: 4 "Alter the Horizon Branch figure at the 5 counter to show the discrepancy. Fujitsu would 6 have to manually write an entry value to the 7 local branch account." 8 Were you aware, at that time, that that was 9 something that Fujitsu had the facility to do, 10 to manually write an entry to the local branch 11 account? 12 A. No. 13 Q. It says: 14 "IMPACT -- When the branch comes to complete 15 next Trading Period they would have a 16 discrepancy, which they have to bring to 17 account. 18 "RISK -- this has significant data integrity 19 concerns and could lead to questions of 20 'tampering' with the branch system and could 21 generate questions around how the discrepancy 22 was caused. This solution could have moral 23 implications of Post Office changing branch data 24 without informing the branch." 25 Were you ever aware throughout your time at 75 1 the Post Office of issues relating to manually 2 altering data within the branch? 3 A. Not to my recollection, no. 4 Q. Solution Two, which is the one that they 5 recommended, was: 6 "P&BA will journal values from the 7 discrepancy account into the Customer Account 8 and recover/refund via normal processes. This 9 will need to be supported by an approved POL 10 communication. Unlike the branch 'POLSAP' 11 remains in balance albeit with an account 12 (discrepancies) that should be cleared. 13 "IMPACT -- Post Office will be required to 14 explain the reason for a debt recovery/refund 15 even though there is no discrepancy at the 16 branch. 17 "RISK -- could potentially highlight to 18 branches that Horizon can lose data." 19 Were you aware of a concern within the Post 20 Office of highlighting to branches that Horizon 21 could lose data? 22 A. No. 23 Q. Solution Three: 24 "It is decided not to correct the data in 25 the branches (ie Post Office would prepare to 76 1 write off the 'lost'). 2 "IMPACT -- Post Office must absorb circa 3 £20K loss. 4 "RISK -- huge moral implications to the 5 integrity of the business, as there are agents 6 that were potentially due a cash gain on their 7 system." 8 Were you aware, at this stage, the 9 significance of these huge moral implications, 10 and debates of this nature taking place within 11 the Post Office? 12 A. Only in a more generalised sense about 13 a concern, you know, after the Computer Weekly 14 article about the robustness of Horizon. And, 15 as I say, the messaging on that down the chain, 16 to me at least, was the system is robust, it's 17 audited and there is monitoring in place. But 18 nothing in terms of this specific problem. 19 Q. If we scroll down to the next page, you're aware 20 that your name is mentioned there in an action 21 point summary. 22 A. Yes. 23 Q. It's the fourth one down, it says: 24 "Branch Performance review: Confirm with 25 Shaun Turner any future audits for Branches and 77 1 any performance issues flagged." 2 Do you know why you're mentioned there? 3 A. I can speculate or take an educated guess, which 4 would be, at this time, part of my job role was 5 the production and maintenance of a sort of 6 branch profile, which had the branches for all 7 the network in there, which was used to -- as 8 one of the tools to guide audits. I suspect 9 that this action is to check with me that 10 there's no impact from this particular issue on 11 that profile. 12 Q. In this period, so we're talking about October 13 2010 or thereabouts, we know that, for example, 14 Seema Misra, a subpostmistress, was being 15 prosecuted. Did anyone ever discuss with you 16 these kinds of concerns and the potential impact 17 that they may have on ongoing legal cases? 18 A. No. 19 Q. Were you aware, at this stage, that people were 20 being prosecuted for shortfalls on Horizon? 21 A. I was aware that there was a business process to 22 do so. But, yeah, in terms of direct 23 involvement or interactions with those 24 individuals in the business that were pursuing 25 those cases, I didn't have any contact with 78 1 them. 2 Q. We saw from the Castleton case, for example, 3 earlier on, an email that had been sent to you 4 being forwarded in the context of legal 5 proceedings. Were you aware at this stage that 6 some people were raising faults with Horizon in 7 their defence in legal proceedings? 8 A. In 2010? 9 Q. Yes. 10 A. I was aware, because of the Justice for 11 Subpostmasters, that yeah, that was being 12 raised, yeah. 13 Q. We've seen earlier about your knowledge of the 14 Callendar Square bug. Was this the time, 15 perhaps, to have raised that issue more widely? 16 A. Perhaps but, you know, my feeling on that bug 17 was that that had been fixed and that the impact 18 of it would have been monitored through the two 19 Problem Management Teams. So if there were 20 concerns to be raised to wider stakeholders, 21 I would have expected that to happen through 22 that process. 23 Q. But by 2010 you had known that the Callendar 24 Square bug had gone on for quite a while, 25 affected quite a few branches, a significant 79 1 sum. You're now informed about an issue 2 relating to the receipts and payments mismatch 3 in Horizon Online. The knowledge that you had 4 developed over those years, was that not 5 something that you thought "Actually, this is 6 an issue now"? 7 A. I don't think it's so much that; I think it is 8 that I had a kind of institutional trust in the 9 processes and teams that were managing these 10 issues. 11 Q. Do you think the teams were acting in silos and 12 weren't really communicating with each other in 13 respect of bugs, errors and defects in Horizon? 14 A. Which teams do you mean? 15 Q. You often refer to other teams dealing with 16 issues and it not being within your sphere at 17 that time. Was that caused by any inherent 18 problem with the structure at the Post Office? 19 A. I think -- I mean, I think the reason, in the 20 Callendar Square case, for example, that 21 I wanted to get that in a Problem Management, 22 was that I saw them as the most appropriate 23 people to resolve it. They were the people with 24 the expertise, the people with the contacts. So 25 I don't think there was -- from my perspective 80 1 at least, at this time -- that there was sort of 2 institutional or process failings, as such. In 3 retrospective, I think you can make that 4 argument, certainly. 5 Q. Because there was this team that existed that 6 you could pass over the problem to and it no 7 longer became your problem? 8 A. Well, I don't see it like that; I see it as 9 passing it to the individuals who are most 10 appropriate to resolve that particular issue for 11 the branch. 12 Q. During this period, so 2010, you were the branch 13 standards data analyst. In that role, you made 14 outbound calls, I think, to branches to tackle 15 common issues; is that right? 16 A. Yeah, conformance and compliance issues, yeah. 17 Q. Can you tell us how such calls would be made? 18 What was the process involved? Why would you 19 call branches? 20 A. So we had streams of compliance data coming in 21 to the team. So this would be things like mails 22 data or cash declaration data, and that was 23 analysed by myself and a colleague, and we would 24 identify sets of branches to pull together 25 a call campaign based on their performance. 81 1 So you'd be looking at the branches that 2 were struggling the most with, say, oversized 3 mails, identify those, upload them for the team 4 and then they would make outbound calls to the 5 branches to do some coaching. 6 Q. Was there a process of making outbound calls to 7 branches that may be affected by bugs, errors 8 and defects in Horizon? 9 A. Not in the team that I was in at this stage, it 10 was conformance and compliance. 11 Q. Was there a team that was carrying out, to the 12 best of your knowledge, those kinds of outbound 13 calls? 14 A. If there was, I -- it would be HSD, I would say, 15 and then further down the line, if there was 16 discrepancies, it may well be that P&BA and 17 Chesterfield or the contracts adviser would 18 become involved. 19 Q. But the idea of making these proactive calls, 20 which is what you were doing for a certain type 21 of problem -- 22 A. Yes. 23 Q. -- are you aware of an equivalent of that in 24 relation to bugs, errors and defects with 25 Horizon? 82 1 A. Proactive calls? No, I wasn't aware of 2 anything. 3 Q. How, in your view, would an issue such as this 4 receipts and payments mismatch issue in 2010 be 5 cascaded down to branches, branches that didn't 6 know that they were affect by the bug? 7 A. Well, if a decision was made to communicate it, 8 there was the sort of weekly Counter News, 9 I would imagine, or Branch Focus, I think it may 10 have been named at that stage, or Memo View, 11 through the -- which is a broadcast, messaging 12 broadcast. Those would be typically the 13 channels down to the branch, if we were going to 14 communicate. 15 Q. What would they typically have said about 16 an issue such as this receipts and payments 17 issue? 18 A. I don't know. I wasn't involved in 19 communications so I can't say what, typically, 20 the messages would say. But I would imagine it 21 would outline the nature of the problem and 22 advise branches on how to avoid any sequences of 23 icon -- you know, icon pressing that would 24 result in that particular bug. 25 Q. Taking this one as an example, if that didn't 83 1 happen, would you be concerned by it? 2 A. It depends on the bug and the level of branches 3 impacted and how robust the monitoring is to 4 ensure that we're picking up any other impacted 5 branches. 6 Q. So we have here "Receipts/Payments Mismatch" 7 affecting circa 40 branches, Horizon Online, 8 overall cash value of £20,000 loss. In those 9 circumstances, would you have expected it to 10 have been communicated via those channels that 11 you've just described? 12 A. Um, yeah, as I say, I wasn't involved in the 13 decision-making process around communications 14 but, to me, it feels significant enough that we 15 would want to communicate something. 16 Q. Are you aware of the processes involved in how 17 that communication took place? 18 A. As I say, it would either be through Branch 19 Focus or, if it was more urgent, it would go 20 through the message broadcast service. But in 21 terms of the sort of process for generating that 22 communication, as I say, that would generally 23 come through the problem management process and 24 then to internal stakeholders and the comms team 25 ultimately to distribute. 84 1 Q. I'm going to take you to an even later problem, 2 and that's the smart ID receipts and payments 3 mismatch you've mentioned in your statement. 4 Can we look at POL00043585, please. So this is 5 a very late issue in the context of this 6 Inquiry. We have the date there, 27 November 7 2018. Are you aware of this occurring? 8 A. Yes. 9 Q. Yes. Are you aware that this was during the 10 Group Litigation against the Post Office? 11 A. Yes. 12 Q. Can you summarise for us, looking there at the 13 incident detail, what issue this caused in 14 recording transactions? 15 A. I believe it caused a receipts and payments 16 mismatch in the branch. 17 Q. You're named there, "Shaun Turner, Enhanced User 18 Management Product Owner". Why are you named as 19 having attended that update? 20 A. At this time, I was working on the Enhanced User 21 Management Programme as a product owner, and 22 that role involved taking the business 23 requirements to the technical teams to build the 24 solution. 25 Q. Having been aware of the Callendar Square bug, 85 1 having been aware of the bug in 2010, the 2 receipts and payments mismatch issue that we 3 just looked at, now being told in 2018 about 4 this particular incident, were you concerned 5 about the integrity of Horizon? 6 A. I was concerned about this bug and the impact it 7 might have. As to why the questions about the 8 integrity, I -- 9 Q. Were people at this meeting discussing historic 10 issues, historic bugs? 11 A. No. 12 Q. Did you, at this meeting, mention those two 13 incidents that you had previous knowledge of? 14 A. Not to my recollection, no. 15 Q. There are a large number of people named here as 16 having attended, both from the Post Office and 17 also we see from Fujitsu. We see Steve Bansal, 18 Pete Newsome. Do you remember Fujitsu people 19 attending? 20 A. Yeah, I do remember there were Fujitsu 21 attendees, yeah. 22 Q. Do you know when the Post Office started raising 23 incidents such as this with such a wide group of 24 individuals? Was there a particular moment in 25 time when these issues were escalated in the way 86 1 that this one seems to have been escalated in 2 2018? 3 A. I don't know. This was my first exposure to 4 a kind of forum like this. 5 Q. Did you question why it was taking place? 6 A. Well, I knew why it was taking place. It was to 7 drive the solution for the bug that we'd 8 encountered in the network. 9 Q. I mean, you have quite significant people from 10 the business there, Angela van den Bogerd we've 11 heard about, and others. You also have the head 12 of Legal, Rodric Williams, attending this 13 meeting. Did it strike you as surprising that 14 an incident was being raised among such 15 a significant group of people? 16 A. Not considering the business context at the 17 time, no. 18 Q. What was that context? 19 A. The ongoing GLO proceedings. 20 Q. Do you think it would have been helpful to have 21 had meetings of this sort involving senior 22 individuals within the Post Office and Fujitsu 23 discussing earlier incidents of the type that 24 we've discussed today? 25 A. In hindsight, yes. 87 1 Q. Do you know by this stage, as late as 2018, how 2 this incident was being cascaded down to 3 branches? 4 A. I don't recall specifically the communications 5 to the branches but I believe there is mention 6 of communication to the branches further down 7 this document. 8 Q. Perhaps we can look down page 3. There's 9 mention of Impact there: 10 "As of [22 January 2019] 19 branches are 11 impacted. Impacts are: 12 "Confusion for the user. 13 "Transactions accounted for against the 14 wrong BP/TP. 15 "Calls into contact centres from 16 users/branches impacted. 17 "Remedial action required by POL to resolve 18 cases." 19 Now, are you aware of there being 20 a significant effort to notify branches, 21 branches that may not know that they're affected 22 by this incident, that it is a known incident. 23 A. In terms of the sort of branches that were 24 identified by Fujitsu, there was a specific 25 process to contact those branches, discuss the 88 1 cause of the issue and the branch and rectify 2 any accounting issues. 3 Q. In terms of branches that hadn't been identified 4 by Fujitsu, was there a way of notifying them of 5 this particular incident, an incident that 6 presumably has an affect on the cash balancing? 7 A. As I say, I think further down in this -- 8 I think it's this document, there is mention of 9 a wider communication to branches that were on 10 Smart ID at this point. I don't recall the 11 content of that communication. 12 Q. We can scroll down, if you like. 13 A. Yeah, sorry the top of that page there, "BAU 14 impact". 15 Q. So: 16 "Contact centres ... minimal impact ... 17 "Live service desk: minimal impact ... 18 "Finance Service centre: minimal impact ... 19 Transaction correction team are contacting 20 branches to ensure accounts balance." 21 Do you know if that is affected branches or 22 those that are known to be affected or do you 23 think that's contacting branches more broadly? 24 A. I think that is the branches that are impacted 25 but there is some documentation that I have 89 1 seen -- apologies, like -- I can't find it at 2 the moment, but it is one of these Horizon issue 3 management reports where it mentions wider 4 communications to branches that are on Smart ID. 5 I think I provided a later version of this 6 particular report. Could we bring that up? 7 Q. We may not need to. I can ask you a very broad 8 question, which is are you aware, as at 2018 or 9 as at now, of a policy being in place within the 10 Post Office that, if you get a significant 11 incident that affects cash account balances, 12 that it is distributed widely across the network 13 to as many subpostmasters as you possibly can, 14 as many branches as you possibly can? 15 A. I'm not aware of a policy to that effect. I can 16 only reflect on my experience of this particular 17 incident, which I believe the project manager 18 tried to manage with maximum transparency, both 19 with internal stakeholders and with the branches 20 that are impacted. 21 Q. Can we look at POL00037819. This is a paper 22 that seems to have been written by you. Do you 23 remember this paper? 24 A. I do. 25 Q. If we look at page 3 we have there what seems to 90 1 be the same issue. Can you tell me if it is in 2 fact the same issue. It's 4.3, "Receipts & 3 Payments Misbalances -- FIXED [September] 2019". 4 A. Yeah, that's the same issue. 5 Q. So it seems there the issue was fixed a year 6 after that earlier paper. So it took quite 7 a while to fix; do you remember it taking quite 8 a while to fix that issue? 9 A. Yeah. 10 Q. I'll just read that for the record. It says: 11 "Multiple login functionality allowed a user 12 with the same Horizon ID ... to be logged in at 13 more than one terminal providing any one of the 14 sessions was active ... Fujitsu identified in 15 cases where an SU or Office Balance was 16 initiated in those circumstances then it could 17 cause receipts and payments misbalances in the 18 branch which had to be corrected by 19 a Transaction Corrections. 20 "These issues were fixed by the Horizon 21 release in September/October 2019, which put in 22 place further controls on multiple logins 23 related to Stock Unit and Branch Balancing and 24 changing Stock Units." 25 You have in your witness statement, it's 91 1 paragraph 282 of your witness statement, listed 2 a wide range of people who were kept in the loop 3 about this particular issue. 4 A. Mm-hm. 5 Q. You said that Esther Harvey sought to ensure 6 various internal stakeholders were informed. 7 Who is Esther Harvey? 8 A. She was the project manager for the Smart ID or 9 Enhanced User Management Programme. 10 Q. Why was she responsible for ensuring various 11 internal stakeholders were informed about this 12 issue? 13 A. That was part of her role as project manager. 14 Q. Which stakeholders do you have in mind when you 15 refer to informing stakeholders? 16 A. The Smart ID project team itself, in the first 17 instance, so that would have been myself and my 18 colleagues that were working on the Smart ID 19 project; Angela van den Bogerd; Julie Thomas, 20 who was the project sponsor; the relevant 21 parties in the FSC, which was the Finance 22 Service Centre in Chesterfield, who would be 23 dealing with any rectification of accounts. So 24 that would kind of be broadly the ones I'm 25 thinking of. 92 1 Q. How high up within the Post Office was this 2 receipts and payments misbalancing issue known, 3 2018/2019? 4 A. At quite senior levels, I would say. I mean 5 Angela van den Bogerd was certainly quite senior 6 and involved with the GLO. 7 Q. Do you think it went above her? 8 A. I couldn't say. 9 Q. What was the relevance of this particular issue 10 to the GLO? 11 A. Well, because the GLO were looking at Horizon 12 Issues, bugs, defects, and obviously this was 13 a defect that was causing -- impacting accounts, 14 branch accounts. 15 Q. Can you remember any conversations with anyone 16 in management within the Post Office about the 17 significance of this issue in that context? 18 A. Not that I had, no. 19 Q. Are you aware of any conversations that others 20 had? 21 A. Not in terms of the content but I'm aware that 22 conversations were held with Julie Thomas and 23 Angela van den Bogerd and, of course, they were 24 on the weekly calls that we were having to track 25 and monitor this, which is sort of documented in 93 1 the reports that we brought up earlier. 2 Q. You say in your statement at paragraph 296 that 3 this is an example which shows that the Post 4 Office's management of Fujitsu was lacking. Can 5 you tell us what you mean by that? 6 A. There was -- when this happened, not least 7 because we were going through the GLO at the 8 time, there was a lot of soul searching within 9 the team about how this had occurred and whether 10 we could have done anything to identify it. 11 And, at least from my perspective, I felt 12 that these kind of issues, given that the 13 functionality that led to this was a change, 14 a significant change to the way that user 15 management was managed on Horizon, I personally 16 felt that this should have been flagged to us as 17 part to the impact assessment of those changes, 18 such that we could have avoided these issues. 19 Q. Have you reflected on previous bugs, errors and 20 defects that you've identified or been involved 21 in and whether, in those particular cases, there 22 were issues with the management of Fujitsu? 23 A. I think particularly since the GLO findings and 24 being aware of some of the points made in those 25 cases, I would say, yes, we weren't finding out 94 1 about bugs quickly enough, in my view. So yeah, 2 I would agree that, following reflection, there 3 should have been better management of Fujitsu. 4 Q. You say you weren't finding out about them 5 enough or quickly enough. Whose responsibility 6 was that on either side? 7 A. My -- as I say, my understanding is, in terms of 8 the incident management and problem management 9 processes, is that it should be flagged through 10 that route. Where there are issues with the 11 system, it should be raised as across the main 12 problem and managed with communications going 13 out to branches as appropriate. 14 Q. Raised by who? 15 A. In cases where the branch is aware of it, the 16 branch, or from Fujitsu themselves through their 17 monitoring. So it should be raised by the 18 branch to Horizon System helpdesk and then by 19 Fujitsu across the Post Office, if it's 20 something that's going to impact multiple 21 branches. 22 Q. Has that improved in recent months, years? 23 A. Well, I can only go on the experience that I've 24 had with the Smart ID bug and I did find, in 25 that particular case, that Fujitsu were very 95 1 quick to react, they did a lot of work in terms 2 of identifying not just the scenarios that had 3 caused issues in the live estate but also other 4 potential scenarios that could potentially cause 5 issues in the live estate from the same sort of 6 functionality. So, yeah, my experience was that 7 that was improved, if you looked back, compared 8 to something like the Callendar Square issue. 9 Q. Was there a marked shift of any sort following, 10 for example, the Group Litigation or since the 11 beginning of this Inquiry? 12 A. My exposure has been through, you know, some 13 very limited windows where I have been dealing 14 with bugs. So I can only say that compared to 15 the Callendar Square, my experience of working 16 with Fujitsu collaboratively to resolve the 17 Smart ID EUM bug in 2018 was much improved. 18 We had a good working relationship with 19 Fujitsu at that time. I found them very helpful 20 and, as I say, there was a -- there was 21 documentation produced that identified all of 22 these various scenarios that could potentially 23 cause the receipts and payments mismatches. 24 Q. I'm going to move on now to a few discrete 25 issues. I won't be very long with them but 96 1 they're ones I have been asked to raise with 2 you. The first is transaction correction 3 process and can we go to POL00039024. You've 4 dealt with this at paragraph 137 of your 5 statement. This is a document from October 6 2007. Can you briefly tell us what your 7 involvement was in issues concerning transaction 8 corrections? 9 A. With this particular document, I think I stated 10 this in my witness statement, my colleague Nicky 11 Barraclough wrote the document. I was -- I'm 12 listed as a contributor here and I believe that 13 my contributions were predominantly reviewing 14 the process flows, the process flows in the 15 document, and also the reporting that are in the 16 appendices the spreadsheets that we use to 17 manage deductions from remuneration. 18 Q. Thank you. I'm just going to take you to two 19 passages in this document and the first is on 20 page 2, and it's the bottom of page 2, 21 "Timeliness of Issuing Transaction Correction", 22 it says: 23 "The time taken to issue Transaction 24 Corrections can be slow, including some cases 25 where it has taken 2 years to clear the ledger. 97 1 Not only does this delay the time taken to 2 recover outstanding debt, but it also leaves the 3 subpostmaster feeling frustrated that they are 4 being notified to rectify a mistake that they 5 may not remember, or where the evidence has been 6 destroyed that would support a dispute to the 7 Transaction Correction." 8 This is 2007. Is this something that you 9 remember, an issue that you remember? 10 A. Yeah, I remember concerns about the length of 11 time it took to issue transaction corrections. 12 Q. Looking at it from 2007, was that a historic 13 issue at a particular point in time, in 2007, or 14 something that continued after this document had 15 been written? 16 A. I don't think it was any different, in my 17 experience at least, before this document. But 18 there were always issues that branches would 19 raise about the length of time it was taking to 20 create an issue, transaction corrections to 21 them. I think some of it, from my recollection, 22 was tied to getting information from clients, 23 which often had, sort of, long lead times. But, 24 yeah, it's not an issue that suddenly appeared 25 in 2007. 98 1 Q. If we go over the page to page 3 there's another 2 section on "Disputing the Transaction 3 Correction". It's the first paragraph there 4 that I would like to ask you about. It says: 5 "The current process for disputing 6 Transaction Corrections allows the subpostmaster 7 to challenge the error at every stage of the 8 process, even after pressing the button on the 9 Horizon System to settle the debt centrally. 10 This delays the process in recovering any 11 outstanding monies, and can be used as 12 a deliberate delaying tactic in order to delay 13 making payment." 14 Is that something that you remember? 15 A. Not specifically. No. 16 Q. The mention there of a deliberate delaying 17 tactic, do you recall a culture within the Post 18 Office to assume that subpostmasters, assistants 19 and others were up to no good or using 20 particular tactics to delay payments? 21 A. No, not to my knowledge, no. 22 Q. Thank you. Totally different topic, the PING 23 Project. Can we look at FUJ00091292, please. 24 Very briefly, because you have dealt with it in 25 your witness statement, what was the PING 99 1 Project? 2 A. The PING Project was a way of improving the 3 accounting for third-party terminals, for 4 example the Camelot Lottery terminal. Before 5 the PING Project was implemented, branches would 6 have to key their transactions into the Lottery 7 terminal as they were doing Lottery sales, and 8 so forth, and then they would go and key that 9 into Horizon for accounting purposes. What the 10 PING Project sought to do was remove that human 11 input by harvesting the transactions directly 12 from the third party terminal, eg the Lottery 13 terminal, and send those over to Horizon as 14 transaction acknowledgements. 15 Q. Thank you. If we turn to page 5, you're listed 16 as a reviewer in respect of this particular 17 document. 18 A. Yes. 19 Q. If we look at page 9, there is the background. 20 I just want to briefly take you to that 21 background. It's the second paragraph and down. 22 It says: 23 "The client data is uploaded into POL-FS and 24 compared with the equivalent HNG-X data which 25 has to be manually input by the agent/counter 100 1 clerk. Ideally the data, when compared, should 2 be the same but number of conformance issues 3 have been identified where agents/counter clerks 4 do not perform end of day routines correctly, do 5 not input the Camelot details into HNG-X as they 6 should, and can key incorrect figures, leaving 7 Product and Branch Accounting with 8 a reconciliation difference. This difference 9 may require the issuing of a transaction 10 correction." 11 Can you tell us what a "conformance issue" 12 is? 13 A. In this particular context, I believe it means 14 either not inputting the lottery figures into 15 Horizon at the end of the day or inputting them 16 incorrectly, such that there would be 17 a mismatch. 18 Q. In terms of non-conformance, is that something 19 that the Post Office generally understood would 20 happen? 21 A. Yeah, I mean, it's reliant on the postmaster in 22 branch keying figures and keying figures 23 correctly. 24 Q. To what extent do you think the Post Office put 25 too much emphasis on conformance issues rather 101 1 than, for example, bugs, errors and defects, 2 historically, or ... 3 A. Yeah, I mean, I was involved in a team that was 4 managing conformance for a number of years, and 5 I think, in retrospective, there should have 6 been -- there certainly should have been more 7 focus on bugs and issues and the management of 8 those bugs and issues but it was still important 9 to manage conformance. We had contracts with 10 clients, Royal Mail being most notable, that if 11 we didn't achieve certain levels of performance, 12 we were financially penalised. 13 But, in terms of the balance between those 14 two things, I think an argument can be made that 15 we should have put more emphasis on bugs and 16 defects, certainly. 17 Q. Thank you. My final issue is IBM, who you've 18 mentioned in your witness statement. Between 19 December 2015 and March 2017 you were business 20 readiness lead and that involved working with 21 IBM to replace Horizon. Do you know why the 22 Post Office wanted to replace Horizon at that 23 stage? 24 A. I think it's just old technology and it was time 25 to move on to something that was quicker to 102 1 develop, easier to develop, so that new products 2 could be brought on more easily. That was the 3 sense that I got when I joined that programme. 4 Q. Do you know why IBM was chosen for that? 5 A. I wasn't party to those sort of contractual 6 discussions. I don't remember anybody talking 7 to me about what the options were and whether 8 there were other alternatives that we'd looked 9 at. 10 Q. Do you know why the decision was taken to 11 withdraw from that project? 12 A. I wasn't directly involved but I did hear 13 suggestions that there were concerns over cost 14 and there were concerns over whether it would be 15 deliverable to the timescales we needed it to 16 be. I think this was not just a simple 17 changeover, like Horizon Online, where we were 18 introducing new software; it was hardware, data 19 centre and the front office as well, so it was 20 a considerable challenge. 21 MR BLAKE: Thank you very much, Mr Turner. Those 22 are all of my questions. 23 I think Mr Stein has some questions. 24 Mr Whittam? No. 25 Oh, and Ms Page has a question as well. 103 1 We'll start with Mr Stein. 2 Questioned by MR STEIN 3 MR STEIN: Sir, I'm grateful, my learned friend, 4 Mr Blake, raised the question, sir, this morning 5 of disclosure and I'm just going to pause to 6 note that on 14 February this year, we wrote to 7 the Inquiry concerning the difficulties with 8 disclosure and, indeed, concerning the 9 difficulties that the lack of closure of scripts 10 and other documents were going to be causing us 11 with putting forward questions for this witness, 12 Mr Turner. 13 So I'm going to ask that that letter be 14 resent so that it goes directly to you, sir, and 15 therefore you can see how it is that we frame 16 matters at that date. That wasn't the first 17 time, sir, we've raised issues regarding 18 scripts. 19 With that as a starting point, I'll now move 20 to my questions for Mr Turner. 21 Mr Turner, as you've heard from Mr Blake and 22 indeed myself, there have been some issues 23 regarding disclosure that may inhibit our 24 ability to ask you some wider questions and that 25 could cause, therefore, a need for you to return 104 1 to the witness box at a later stage. 2 A. Okay. 3 Q. I'm just going to refer you to your own 4 statement, I'll read a part of it, paragraph 12, 5 page 5. You refer to your role to analyse the 6 data arising from calls to identify trends and 7 common issues experienced by subpostmasters. 8 A. Yes. 9 Q. Okay. Do you call yourself a data analyst? 10 What would be the term that you would use? 11 A. At that time it was incident analyst. 12 Q. Incident analyst. Right. Now, help us on the 13 other side of the calls from the subpostmasters. 14 You were aware at that time that Fujitsu also 15 had their own helpdesk taking calls? 16 A. Yes. 17 Q. We know that because indeed you have spoken 18 today about the fact that some calls were taken 19 by the Post Office team and some were taken by 20 Fujitsu; is that correct? 21 A. That's correct. 22 Q. Right. Now the analysis of common issues 23 experienced by subpostmasters, that job would 24 need to be shared, I assume, between the Post 25 Office and Fujitsu? 105 1 A. Yes, so the -- both helpdesks should have had 2 an incident management framework in place that 3 included analysis of calls coming in, trends, 4 and I would expect that if system issues were 5 coming through, common system issues, that that 6 would manifest in HSD and be raised to their 7 Problem Management Team, if appropriate. 8 Q. Now, within that answer you used the word 9 "should" and the words "I would have 10 expected" -- 11 A. Yes. 12 Q. -- in relation to those sorts of systems. Now, 13 clearly you can give evidence about the Post 14 Office system. What knowledge did you have at 15 this time about the Fujitsu equivalent systems? 16 A. I'm trying to remember. I know there was 17 documentation that showed how Fujitsu should 18 manage incidents and track common themes, and 19 that would have been the basis from my 20 knowledge. My sort of layman's view of it was 21 that they would be replicating something very 22 similar to what we were doing in the NBSC. In 23 other words, they would have some kind of data 24 team that was looking at trends. I didn't have 25 any direct contact with their own incident 106 1 analysts myself. 2 Q. That was going to be my next question which is: 3 how were these things tied together? If it 4 wasn't tight together at your level at that 5 time, at what level was it tied together and who 6 can you direct us to, who should be able to know 7 how these things came together? 8 A. There were forums at our level where common 9 issues were discussed. I didn't have any direct 10 contact with their equivalent of the incident 11 analysts but there were forums between the 12 helpdesks. But I would imagine if there were 13 common themes coming through, these should have 14 been raised throughout the problem management 15 process and then managed at that level between 16 the two Problem Management Teams. 17 And then there was a Service Management 18 Forum, as well, where the two parties met to 19 similarly discuss problems. 20 Q. The numbers of calls, in terms of working out 21 numbers of calls coming into the Post Office 22 helpdesk versus Fujitsu, can you give us an idea 23 whether this was an even spread of 50/50 that 24 came into the Post Office versus Fujitsu, or 25 whether it was more Post Office or more Fujitsu? 107 1 Have you got any ... 2 A. I'm sorry, I don't recall the numbers or the 3 sort of volumetric data at this stage. 4 Q. Just remaining with this question, you've 5 explained that there should have been something 6 operating in a similar way by way of incident 7 analysis with Fujitsu? 8 A. Yes. 9 Q. Did you see documents that would have provided 10 the other side of the coin, the Fujitsu side of 11 the coin, and your documents synthesised, put 12 together, so that you've got a picture? Did you 13 see such things? 14 A. Not to my recollection, no. 15 Q. Were there such things? 16 A. I don't know after this amount of time. There 17 were forums where common issues were discussed, 18 so there was some interaction between the two 19 helpdesks, and we tried to resolve issues, 20 operational issues through those forums as well. 21 But I don't remember seeing specific 22 documentation other than what I've referred to 23 in my witness statement. 24 Q. Just before I then ask you a little bit more 25 about how documents were kept, shouldn't there 108 1 have been documents that provided a combination 2 of these are the issues we are seeing within POL 3 and these are the issues we are seeing within 4 Fujitsu, and they are showing similar issues 5 being raised or similar problems or good things 6 or indeed bad things. Shouldn't such things 7 have existed and shouldn't they have come to 8 you? 9 A. Yes, yeah. 10 Q. Now, you have remained at the Post Office for 11 quite some time, indeed I think it's essentially 12 your career so far? 13 A. That's right. 14 Q. You describe in your statement the analyst work 15 you carried out in relation to incidents. Can 16 you just help us with access to those, the 17 analysis that you've carried out. There must 18 be, going back now to what, the early days of 19 Horizon and then through, there must be hundreds 20 of these documents that you've been created or 21 been part of the creation of; where are they? 22 A. What type of documents are you talking about, 23 specifically? 24 Q. Well, let's go with the analysis of the NBSC 25 call performance, call operating performance, as 109 1 an example, or trends within difficulties being 2 experienced by subpostmasters. Where are these 3 documents? 4 A. I don't know at the moment. There certainly 5 were reports produced. I know where problems 6 were raised off the back of those reports they 7 were logged in a system that we used at that 8 time called Remedy, which was -- 9 Q. Did you say "Remedy"? 10 A. Remedy, yes, which was both the incident 11 management tool and also the problem management 12 tool so that's where the calls were logged, and 13 any problems that were raised to the Problem 14 Management Team were also logged in there with 15 the originator. 16 Q. We've encountered some problems with terminology 17 being used, so scripts for helpdesks are 18 sometimes called other names other than scripts? 19 A. Right. 20 Q. Okay? So help us identify, the analysis work 21 you would have carried out, so this was the 22 analysis to identify trends and common issues 23 experienced by subpostmasters, what would have 24 been the title or what would be the title of 25 such analysis reports? 110 1 A. It would be an incident analysis report on 2 topic X and, if the problem was raised, it would 3 be a problem record on that topic. 4 Q. Do you have access to systems that contain this 5 documentation? 6 A. I don't. 7 Q. Help us why not? 8 A. The passage of time. I've moved through various 9 roles since then. 10 Q. Have you been asked by either the Post Office or 11 lawyers on their behalf regarding your knowledge 12 of access to this material where it might be? 13 A. No. 14 Q. Despite the fact that you're the author of such 15 documentation over time? 16 A. Yes. 17 Q. Now, I'm just going to complete this particular 18 topic range. Oversight of the helpdesk, this is 19 the Fujitsu helpdesk. Now, you would expect, 20 I imagine, that Fujitsu should carry out its own 21 oversight, governance operation regarding 22 helpdesk; okay? Who within Post Office 23 monitored Fujitsu's oversight of its own 24 helpdesk? 25 A. The service management team, I think, from 111 1 recollection. 2 Q. Service management team? 3 A. Yeah. 4 Q. Right, and a particular name of an individual 5 you would recall at any given time? 6 A. At any given time, somebody like Dave Hulbert 7 would have been heading up that team. That's 8 where I believe the monthly service management 9 reviews were managed. 10 Q. There's a term used, which is used within your 11 statement and referred to in a document that you 12 were taken to today by Mr Blake, which is 13 a cross-domain problem. 14 A. Yes. 15 Q. Is that a term of art within the Post Office to 16 describe a type of document or report or system 17 where there is a problem that is known to both 18 Fujitsu and the Post Office? 19 A. It's terminology to describe any problem that 20 requires interaction with a third-party 21 supplier, ie it's not just within Post Office's 22 gift to resolve. 23 Q. Right. Is there any sort of risk analysis 24 system that has been developed to decide when 25 such a thing does reach the stage of it being 112 1 a cross-domain problem? Is there a trigger for 2 this to say "Look, this is cross-domain"? 3 A. There were problem management documents around 4 prioritisation, and so forth, in terms of 5 issues. 6 Q. Right. And risk analysis similar? Were there 7 documents around that considered risk analysis 8 in the way that traditionally regulators often 9 do, thinking about either use of a traffic light 10 system, like green or amber or red, to analyse 11 risk? 12 A. I don't recall specific documents but I do 13 remember conversations with problem managers 14 about rating problems based on the number of 15 branches impacted, and so forth. 16 Q. Thank you. One remaining question. You were 17 asked a number of questions by Mr Blake 18 regarding the receipts/payments mismatch issue. 19 A. Yeah. 20 Q. He referred you to a document, which is a Post 21 Office Fujitsu document, that discussed that. 22 That's the "Receipts/Payments Mismatch Issue". 23 He referred to that at around the period of time 24 which is 2010 and ongoing. 25 One thing that is unclear from your evidence 113 1 so far is what were you actually told about that 2 issue, the receipts/payments mismatch issue, 3 around about that time? What were you told 4 about it? 5 A. I don't remember. I remember being told that 6 there was an issue, having seen the document 7 last week. But I don't remember exactly what 8 I was told, whether the nature of that 9 conversation was just as per the action point 10 out of that meeting, can you check whether this 11 would cause an issue with the profile that 12 I managed at that time, the branch profile? But 13 I'm sorry, I don't remember what I was told. 14 Q. I am grateful, Mr Turner because your evidence 15 was clearly hesitating -- you were hesitating in 16 your evidence at the time regarding this and so 17 it led to these questions, which is: what access 18 did you actual have to this information? 19 You refer to that as a conversation. Was it 20 actually a discussion, an oral discussion or was 21 this is an email discussion or some other means 22 of discussing it? 23 A. As I say, I don't remember and, using sort of 24 "conversation" in a more informal context, 25 I don't remember whether it would have been 114 1 somebody dropped me a note, saying, "Can you 2 look at this?" I hope if it was, we'd be able 3 to find documentary evidence of that but it may 4 very well have been a phone call. 5 Q. Yes, and if I ask whether you recall being 6 briefed, in other words being given a written 7 document "Mr Turner, this is something you need 8 to be aware of, this is the issue, we want you 9 to do X action", do you recall getting anything 10 like that? 11 A. I'm sorry, I don't. 12 Q. No. If someone were to send you such 13 a document, who would that -- which team would 14 have given you that kind of heads-up or that 15 information? 16 A. Well, I think considering that the action listed 17 was against Julia Marwood, who was an attendee 18 of that meeting, I would imagine that any 19 interactions that I had had, and if there was 20 an email trail, I imagine that would be from 21 Julia to me. 22 Q. Yes, because she's listed as being the owner of 23 that particular issue, under her initials "JM"? 24 A. That's right. 25 Q. Yes. Excuse me one moment. 115 1 A. Sure. 2 MR STEIN: Sir, I'm grateful, thank you. 3 SIR WYN WILLIAMS: Thank you. 4 Ms Page, I think, has a question. 5 Questioned by MS PAGE 6 MS PAGE: Thank you, sir. 7 In your witness statement, Mr Turner, you 8 note that because of a baseline faith in Horizon 9 as a robust system, POL was perhaps not as 10 attuned to concerns raised by SPMs as it should 11 have been? 12 A. Yes. 13 Q. Can I just ask you about that when you were in 14 your role when the Callendar Square bug was 15 brought to your attention. I don't know if 16 you're particularly aware of this, but in 17 Scotland it was a rather different prosecutorial 18 process and Post Office were not allowed to 19 bring their own prosecutions; were you aware of 20 that? 21 A. I was not no. 22 Q. Well, what we've obviously got here is 23 a Scottish problem with a Sandra MacKay picking 24 up the issue in Callendar Square in Falkirk. It 25 was first noticed then in Scotland. 116 1 A. Yes. 2 Q. Does that make sense? 3 A. It does. 4 Q. Did you ever receive -- what I'm interested in 5 is the possibility that people in Scotland were 6 more attuned to the problems that subpostmasters 7 faced because their experience of having to deal 8 with third-party prosecutors may have made them 9 more attuned, if that makes sense? 10 A. I never received or had a conversation to 11 suggest that, no. 12 Q. Did you ever receive any problems like the 13 Callendar Square bug type problem, or indeed 14 problems generally? Did they ever come across 15 your desk from area managers working in England 16 and Wales? 17 A. I don't recall any specific instances but 18 I would expect, yes, there were cases from 19 England and Wales. My role in terms of that 20 escalation role covered the entire country, so 21 it could come from any area of the country. 22 Q. You described earlier in your evidence that the 23 Callendar Square problem was perhaps somewhat 24 unique. What was the qualities of it that made 25 it unique? 117 1 A. Unique in my experience, at least? 2 Q. Yes. 3 A. You know, I think it was -- in terms of my 4 exposure, I'd not come across anything like that 5 previously. I think I said in my witness 6 statement that of the issues I dealt with during 7 my time as network co-ordination manager, 8 I didn't remember the details, but I still 9 remembered that I'd dealt with this particular 10 bug at Callendar Square. 11 Q. But do you not particularly have memories of 12 other bugs that you dealt with? 13 A. No. I remember other issues with Horizon and 14 other parts of the business but this, to me, 15 would have been an unusual and anomalous kind of 16 case, at least in terms of my exposure to it. 17 Q. So in terms of your exposure, bugs of this 18 nature were not raised to you at all from 19 England and Wales? 20 A. No, that's -- I'm not saying that I didn't have 21 some issues raised, but I think -- I don't 22 remember whether there were issues raised from 23 England and Wales, but I would expect that there 24 would be. 25 Q. But they didn't stand out in your mind as 118 1 a significant problem in this way, in the way 2 that the Callendar Square bug did? 3 A. No. 4 MS PAGE: Thank you. Those are my questions. 5 SIR WYN WILLIAMS: Thank you, Ms Page. 6 Is that it, Mr Blake? 7 MR BLAKE: I believe that is, sir, yes, unless you, 8 sir, have questions. 9 SIR WYN WILLIAMS: No, thank you very much. 10 Mr Turner, I'm very grateful to you for 11 providing such a detailed witness statement in 12 response to very many questions, and I am 13 equally grateful that you've come before the 14 Inquiry to give oral evidence. So thank you. 15 THE WITNESS: Thank you, sir. 16 SIR WYN WILLIAMS: I should say that in view of the 17 fact that, as you will have gathered, there 18 appears to be a glitch in the disclosure 19 process. It is not inconceivable that you will 20 be asked to return, but we will keep that under 21 review, all right? 22 THE WITNESS: Okay. 23 MR BLAKE: Sir, shall we say 1.50? 24 SIR WYN WILLIAMS: Yes, by all means Mr Blake, yes. 25 MR BLAKE: Thank you very much. 119 1 (12.50 pm) 2 (The Short Adjournment) 3 (1.50 pm) 4 MR STEVENS: Good afternoon, sir. Can you see and 5 hear me? 6 SIR WYN WILLIAMS: Thank you. 7 MR STEVENS: If I may call Mr Blackburn. 8 GARY DAVID BLACKBURN (affirmed) 9 Questioned by MR STEVENS 10 MR STEVENS: Mr Blackburn, as you know, my name is 11 Sam Stevens and I ask questions on behalf of the 12 Inquiry. Please could I ask you to state your 13 full name? 14 A. Gary David Blackburn. 15 Q. Firstly, thank you for giving evidence to the 16 Inquiry today and thank you for providing 17 a written statement to which I'd like to turn 18 now. Do you have a bundle of documents in front 19 of you? 20 A. I don't. 21 Q. You don't. Right. Bear with me, I'll see where 22 that is. 23 Apologies, sir. We will just wait while 24 that bundle arrives. 25 SIR WYN WILLIAMS: Sure. Yes. 120 1 MR STEVENS: Thank you, Mr Blackburn. Now, one of 2 those bundles should have your witness statement 3 at the front of it, behind tab A. 4 A. Yes, got that. 5 Q. Excellent, apologies for that and thank you. So 6 that witness statement should run to 16 pages. 7 A. Correct. 8 Q. If you turn to page 15, you'll see paragraph 28 9 being the last paragraph, and at the bottom is 10 that your signature? 11 A. It is. 12 Q. Can I ask you to confirm that the facts within 13 that statement are true to the best of your 14 knowledge and belief? 15 A. They are. 16 Q. Thank you, Mr Blackburn, that stands as your 17 evidence in the Inquiry. I am going to ask you 18 some questions about that. You joined the Post 19 Office in 1985? 20 A. Yes, I did. 21 Q. From 1985 to 1994 you worked, I understand, in 22 Crown Office branches around Huddersfield? 23 A. Correct. 24 Q. Then in 1994, you were posted to the northeast 25 regional office? 121 1 A. That's correct. 2 Q. At that stage, you say in your evidence that you 3 worked on a relatively new helpdesk that was 4 created to support the region? 5 A. That is also correct. 6 Q. At this point, the Post Office was split into 7 seven regions? 8 A. It was. 9 Q. You say that, subsequently, those individual 10 regional helplines merged in to become what we 11 now know as the Network Business Support 12 Centre -- 13 A. That is correct. 14 Q. -- or the NBSC? 15 A. The NBSC, yes. 16 Q. Do you remember when that was? 17 A. It coincided with the introduction of the 18 Horizon solution, so prior to that, it was -- 19 the northeast regional desk or the various 20 regional desks were there pre-automation and 21 then also for -- I think it was called the ECCO+ 22 system. So I'm fairly certain it was a part of 23 a larger reorganisation of the Post Office and 24 the introduction of Legacy Horizon. 25 Q. Casting your mind back to when you worked on the 122 1 pre-NBSC desk and pre-automation desks, what 2 were the types of enquiries you would most often 3 receive? 4 A. It was very transactional in nature it was 5 really supporting the branches in terms of "How 6 do I complete it", particularly a transaction 7 for a member of the public, interspersed with 8 accounting queries but, as we may well end up 9 discussing, nowhere near to the extent it was 10 post-automation. That was the old paper-based 11 cash account, which was relatively 12 straightforward and simple to follow. 13 We also supported the Retail Network 14 Managers at the time in any enquiries that they 15 may have had and we were also there as a bit of 16 an emergency point of contact for events such as 17 burglaries and robberies, that kind of thing. 18 Q. What did you think of the quality of the advice 19 and assistance that was able to provided by the 20 regional helpdesk in comparison to the national 21 one? 22 A. Yeah, it was, in my personal opinion, better, 23 and purely based upon the fact that it was 24 staffed with people who had similar career 25 profiles to myself, so that all worked in the 123 1 branch network, primarily in the directly 2 managed branches in the Crown Offices, but one 3 or two of my colleagues at the time had also 4 worked in the independent branches, so they had 5 got a wealth of experience of completing the 6 varied transactions that we had at the time in 7 the business. 8 Q. So when you say in the regional helpdesk, you 9 pointed to the experience of the people within 10 it as a strength. 11 A. Yes. 12 Q. Can you explain why that was different from the 13 NBSC? 14 A. Well, when -- so the regional helpline converted 15 into the Horizon trial desk. So I was part of 16 the trial desk then, which was run out of 17 an office in Leeds. I think the branches in the 18 trial were from both the Leeds and Bristol 19 regions, I want to say. It was still very much 20 the same group of people, albeit everything was 21 new to all of us at that time, in terms of 22 Horizon itself. But the business knowledge was 23 still the same, the transactions were still the 24 same, although completed in a different manner. 25 As we moved then into what was the 124 1 national -- the Network Business Support Centre 2 and the national -- the only helpdesk for 3 business enquiries, naturally we had to expand 4 and recruit more people. We moved from Leeds to 5 a greenfield site in the Dearne Valley, it was 6 fairly isolated at the time. There was one or 7 two other contact centres in the environment, so 8 we picked up new recruits from the surrounding 9 area, people, in effect, who had -- it was their 10 first experience of working for the Post Office, 11 they'd no prior experience. 12 So, for me, naturally, there was just 13 a slight diluting of the quality of the 14 individual on the desk, as they obviously that 15 to go through a steep learning curve themselves. 16 Q. Is it fair to say there was a lot of 17 institutional knowledge when moving to the NBSC? 18 A. That is my opinion, yes. 19 Q. For that reason, is it fair to say that, because 20 of that loss of institutional knowledge, it was 21 important to ensure that the new members of the 22 NBSC were adequately trained? 23 A. Yes, vital. Yes. 24 Q. We'll come on to training shortly. Before I do, 25 the pre-NBSC regional helplines, I appreciate 125 1 you can only speak for the northeast, but did 2 they ever have any communication, the helpline 3 as such, with people involved in bringing 4 prosecutions against subpostmasters for false 5 accounting or theft? 6 A. No, not at that time, no. It wasn't part of the 7 remit. 8 Q. The same question, but in respect of auditors: 9 was there any communication between the regional 10 helpline and the audit teams? 11 A. I can't remember, if I'm honest. There may well 12 have been, because regional auditing was a very, 13 you know, business as usual part of everyday 14 Post Office life. So there may have been. But 15 it's not something I recollect. 16 Q. So in paragraph 4 of your statement, which we 17 don't need to turn to, you say that you were 18 a team leader on the NBSC. Could you just 19 summarise, briefly, what that role entailed? 20 A. It was line management of a group of individuals 21 who would be actually manning the telephones and 22 taking the calls from the branches. When we had 23 busy times, I would also take calls from the 24 branches but it was providing a layer of people 25 management and support and guidance for any 126 1 calls that they were struggling to answer. 2 Q. When did you finish that role as a team leader? 3 A. I think -- now this is where chronology becomes 4 quite hazy for me but I think it was -- I'm 5 going to say around about 2001/2, I think, and 6 I went down -- what was downstairs into business 7 service management, as it was called at the 8 time, and I -- 9 Q. Pausing there, could you say what business 10 service management's role was? 11 A. It's what I today would refer to as IT support, 12 some vital processes, change management, problem 13 management, secondary layers of incident 14 management. Things like that. 15 Q. What was your first role? 16 A. I started as a problem manager when I first went 17 down there. 18 Q. For how long did you hold that role? 19 A. Now, that's the bit I really can't recall. From 20 memory, even looking through all of the 21 documentation, I really have very little 22 recollection of my time in that team. I don't 23 think it was for very long, because I seemed to 24 move on quite quickly into other incident facing 25 and live service facing roles, so I'm sorry 127 1 I couldn't tell you. 2 Q. We'll come to those roles in a moment. In brief 3 terms, could you summarise what your role as 4 a problem manager was? 5 A. Yeah, and I think, as I've said in my statement, 6 it felt, looking back with hindsight, reactive 7 rather than sort of proactive, in the sense that 8 we would literally take trends and analysis from 9 the Network Business Support Centre, we'd also 10 have it provided to us by the Horizon System 11 Helpdesk and we would look for where there was 12 any remediation activities that might be 13 required to prevent future occurrences of 14 whatever the incidents had been. 15 Q. Your role after problem -- we don't know the 16 date, specifically, but what was your role after 17 that of a problem manager? 18 A. So I -- this is where my career, I do apologise, 19 becomes quite hazy, because I did so many roles 20 including interspersed with taking parts in 21 various programmes from a support perspective. 22 My next recollection was a live service desk 23 recall, that's the key role that I remember, 24 which was also a new desk which we'd set up at 25 the time, an internal facing desk. 128 1 Q. We'll come to that again in due course but, just 2 for chronology purposes, you were transferred 3 understanding TUPE to ATOS in 2014? 4 A. That's correct. 5 Q. Could you just explain what ATOS was doing at 6 that stage for you to be TUPE transferred 7 across? 8 A. Yes, so ATOS in -- for me, in simple terms, it 9 was Post Office outsourcing that -- what I'd 10 just described as the business service 11 management functionality and it was implementing 12 a new operating model for management of Post 13 Office's IT supply chain. 14 Q. I understand you left that role in 2017 -- 15 A. I did, yes. 16 Q. -- and you no longer work for the Post Office? 17 A. That's correct. 18 Q. Could we, please, turn up your witness statement 19 on the screen. It's WITN04650100, and 20 paragraph 10, please, on page 6. Thank you. 21 You state that: 22 "The purpose of the NBSC was to support the 23 branch network through answering 'how do I' 24 related transactional questions alongside the 25 Fujitsu Horizon Service Desk ... which was there 129 1 to support the branch network with technical 2 questions and queries in relation to the 3 technology (Hardware, Software and Network) that 4 had been provided." 5 What do you mean by "'How do I' 6 transactional questions"? 7 A. I mean literally a member of the public, being 8 in the branch, wanting to conduct a particular 9 transaction type and the branch not being quite 10 sure how to do that. At the time of Horizon, 11 that also involved, you know, what were more 12 complex navigational type questions through the 13 solution. So that was -- that is what I would 14 refer to as "How do I complete this transaction 15 for, you know, an item going overseas" or 16 something to that effect. That was the type of 17 things we did. 18 Q. That document can come down for the time being. 19 Thank you. 20 I think you've mentioned already a large 21 part of the NBSC's role was to assist with 22 balancing as well. 23 A. Very much so, yes. Yes. 24 Q. What training did you receive on joining the 25 NBSC to enable you to carry out your role as 130 1 team leader? 2 A. I think the only training that I remember was 3 training on the Horizon solution. So there was 4 no additional business training or anything like 5 that. It was a standard counter training 6 course, as it was at that time, which included 7 use of the Horizon System. 8 Q. Do you recall what that training involved? 9 A. I'm sorry, I don't. 10 Q. What about the new candidates, the people who 11 hadn't the institutional knowledge that you had? 12 Did they receive the same training or different? 13 A. Number 1, I believe, you know, they received 14 that element of training because the Horizon 15 System was now ultra-important but they also 16 received what I would call the standard Post 17 Office training, as if they were going to work 18 in a directly managed branch, for example. 19 Q. Do you recall if there was ever updater or 20 refresher training given to members of the NBSC 21 on how to use the Horizon System? 22 A. Not specifically, no, I don't. 23 Q. I'd like now to turn to the relationship between 24 the NBSC and the Fujitsu helpdesk, which you've 25 mentioned in your statement. Please can we turn 131 1 up FUJ00080405. 2 This is the "ICL Pathway/Post Office Counter 3 Limited Interface Agreement for the Network 4 Business Support Centre and the Horizon System 5 Helpdesk". If we could turn to page 4, please, 6 towards the bottom. Under "Contributors" we can 7 see that you -- the name on the right towards 8 the bottom, you contributed towards this 9 document? 10 A. Yeah. 11 Q. Do you recall the level of input you had on it? 12 A. I don't but I would assume, given my role at the 13 time, it was one really, you know, coming from 14 a point of managing a team of people who might 15 possibly have to interact with the Horizon 16 Service Desk or exchange ownership of 17 an incident that had been logged. 18 Q. Please can we turn to page 7 of that document. 19 So section 5 sets out general responsibilities, 20 and under (h) it says -- I'll summarise -- Post 21 Office Counters Limited and ICL Pathway: 22 "... are responsible for ensuring that known 23 problems or events, that may impact on the 24 everyday business of NBSC and HSH, are made 25 known to both helpdesks." 132 1 Can you recall how the bodies, the NBSC and 2 the Horizon Service Helpdesk, communicated with 3 each other regarding problems with the Horizon 4 System. 5 A. At that time, when we were talking of Legacy 6 Horizon, as opposed to HNG-X, the two desks were 7 very distinct, in fact the responsibility, in 8 effect, had been passed to the branch network to 9 determine which number they rang and, therefore, 10 who they spoke to. So I do remember there being 11 a lot of interaction between the two desks in 12 terms of swapping ownership. We also had -- the 13 NBSC had an admin team that, if there was 14 a wider sort of unplanned event that was 15 impacting either desk or a large volume of calls 16 suddenly started coming into the desk, they 17 would take the responsibility for the interface 18 and the communication. 19 I don't really remember anything further, 20 I'm afraid. 21 Q. Well, it may help assist your memory if we look 22 at page 16. Section 9 describes daily 23 interactions, which I don't need to trouble you 24 with but, if we can go to 10 the "NBSC/HSH 25 Review Forum" says: 133 1 "The performance of work undertaken across 2 the NBSC/HSH interface will be the subject of 3 monthly review. The output of the review forum 4 will provide input to: 5 "The Horizon Service Review Forum." 6 A. Right. 7 Q. Were you involved or do you recall being 8 involved in these monthly reviews? 9 A. I don't but, given my role, I would be very 10 likely to have been involved in at least some of 11 them, yeah. 12 Q. I'll ask the question but I can see the answer: 13 you don't recall the types of matters which 14 would have been discussed in these meetings? 15 A. It would have been primarily contact centre 16 focused, I would imagine. But it wasn't really 17 the -- as I recall, it wasn't really the forum 18 for raising of concern around, let's say, the 19 performance of the solution. If we'd have had 20 an incident that we couldn't answer with the 21 information that we had and the knowledge that 22 we had on the desk and we felt it was the 23 system, let me say, not working in the way we 24 had understood it should work, we'd have passed 25 that, at the time, down into the business 134 1 service management function, a second level of 2 incident management and problem management, for 3 them to investigate and determine whether or not 4 there was something deeper that needed Fujitsu 5 support and investigation. 6 It was very contact centre focused, was the 7 NBSC/HSH and the reviews. 8 Q. So when you said it would -- if there was 9 a problem -- you feared problem with the system, 10 it would be passed to a different team. Would 11 that be the Problem Management Team? 12 A. It would have been and if it would have gone in 13 a much more timely manner. It would have gone 14 at the time that the event occurred rather than, 15 you know, post-monthly review. 16 Q. Your evidence is, to the best of your 17 recollection, this review forum mainly concerned 18 maybe operational contact centre elements of 19 types -- 20 A. Volume of calls, types of calls, absolutely, to 21 see whether there was a need to produce 22 knowledge articles for the advisers in the NBSC, 23 or perhaps even suggest that there may be 24 reminders that need to go in the Counter News 25 article out to the wider network but not really 135 1 focused at all on the technology. 2 Q. Another purpose of this document was to 3 delineate the role, which calls would go to the 4 NBSC and which to the Horizon or the HSH, later 5 HSD. Can we turn to page 18, please, which 6 should be an appendix. Yes, that's right, it's 7 not formally an appendix but this is the table 8 I was looking at. We see here, on the left 9 column "Postmaster Incident", and the first two 10 are "[unable] to log on" and, if it was because 11 of a system failure or user error the Horizon 12 Service Helpdesk would deal with it but, if it 13 was a lost password, we see the third line, it 14 was the NBSC. So this was separating those 15 roles. 16 The last entry "Cannot use the Horizon 17 counter system or part of the system" refers to 18 matters such as -- sorry, we'll just wait for 19 that to come back on screen. Thank you. 20 The last entry, here we're talking about 21 monitor failure or equipment failure, which 22 results in the subpostmaster not being able to 23 use the system and that's clearly a helpdesk 24 issue, a Fujitsu issue. 25 A. Yes. 136 1 Q. Can we turn to page 20, please. The bottom two 2 entries, one says, "Has an EPOSS discrepancy" 3 and the next is a weekly one. That's referring 4 to an issue in the cash account when balancing, 5 isn't it? 6 A. Yes. 7 Q. In both cases, it says the first contact is the 8 Fujitsu helpdesk, and the comment says: 9 "HSH are responsible for assisting the PM in 10 the correction of the discrepancy. However, if 11 HSH cannot resolve an EPOSS discrepancy the PM 12 will be referred to NBSC for approval to accept 13 the discrepancy." 14 In both situations, the sole cause is listed 15 as "user error". Can you explain why this 16 document doesn't refer to what to do if there's 17 an EPOSS discrepancy caused by the Horizon IT 18 System? 19 A. No, I can't. All I can say at the time, and 20 this is a long time ago when we were first 21 automating the network, there was no belief or 22 understanding at my level, at my operational 23 level, at my team's level, that there was any 24 reason to distrust the technology. So, you 25 know, we were told quite clearly, in fact on 137 1 numerous occasions throughout my time, that 2 there were no Horizon integrity issues, and 3 there were no systemic issues. So at the time, 4 I think the stance would probably have been the 5 solution works as per Post Office's 6 requirements, as per the design, therefore there 7 wouldn't be -- such a situation arise, I'm 8 assuming that would have been the stance, 9 rightly or wrongly. 10 Q. At this time, can you recall who was saying or 11 telling you that the system was robust? So this 12 is in 2000? 13 A. I can't be specific but I'd have to say the 14 wider business. I can't remember the names of 15 the senior managers or the leaders of the 16 programme and the rollout, I'm afraid, at the 17 time, although I do remember Don Grey because 18 Don Grey was northeast regional office so that's 19 where I came across Don. I knew he had some 20 role to play within that. 21 But the general message was one, which I can 22 understand, even looking back, of trying to 23 ensure that, you know, operational people had 24 confidence to go about the processes that we'd 25 been -- that had been implemented and what we 138 1 were asked to do on a daily basis. 2 Q. You've mentioned Don Grey. Are you aware of 3 whether this message that Horizon was robust, 4 did that -- are you aware whether that came from 5 any higher than Don Grey? 6 A. I'm sorry, I couldn't say. 7 Q. You said as well in your evidence that it wasn't 8 just then that message was repeated. 9 A. Yes. 10 Q. Can you provide other examples? I know we're 11 jumping ahead but other examples of when you 12 were told that the system was robust? 13 A. Well, I guess as well there was a -- I suppose 14 I've got to try and convey the -- and give 15 context. So, you know, it was a massive 16 transformation for the Post Office from manual 17 to automation. Of course, we were all 18 understanding of it was that or Post Office 19 ceases to be relevant and probably exist. So we 20 understood the journey and the strategy. We 21 didn't go into it doubting it, I guess, is what 22 we were saying. We went into it accepting that 23 it was going to work. No reason to challenge it 24 at that stage. 25 Now, over time, evidence obviously started 139 1 to came in. We had improved knowledge, improved 2 experience over that same time frame, so, 3 therefore, we became much more comfortable with 4 our ability to question or challenge or 5 certainly escalate to Fujitsu to investigate 6 something when it wasn't working correctly. 7 Legacy Horizon, that was incredibly difficult 8 because of the, you know, the steep learning 9 curve that we were all on. Probably not brave 10 enough either, if I'm honest, looking back to 11 challenge some of the things when maybe we 12 weren't sure. 13 So I can think of, when it started to be in 14 the press, when there were postmasters who were 15 starting to take, I think, personal litigation, 16 then, I think, there were more frequent messages 17 to reaffirm that Horizon was robust, Horizon 18 integrity was there and there weren't issues. 19 I think there was also a third party -- I think 20 it was called Second Sight -- enquiry. Even 21 post that I remember the message being the same, 22 and I think there were also subsequent internal 23 inquiries undertaken, and I might be in HNG-X 24 chronology now, to also check -- and I think 25 that might have been done from my colleagues in 140 1 Product and Branch Accounting, I can't remember 2 the names of any of the individuals, I'm afraid. 3 And that also confirmed -- and those 4 messages were filtered down to the operational 5 teams. 6 Q. Actually, we'll come to that bit of the 7 chronology later on. Back to 2000, when this 8 document was being created and the message 9 was -- and I think you said the belief was that 10 the system was robust. 11 A. Yes. 12 Q. Do you accept that document is one of many which 13 is setting up support services that were to be 14 made available to subpostmasters using the 15 Horizon Helpdesk -- 16 A. Yes. 17 Q. -- or the NBSC? 18 A. Yes. 19 Q. Is it fair to say that those systems that were 20 put in place to provide assistance to them, were 21 built on an assumption that the Horizon System 22 was robust? 23 A. Correct. 24 Q. Moving, then, to knowledge sharing within the 25 NBSC. We're going to come to some examples of 141 1 problems later on. It's uncontroversial that 2 there were problems in the Horizon IT System. 3 As a matter of generality, how -- when someone 4 in the NBSC at the top became aware of 5 a problem, how was that shared amongst the other 6 members of the team? 7 A. Within the NBSC environment itself? 8 Q. Yes. 9 A. It was verbally cascaded. We also had bulletin 10 boards, or there was ad hoc, infrequent team 11 meetings where the information was provided. 12 Now, I can't remember the exact introduction but 13 I don't think it was at the very, very beginning 14 of the NBSC, but we also introduced knowledge 15 articles, and those knowledge articles 16 ultimately, over time, became, I would say, 17 mandatory in terms of their usage. 18 You had to use a knowledge article. You had 19 to associate the call that you'd had, the ticket 20 that you'd logged, with the knowledge article 21 that you had used to advise and guide the 22 caller. And there was a team of people set up 23 to produce those articles, manage those 24 articles, maintain those articles. So that was 25 the other method. 142 1 Now, naturally, if there was something fresh 2 or new came in that was slightly unexpected or 3 the timing wasn't great in terms of internal 4 communications, there would sometimes be a gap 5 between the Go Live of that, say, new 6 transaction or that knowledge, whatever it was, 7 and the creation of that knowledge article. So 8 that's where we fell back on the more manual 9 methodologies that I just mentioned, which would 10 really be ensuring people were informed word of 11 mouth, emails and bulletin boards. 12 Q. Do you recall when this change occurred, when it 13 was mandatory to rely on knowledge articles? 14 A. I think it was after I had gone down to the 15 business service management function and, in 16 fact, myself and one of my colleagues, Shaun 17 Turner, were involved in supporting the initial 18 set-up of a number of, you know, existing 19 processes in terms of the documentation. 20 Q. Could you just describe what a knowledge article 21 looks like? 22 A. It would literally be a Word document. It was 23 put on to the Remedy -- it sat independently but 24 accessed via what was the call log-in system, 25 the Remedy system. It could involve anything. 143 1 It could involve Counter News articles that were 2 literally just almost copy and pasted into 3 an electronic format. It could involve process 4 flow maps with, you know, swim lanes in terms of 5 start and finishing and different people might 6 need to be -- interact or involved. It could 7 involve diagrams. There were pictorial evidence 8 as well to show what screens looked like in 9 terms of trying to guide the NBSC agent and the 10 office. 11 Q. To what extent were the use of scripts used in 12 the NBSC, call scripts? 13 A. There may have been a script on occasion. 14 Normally, I would have said that would come off 15 the back of an unplanned event of some 16 description, something that's happened untoward 17 and, therefore, there would be a scripted 18 response given. I don't actually remember 19 a time where we were told quite, you know, 20 precisely to follow, almost word for word, 21 a script, but that would be the situation, where 22 we'd have something that was perhaps more 23 temporary in nature than the more permanent way 24 of responding to the enquiries. 25 Q. So perhaps in response to a major incident? 144 1 A. Perhaps response to a major incident yes, 2 something that's occurred that branches may call 3 in about and need to be aware of. 4 Q. Were you ever aware of subpostmasters being told 5 by members of the NBSC that they were the only 6 person experiencing problems using the Horizon 7 IT System? 8 A. No. I mean, that's not something I certainly 9 would have told my team. That sounds -- I'm 10 offering an opinion here but that sounds 11 slightly rogue to me, rather than anything that 12 would have been directed. 13 Q. That's internally to Post Office. Were you 14 aware of anything along the same lines being 15 communicated to people working on the Fujitsu 16 helpdesk? 17 A. No, I didn't really have a great deal of 18 visibility of their internal ways of working. 19 Q. I want to move now to your problem management 20 role, and please can we been up paragraph 5 of 21 your witness statement on page 2. I'll read the 22 last line and then go over the page. What you 23 say is: 24 "But what I do remember is that the role was 25 largely reactive in nature not particularly 145 1 predictive or preventative and therefore a lot 2 of the work that I would have undertaken 3 alongside my colleagues came from analysis of 4 the calls that had been received by the NBSC or 5 thematic incidents." 6 In general terms, what were the types of 7 problems that you were seeking to analyse from 8 these call records? 9 A. You were looking for any trend at all in terms 10 of a volume of call, a particular type of call, 11 business in nature, potentially technical in 12 nature and absolutely in relation to accounting. 13 Anything that might indicate there was a way of 14 us operating better, doing something better or 15 something potentially that needed fixing. 16 Q. When you say the role was not particularly 17 predictive, what do you mean by that? 18 A. I think -- again, I come back to a bit of 19 a chronology and a bit about the learning curve 20 and everyone being, you know, on the same 21 journey, very difficult back in the time -- in 22 the day, when I think about it. I guess I'm 23 trying to suggest that we probably had limited 24 capability of predicting what might happen and 25 we were learning through experience what could 146 1 happen. 2 So to start off with, the problem management 3 role in my time was very much just simply 4 reacting to what had already occurred. 5 Hopefully, we would learn from that lesson, 6 though, and increase and mature over time in 7 terms of the ability to look out for certain 8 things and prevent things from being repeated. 9 That's kind of what I mean about it being 10 largely reactive at the time I was the problem 11 manager. 12 Q. In your role as problem manager, would you have 13 benefited from more internal technical support, 14 technical IT support? 15 A. Oh, without doubt. I have no personal technical 16 background whatsoever: Post Office man and boy, 17 business training. Yes, I could answer 18 transactional queries and help postmasters, I'd 19 done the job. Technology-wise, no. I felt that 20 we were in a position where Fujitsu were the 21 chosen supplier. If we needed to go to Fujitsu, 22 I'd go to Fujitsu. Whatever Fujitsu came back 23 with, it was incredibly difficult to challenge, 24 if at all, if it was technical. 25 If it didn't feel right, if there was 147 1 something that wasn't sitting well with you, we 2 had a few strategic individuals in the business 3 that we could go to and ask them to have a look, 4 people like Ian Trundle and Bob Booth are two 5 people I know particularly supported me, not 6 necessarily with accounting issues but with 7 certainly large-scale geographic incidents that 8 we had, but no, we were very much trusting the 9 supplier for the technical knowledge. 10 Q. Just on that, please could we bring up page 5 of 11 your witness statement and the paragraph at the 12 top. I think this relates to evidence you just 13 gave. You say that: 14 "The Post Office IT function did have a team 15 of Business Relationship Managers run by Chris 16 Taylor that were technical." 17 A. Yes. 18 Q. Just to clarify, those are different, are they, 19 from the managers, the contract managers who -- 20 A. Oh yes. Yeah, yeah, those guys were all based 21 down in London. They were very much working 22 with the business from an innovation and 23 a future transactional perspective but all of 24 them naturally very technical in understanding 25 in terms of the Horizon System. So they weren't 148 1 there for us to be used on the process. You 2 won't find it written in a document, you know, 3 "Escalate to Chris or the team", but we did lean 4 on them when we felt it was necessary. 5 Q. Again, you refer to Ian Trundle and Bob Booth. 6 You say: 7 "Whilst also supportive, these 8 individuals/teams were primarily strategically 9 focused and not designed to be regularly engaged 10 in the operational day-to-day running." 11 What do you mean by "strategically focused"? 12 A. On next steps, on the future, on development of 13 software, it tended, as I say, to be linked to 14 my knowledge to business activity, rather than, 15 you know, focusing entirely on the technology. 16 But they were the people who would, from my 17 perspective, interact with Fujitsu on a very 18 regular basis in terms of how of the solution 19 worked. I'm sure they were both involved in 20 an awful lot of programme activity and an awful 21 lot of determining Post Office's requirements of 22 the technology. 23 Q. Did you ever request at the time for more IT, 24 internal IT support on the operational 25 day-to-day matters to be able to test what 149 1 Fujitsu was saying to you? 2 A. No, I don't believe that we did. Again, that 3 sort of stems from a belief of there were no 4 issues with the Horizon System. So back in HNG 5 or Horizon Legacy, I think you refer to it, it 6 was part of the learning curve again. So it 7 just didn't really necessarily occur at the 8 time. 9 Q. I'm going to come on now to some of the 10 processes involved in problem management and we 11 have a distinction between incident management 12 and problem management. 13 A. Yes. 14 Q. The Inquiry has heard evidence that incident 15 management relates to dealing with the symptoms 16 of a particular issue such as a server failure, 17 whereas the problem looks at the underlying 18 causes of the incident. Do you broadly agree 19 with that? 20 A. Yes. My role in incident management, we always 21 focused on service restoration and then, you 22 know, ideally we'd identify the root cause but, 23 if not, that was always part of problem 24 management's job once handed off. 25 Q. Please could we bring up the document 150 1 FUJ00079946. This is "[Post Office Account] 2 Customer Service Major Incident Escalation 3 Process". Is this something you would have 4 worked towards -- well, worked with when dealing 5 with major incidents? 6 A. Yes, I'm fairly sure, looking at the Post Office 7 distribution list, those two gentlemen were 8 either my line managers at one point in time or 9 line manager's line manager, so Dave Hulbert and 10 Richard Ashcroft. 11 Q. Can we turn to page 7, please. I should say 12 that document was dated 3 October 2006. Under 13 the heading "Process Objective", the fourth 14 bullet point says one of the objectives was to: 15 "Avoid unnecessary alerting of the 16 customer." 17 In this context, who was the customer: 18 a subpostmaster or the Post Office itself? 19 A. My reading of that would have been "clients" 20 would be another way of looking at it. So at 21 the time a major one was -- well, they were 22 Alliance & Leicester at the time, I think. So 23 I think it's referring to them rather than to 24 the branch network. 25 Q. I see. So the clients of the Post Office. 151 1 A. Yes. 2 Q. What do you understand then of the purpose of 3 "avoiding unnecessary alerting"? 4 A. It just doesn't sound right, does it, when 5 I read it today at all. I'd be guessing, to be 6 perfectly honest. I'm not entirely clear. It's 7 a bit ambiguous, isn't it? I don't know if it 8 was just purely wanting to protect brand. 9 Q. We have the last, and four bullet points up: 10 "Demonstrate to the Post Office a more 11 professional approach [and] 12 "Improved governance." 13 Had there been, before this document, what 14 was perceived to be a lack of a professional 15 approach from Fujitsu, from your perspective? 16 A. I don't actually know, if I'm honest, because 17 this -- is this 2006? Sorry. 18 Q. I think if we can just come just to double 19 check, if we can come out to the full page, 20 please. Thank you. Ah, sorry, 27 June 2005. 21 A. It's near enough, isn't it? No, if I'm being 22 honest, and it is recollection, no, I don't 23 remember there being poor governance. Perhaps 24 on occasions communication was inconsistent, if 25 we want to look at it that way. But I don't 152 1 remember there being governance issues, but it 2 might be a more appropriate question for David 3 Hulbert or Richard Ashcroft. 4 Q. Could we turn to page 8, please. The bullet 5 point at the bottom of this -- well, the 6 penultimate one now, says: 7 "The Fujitsu Service Delivery Manager (or 8 Duty Manager out of hours) is responsible for 9 communicating both up the Fujitsu Organisation 10 and across (see appendix A) to their counterpart 11 in POL." 12 Now, in your experience of dealing with 13 incident management, did you feel that Fujitsu 14 kept you or the Post Office appropriately aware 15 of any incidents as they arose? 16 A. Well, they certainly made us aware of the 17 incidents when they arose. It would be 18 difficult, I guess, for me to second-guess that 19 and challenge whether or not it was always done 20 in a timely manner or whether it was always 21 exactly what they knew at the time. It would 22 be -- the communication, basic communication, 23 ie would they call our duty manager out of hours 24 to tell us there was an issue? The answer's 25 yes. 153 1 Q. Well, at the time, do you recall having any 2 concerns that, to put it bluntly, you were being 3 kept in the dark about any incidents? 4 A. I don't have any particular evidence but, on 5 occasions, when you finally ended up with the 6 full root cause analysis and the full 7 documentation of the events, the timeline, 8 et cetera, there would be an ability to reflect 9 and think that it hadn't been your experience, 10 in terms of the timing of, in terms of the 11 quality of the information you were being 12 provided. 13 Q. When you say "on occasions", how often would 14 that be? 15 A. Rarely. Rarely. 16 Q. Could we please turn to page 9. Under "Incident 17 classification", it says: 18 "As a general rule a Major Incident will 19 always be an incident rated with severity level 20 A (critical) in the POA Customer Service 21 Incident Management Process Details document ... 22 However not all incidents rated as severity 23 level A qualify. This is because the severity 24 levels do not necessarily translate to the 25 global business impact on POL's business. For 154 1 example a single counter post office which is 2 unable to transact, regardless of its business 3 volumes, is rated as a severity A." 4 It then goes on to say: 5 "For simplicity, Incidents are classified 6 into three impact levels", and uses high, 7 medium, low. 8 Did an incident have to be a high impact to 9 be a major incident? 10 A. I think this is where -- I think the answer is 11 yes and no and sorry for that but the way 12 I remember the incident process management 13 working was it's based upon knowledge and 14 information at the time so it would be perfectly 15 reasonable for us on occasions to have what 16 might turn out to be false alarms that have been 17 raised through that way and a decision taken by 18 Fujitsu in the first instance that it was 19 a high, and therefore the process followed and 20 it might end up high and remain high 21 post-understanding what the impact was, 22 resolution and root cause. 23 Equally, you know -- or conversely, it could 24 work the other way -- it could turn out that 25 a decision was made by the Fujitsu duty manager 155 1 that the incident was medium or low, and it 2 might not have been rung through but, 3 subsequently, called through the following 4 morning, often triggered by calls from the 5 branch network, for example. 6 Q. If the incident wasn't declared a major incident 7 and just an incident, how did that affect the 8 way it was investigated? 9 A. It would be timing, more than anything, in terms 10 of waiting for the next -- you know, the 11 following working day, excluding weekends, as 12 well, so there could be some significant delay 13 if a wrong diagnosis was made in the first 14 instance. But that would never stop. 15 It's that thing about reacting and 16 predicting and preventing again. You get into 17 that kind of scenario where you might have been 18 able to look back and think we could have done 19 something sooner if it had been identified as 20 high and therefore communicated. 21 But branches, you know, if they were open 22 and they didn't have the ability to trade, they 23 would call into one of the two desks. That in 24 itself could also then trigger what would turn 25 out to be the major incident management 156 1 response. 2 Q. Looking at problem management, if we could go to 3 FUJ00079886. This is 23 December 2002, "Fujitsu 4 Services/Post Office Limited Interface Agreement 5 for the Problem Management Interface", which is 6 presumably a process you would have used as 7 a problem manager? 8 A. Yeah, I think that's about the right timing. 9 Yes. 10 Q. Can we turn to page 7 of this document and, 11 under 5.3, there is "Fujitsu Services specific 12 responsibilities". It says: 13 "Fujitsu Services will update the problem 14 Management Database daily (as problems are 15 updated). 16 "Fujitsu Services will provide POL with 17 remote access (via dial up) to the Fujitsu 18 Services problem management database." 19 Do you recall whether, in your time as 20 a problem manager, you had access to that 21 Fujitsu database? 22 A. Yes. I've got to say it wasn't the best 23 experience. I seem to remember we had two 24 machines that allowed us that dial-up access to 25 the tool, but it was a way of audit trailing and 157 1 recording updates to whatever the problem record 2 was. It wasn't -- you know, you had to get up, 3 if you like, and go sit there and go and look 4 for an update, though often, is how I felt about 5 the process rather than receipt of an email, 6 receipt of a phone call, there's an update in 7 the problem record or things like that. But 8 that might be my memory and just my 9 recollection. 10 Q. Did you consider that the information on 11 problems stored within it at the time was 12 adequate for your purposes as a problem manager? 13 A. I'm going to have to say yes but that is not 14 based upon an awful lot of recollection. You 15 know, I can't think of any particular instance, 16 for example, that would allow me to be more 17 specific but certainly, you know, the basic 18 intention was that the update would be provided 19 for and, you know, if it wasn't appropriate, 20 I certainly would have expected to have been 21 challenging it and/or escalating it. 22 Q. Was this another point when you were relying on 23 Fujitsu's technical expertise? 24 A. Oh, totally, yes. Yes. 25 Q. Please can we turn to page 12 of the same 158 1 document. Under section 11.1, it refers to a: 2 "... Cross Domain Problem Management Forum 3 [being] held monthly prior to the Service 4 Management Forum ... and is intended to 5 highlight and discuss all problems if time 6 allows." 7 Would you have attended that? 8 A. No, I don't think I did, actually. Because of 9 the age of the document, I was relatively new to 10 the team and junior, therefore. Some of names 11 on the front of the document I recognise, 12 I notice Stephen Potter's name there, for 13 example. He was a colleague of mine. He would 14 have been much more likely but I would have said 15 I do think it was Richard Ashcroft that was 16 managing this team at the time, and Bethany 17 Newton, that would have been their forum. 18 Q. The last sort of process I wanted to go over 19 which I've been asked to go over with you is the 20 branch issue management process. Can we please 21 turn up FUJ00080015. So this describes the 22 branch issue management process. Do you recall 23 the role of the Fujitsu branch issue management 24 process? 25 A. Only very high level. I certainly remember many 159 1 dealings with Nick Crow in particular. But for 2 me, the difference for this role was it was 3 meant to be much more proactive, from Fujitsu's 4 part, in regards to actively going out and 5 looking for issues. Again, I'm sure Nick would 6 use lots and lots of helpdesk data to guide him 7 and help him, but he would be very, very field 8 based in regards to his role. Network more than 9 accounting discrepancies being, what I remember, 10 are lots and lots of challenges with the various 11 network methodologies that we used at the time 12 and Nick was an expert in that area and 13 particularly assessing the geographical location 14 of the branch environment which could, on 15 occasions, cause issues with the connectivity. 16 Q. So just to clarify that, when you say "network" 17 rather than "accounting", does it mean, in your 18 experience, this process was looking more at, 19 say, issues where a rural branch may not be able 20 to connect to the servers rather than 21 a balancing discrepancy? 22 A. I think the scope would include both but my 23 recollection was that volume-wise would be much, 24 much more challenging and high numbers of 25 network connectivity issues than accounting. 160 1 Q. Can we turn to page 12 of the document, please. 2 At the top, the Proactive BIM process. It says: 3 "The BIM is also responsible for analysing 4 trends and anomalies experienced at Branch 5 level. The BIM will review the monthly 6 statistics, ie the Branch League Tables to 7 identify exceptionally high instances of call 8 numbers from Branches or other possible 9 indicators of potential issues." 10 Did you or anyone at Post Office have access 11 to the branch league tables? 12 A. Not that I recall. It might have been something 13 we'd have done similar from an NBSC perspective. 14 I don't actually remember having visibility of 15 Fujitsu's branch league table. That's not to 16 say that I didn't. I just can't really 17 remember. 18 Q. So these sort of issues, where there's problem 19 management, incident management and this 20 process, the BIM process, how did that work, or 21 the problem management team's work feed into the 22 NBSC and the advice that the members of that 23 team would provide? 24 A. Of course, all of the roles are supposed to 25 interact and share knowledge, share information, 161 1 update as I mentioned earlier, knowledge 2 articles, for example, ensure that end users on 3 the helpdesk were appropriately informed of 4 anything that was particularly important, out 5 there at the moment, major incidents. 6 Incident management was, for me, very much 7 what it sort of says on the tin: it was service 8 restoration, hand off, root cause, problem 9 management. 10 BIM was very much more, as I say, proactive 11 is how I thought about it but Nick and the team, 12 I feel, would interact with any relevant section 13 of the Post Office. So if what Nick had have 14 found or was asked to investigate was accounting 15 discrepancies, for example, it's just as likely 16 that he would have been involved with Product 17 and Branch Accounting within the Post Office as 18 he would be the business service management 19 function. 20 Q. But are you aware of a person within the Post 21 Office who was to draw all of that information 22 together to provide it to the NBSC to update the 23 knowledge articles, and things like that? 24 A. No, no. I'm sorry, I'm not. I think it would 25 have fallen to each functional area to ensure 162 1 that they were informing the NBSC of -- or 2 advising the NBSC at a later stage of the 3 benefits of a new knowledge article or updating 4 an existing. 5 Q. Was there a risk or did you think there was 6 a risk at the time that this knowledge wouldn't 7 be passed on to the NBSC without a co-ordinating 8 role? 9 A. I think on reflection, yes, there would have 10 been definitely been a risk that some 11 information may not have -- may have been lost 12 in translation or may not have been communicated 13 or articulated in a way that it meant to be. So 14 it's entirely possible. 15 Q. Are you aware if the training for members of the 16 NBSC was ever updated to take account of 17 findings or information gleaned through the 18 problem management process? 19 A. I'm not. Post my time, I believe I'm correct in 20 this, the NBSC was, in effect, outsourced, 21 albeit to Royal Mail. They were the owners of 22 Dearne House, they provided a range of contact 23 centre services to other clients of theirs, and 24 were perceived as contact centre experts. So 25 once that was relinquished, if you like, as 163 1 a service, they, in effect, became another 2 supplier to the Post Office and I think they 3 were managed accordingly. I can't comment on 4 Post Office's contribution to training for those 5 agents post that time. 6 Q. In your recollection, when did that transfer 7 occur? 8 A. It would be a wild guess, I'm afraid, but I'm 9 going to say around about, I think, 2006/7 might 10 be the time frame. 11 Q. The same questions really about these findings 12 that problem management would make or incident 13 management would make. How were those 14 findings -- let me rephrase that, sorry. 15 Was there one person who would be 16 responsible for cascading that information to 17 the people responsible for prosecuting 18 subpostmasters? 19 A. I don't know the name or whether they would but 20 I wouldn't have expected it to have been the 21 individual problem managers. I would have 22 expected it to have been the leadership of that 23 function. 24 Q. When you were a problem manager, were you ever 25 approached by anyone in the team responsible for 164 1 prosecuting subpostmasters to provide 2 information on potential problems in the Horizon 3 IT System? 4 A. I remember receiving requests to investigate 5 and, therefore, engage Fujitsu, gather evidence 6 potentially from the branch in regards to, you 7 know, trial balances, et cetera, transaction 8 logs and relay that information. So I certainly 9 remember having those on occasion. I do, from 10 perception, believe that they were quite low in 11 comparison to the other types of problems that 12 we handled at the time. But I don't then have 13 any recollection, I'm afraid, at all about what 14 happened post the conclusion of that problem, 15 particularly if it was inconclusive, perhaps, 16 rather than a very black and white response from 17 Fujitsu that declared that the system was 18 working as per design and, therefore, there was 19 no issue. 20 Those would go back to, I don't know who 21 would send them, but they would go back to and 22 be due to go back to the appropriate retail 23 network manager. 24 Q. So you might have -- your evidence is you might 25 have or you remember someone involved in 165 1 prosecutions asking you about particular cases. 2 Do you have any recollection of a more proactive 3 approach of someone in the prosecution 4 department saying, "Can you provide us 5 information generally on problems in the Horizon 6 IT System?" 7 A. Not that I can remember but all of those types 8 of enquiries would come from internal teams and 9 that, for me, from my recollection, was the 10 Retail Network Manager or the Field Intervention 11 Officers, I think, some of them were called at 12 a later date. So they would be the people out 13 in the field who would have that day-to-day 14 relationship and interaction with the branches. 15 They may come to us for -- to request 16 Fujitsu support and, in those examples, once 17 we'd a conclusion of the investigation, it would 18 go back from whence it came, so back to the 19 Retail Network Manager for consideration. 20 I think it's fair to say that, if it wasn't 21 the answer that we'd all hoped for, in the sense 22 that we'd found something that would explain 23 whatever was occurring, that the retail network 24 manager may have needed to then engage other 25 parts of the business, again including Product 166 1 and Branch Accounting who, you know, naturally 2 had visibility of the general accounting 3 procedures, the records for every single branch 4 in the network. 5 Q. Again, I apologise if I misheard this but 6 I asked about someone in the prosecution team 7 asking you directly for that broader 8 information -- 9 A. No. 10 Q. -- I think you said no. You then pointed to 11 Retail Network Managers. 12 A. Yes. 13 Q. I think again you gave examples of when it would 14 be specific cases that they would ask you about. 15 Do you recall the Retail Network Managers ever 16 asking you proactively and broadly about general 17 problems in the Horizon IT System? 18 A. Not proactively. It would be off the back of 19 them being approached by a branch or them 20 noticing a trend within a branch. Sorry, if 21 I've understood that question, they would react 22 to the knowledge and them becoming aware of 23 a situation that they felt warranted further 24 investigation. 25 MR STEVENS: Sir, that might be a suitable time to 167 1 take a break. 2 SIR WYN WILLIAMS: Let me just unmute myself to say 3 yes. 4 MR STEVENS: Excellent, sir, if we could say 3.05? 5 SIR WYN WILLIAMS: Fine. 6 (2.55 pm) 7 (A short break) 8 (3.05 pm) 9 MR STEVENS: Mr Blackburn, we've talked to about 10 general matters, I now want to get into some of 11 the specifics and the first one is a major 12 incident on 9 May 2005. Please could I ask for 13 POL00091917 to be brought up. This is the major 14 incident report and we can see in the external 15 distribution list, which is at the bottom, that 16 you were included in this. Does that mean it's 17 an incident that you would have dealt with or 18 was it just for information only? 19 A. No, it's quite possible that I would have played 20 a major incident management role during the 21 event itself. 22 Q. Do you recall this incident at all? 23 A. I'm sorry, I don't. Unfortunately, there were 24 quite a few, from memory. 25 Q. Could we turn to page 6, please. This says 168 1 that -- the "Introduction" says that the: 2 "... document details the initial incident 3 that occurred within the live estate between 4 09.00 and 10.00 for a four minute period on 5 Monday 9th May 2005 and the activities that were 6 carried out for the remainder of the business 7 day by the appropriate POA Service Management 8 and support teams." 9 If we go to page 8 now, please. Under the 10 "Description of the fault", I'll read this for 11 the record. It says: 12 "The incident that occurred presently hasn't 13 a determined root cause and occurred within the 14 live ate whilst the SSC word analysing the 15 Correspondence Server volume capacity using the 16 RiposteVolume Command. The purpose of this 17 activity was to ensure that there available 18 spare disk capacity across the Correspondence 19 Server disk volumes. 20 "The work activity was carried out as 21 a precautionary measure as a result of 22 an archiving job not completing following 23 an event storm that occurred the previous week. 24 "The RiposteVolume command has been used on 25 numerous occasions before with no impact to 169 1 service, as it simply displays the details of 2 Riposte Volumes. 3 "Running this command stops Riposte services 4 running for a micro-second and then starts 5 (unlocks) the services again after the volume is 6 taken. It is believed that in this instance, 7 due to a bug in Riposte, the services did not 8 start again. A PEAK [with the number there] has 9 been raised with Escher development." 10 Would you accept this appeared to be 11 a significant problem, in that it's referring to 12 a bug in the Riposte code? 13 A. Yes. 14 Q. Did this concern you at the time? 15 A. Yes. 16 Q. Is it right that, in your evidence, that these 17 types of incidents would not be -- the outcomes 18 of them would not be reported back to the branch 19 network? 20 A. I certainly wouldn't have reported them back to 21 the branch network. It's unlikely, I would have 22 said, but I can't remember the particular event. 23 There may have often been other circumstances 24 where it was a necessity, because, you know, we 25 couldn't perhaps establish which branches were 170 1 impacted to what degree, so we may have 2 understood the impact, but not been able to 3 identify it was branch A, B and C, for example, 4 and therefore you may have been on alert for 5 waiting for a call. And that's the kind of 6 thing that we referred to earlier that might 7 have generated a script on the Network Business 8 Support Centre, to capture that information and 9 then make sure it's processed accordingly, 10 ie a correction, if a correction was required, 11 is applied. 12 But I'm sorry, I really don't remember the 13 detail of this particular incident. 14 Q. Can we turn to paragraph 13 -- sorry, page 13. 15 My apologies. Under the table, the first three 16 paragraphs refer to "E Top Up Transaction 17 failures", which presumably relates to mobile 18 phone top-ups? 19 A. Yes. 20 Q. "... displayed in the table above occurred due 21 to the interface between Fujitsu and Epay. When 22 authorised transactions timeout at the counter 23 a reversal is automatically generated by the 24 counter. Transactions were timing out due to 25 the correspondence server problems on this day. 171 1 Epay's systems have to match the reversal to the 2 original authorisation request. The reversal 3 has to get to Epay within a time limit of 10 4 minutes." 5 Then the paragraph, skip that paragraph. 6 The next one is: 7 "Because of the problems with the 8 Correspondence server replication on this day, 9 a number of these reversals did not get through 10 within the time limits." 11 Do you think because of that, because of 12 reversals not being to be made, that could have 13 led to discrepancies within accounts -- 14 A. Yes. 15 Q. -- and this would affect branches? 16 A. Yes. 17 Q. But if we now turn back to your witness 18 statement, please, at paragraph 5, page 3 of 19 your statement, in the middle, you're referring 20 to a different incident but say: 21 "Post incident findings such as March 2005 22 were not shared with the branch network to my 23 knowledge ..." 24 A. Yes. 25 Q. So you said you don't have knowledge of this 172 1 being shared. 2 A. Correct. 3 Q. Is there anyone else who would have been dealing 4 with this at the time who might have shared it 5 without your knowledge? 6 A. Not directly, no. I don't have any recollection 7 of it being communicated. At least what 8 I probably alluded to just a minute ago, in 9 regards to there'd have been a number of people 10 NBSC, HSH included, who would be looking out for 11 calls. I think this is where Product and Branch 12 Accounting would have played a large role as 13 well, in regards to observing the accounts that 14 were being completed. They would have had the 15 ability to focus in on E top-up transactions as 16 an example and see whether there were 17 discrepancies that I guess, in theory, could 18 align with that event. 19 Q. If a subpostmaster had a discrepancy when they 20 came to balance, that may be because of 21 a transaction that occurred some time before? 22 A. Yes. 23 Q. So when they come to balance, some time has 24 passed, and it might be that a subpostmaster 25 simply accepts a discrepancy as an error? 173 1 A. As has always been problem, I think, yes, often 2 the value would be a trigger. Large values, you 3 know, would likely be reported. Low values may 4 well have just been accepted without the 5 knowledge, yes. 6 Q. What's the justification for the general policy 7 that incidents such as this wouldn't have been 8 shared with the branch network, so that the SPMs 9 could be aware of potential errors in the 10 system? 11 A. I couldn't comment on what the policy was or the 12 reasoning beyond, you know, operational role for 13 me, restore service, and the belief that there 14 were no issues that were generated with the 15 solution. Clearly, this was a major incident 16 not necessarily software related, per se. It 17 was something unintended that occurred that 18 generated a discrepancy. I'd have expected 19 Product and Branch Accounting to have cleaned up 20 this particular situation but can't say 21 categorically that was what occurred. 22 Q. So is it your evidence that the communication -- 23 the responsibility for the communication lay 24 elsewhere? 25 A. It -- exactly that, yes. 174 1 Q. I want to now look at what's been called the 2 Callendar Square bug. Please can we turn to 3 POL00028984 and page 10. At the bottom there's 4 an email from Sandra MacKay. Do you remember 5 who she was? 6 A. I'm sorry, no, I don't. 7 Q. The project is "Callendar Square", which was 8 a sub post office and the first line says -- 9 this is to Shaun Turner, I should say, sorry: 10 "You may recall that in September the above 11 office had major problems with their Horizon 12 system relating to transfers between stock 13 units." 14 Go over the page. 15 "The SPMR has reported that he is again 16 experiencing problems with transfers ... which 17 resulted in a loss of around [£43,000] which has 18 subsequently rectified itself. I know that the 19 SPMR has reported this to Horizon Support, who 20 have come back to him stating they cannot find 21 any problem." 22 Firstly, did you have any dealings with the 23 Callendar Square branch in September? 24 A. No, my dealing with this branch was triggered by 25 Shaun Turner reaching out to me. 175 1 Q. We'll get to there in a moment. Could we go to 2 page 10, please. Just to follow the trail, we 3 have -- we had Sandra MacKay's email at the 4 bottom. We then have an email to Shaun Turner 5 from Brian Trotter. Do you remember who Brian 6 Trotter was? 7 A. I'm going to say retail network management but 8 I couldn't say specifically -- oh he's contract 9 and service manager there, isn't he? But yes, 10 I do recognise the name. 11 Q. It says: 12 "... I visited the branch with Sandra last 13 week and the SPMR provided clear documented 14 evidence that something very wrong is occurring 15 with some of the processors when carrying out 16 transfers between stock units." 17 If we go over the page to page 9, we see 18 what you just referred to as your involvement. 19 It's an email from Shaun Turner dated 6 January 20 2006. He asked for your advice on this branch. 21 Why would Shaun Turner have come to you for this 22 particular issue? 23 A. Purely the interaction and relationship with 24 Fujitsu, so we were an escalation team. We 25 weren't outwardly facing to branches. We were 176 1 a level 2, if you like. And the Retail Network 2 Management, all of them had the opportunity to 3 raise escalations with us for us to engage 4 Fujitsu and get them to investigate a particular 5 occurrence. And that's what happened on this 6 occasion. 7 Q. Indeed, at the bottom, he describes a problem 8 and his last sentence is: 9 "I am concerned that there is a fundamental 10 flaw with the branches configuration, and would 11 be interested to know how Fujitsu Services put 12 the first issue to bed." 13 At this point in time, did you think this 14 was an issue that was affecting a single branch 15 or a wider issue? 16 A. Single. 17 Q. If we go up the page, I think it's to the next 18 page, we'll see you email Liz Evans-Jones. Do 19 you remember who she was? 20 A. Yeah, she was a higher level than me. In 21 effect, she was my line manager's opposite. So 22 I went straight to an escalated level within 23 Fujitsu rather than go to the people I would 24 ordinarily engage with. 25 Q. Why did you decide to escalate it? 177 1 A. Shaun was somebody I'd got the greatest respect 2 for and trust for, I'd known him a very long 3 time. His email was casting enough concern for 4 me to do this on this occasion. 5 Q. The £43,000 discrepancy, that's a large 6 discrepancy? 7 A. Yes, I mean obviously it is ridiculously large. 8 Having said that, numbers are everywhere in the 9 Post Office, and so it isn't necessarily 10 a trigger. It was really the fact that Sean was 11 coming to me. Number 1, he was escalating to me 12 which meant it was important and, number 2, it 13 just looked like something on the basis of his 14 knowledge, his experience, that he was saying 15 the system not working as it was intended. So 16 I went straight to an escalated level on that 17 occasion. 18 Q. Liz Evans-Jones's response is immediately above 19 this. 20 A. Yeah. 21 Q. She says: 22 "I have checked the call and this issue is 23 schedule to be resolved in S90." 24 What was your understanding of S90 at that 25 point? 178 1 A. A new software release that the business was 2 working on. It would have contained many 3 things. I am assuming there are lots of 4 business related reasons for a new software 5 release but it would also have contained 6 a backlog of fixes for any bugs or defects that 7 had been found. I couldn't say on over what 8 time frame but, for me, I do remember S90 being 9 very important. I do seem to remember a lot of 10 communication about its need being rolled out to 11 the branch network. 12 Q. At page 7, if we could go up, we see that you 13 pass this on to Shaun Turner and then Shaun 14 Turner goes on to reply to you. 15 A. Yes. 16 Q. If we carry on up. Thank you. This is 17 17 February now. The first question -- Shaun 18 Turner asks three questions, the first of which 19 says: 20 "Do we understand why this particular branch 21 has been having problems? Or are there other 22 branches in the network that have been having 23 this problem?" 24 Now at this stage, had your view changed on 25 from it being a single branch issue to having 179 1 concerns with it being a wider issue or not? 2 A. It wasn't actually, no. So whilst I'd taken 3 from Liz Melrose-Jones (sic) response that 4 clearly Fujitsu were aware of it and/or a risk 5 of it, this was my first experience of it 6 actually occurring within the network. So at 7 this point in time, clearly I was aware now that 8 there was a risk of others being impacted that 9 this was the only branch specifically that I was 10 aware of. 11 Q. If we look at page 3 on this email chain we see 12 on 23 February 2006, there's an email from Anne 13 Chambers to Mike Stewart, which you were 14 subsequently sent this email. 15 A. Yes. 16 Q. You would have read it at the time? 17 A. Yes. 18 Q. The second paragraph says: 19 "Haven't looked at the recent evidence, but 20 I know that in the past this site had hit this 21 Riposte lock problem 2 or 3 times within a few 22 weeks. This problem has been around for years 23 and affects a number of sites most weeks, and 24 finally Escher say they have done something 25 about it." 180 1 What did this mean to you when you read it? 2 A. News. It's not something -- you know, I do 3 actually -- when I've looked at all the evidence 4 you've provided, whilst I couldn't remember the 5 detail until I went back through the evidence, 6 Callendar Square was a name and a Post Office 7 that I remembered immediately because this was 8 long running from my involvement, but that, to 9 me, that statement was, I believe at the time, 10 news to me. Not something I was aware of that 11 branches could have this issue from Smart post 12 transactional problems. 13 Q. Would this problem concern transfers from the 14 stock units, Callendar Square? 15 A. I'd understood it was -- yes, but I thought I'd 16 understood it was triggered by a particular 17 transaction. I may be incorrect on that. 18 Q. But in terms of the problem itself, the fact 19 that it's saying it's been around for years and 20 it affects a number of sites was weeks, did you 21 believe that was significant? 22 A. It's shocking. Yes, shocking. If -- especially 23 on the basis of, as you could see from Liz's 24 response, that if there was an awareness of it, 25 there wasn't enough awareness of it, and it had 181 1 been parked, obviously, for quite a while in the 2 backlog, if we were only going to look to fix it 3 in the S90 software release. 4 Q. Did you have any concerns at the fact that, 5 where we went the first email we went to, 6 Callendar Square branch were initially told that 7 there was no problem in the system? 8 A. I wouldn't say concerned until you get, you 9 know, you get the rest of the detail and then 10 you understand and then you can reflect back and 11 go, yes, you would be concerned. I think it 12 points to my earlier point around having to 13 trust, you know, the expertise in terms of the 14 advice and guidance that had been given. 15 But, again, I'll also say this was my first 16 awareness of this particular problem and this 17 particular branch experiencing this problem. 18 What had gone before, I couldn't say whether 19 that was just calls logged into the Horizon 20 Service Desk and how those had been resolved and 21 managed. 22 Q. The last paragraph of this email says: 23 "Please note that KELs tell SMC that they 24 must contact sites and warn them of balancing 25 problems if they notice the event storms caused 182 1 by the held lock, and advise them to reboot the 2 affect counter before continuing with the 3 balance. Unfortunately in practice it seems to 4 take SMC several hours to notice these storms, 5 by which time the damage may have been done." 6 Do you know what this refers to when it's 7 talking about "event storms" or what was your 8 understanding at the time? 9 A. At the time, that would have been -- as I've 10 just said, my first reading of it, it would be 11 a question for Fujitsu, I'm afraid, as to what 12 an "event storm" actually is or contains. But 13 it clearly shows that Fujitsu were aware of it, 14 they'd created a Known Error Log and were 15 managing the calls, at least reactively, but 16 I don't equally remember any proactive 17 communication on the -- from when this was 18 originally identified. 19 Q. Did you discuss this issue with anyone within 20 Post Office once it came to your attention? 21 A. No, but I think not long after this particular 22 email, I think more evidence became available to 23 me, I think three or four, maybe five post 24 offices having the same problem, for the same 25 reasons. At that point, I'd made sure that it 183 1 was escalated into our problem management 2 function to get greater visibility and 3 awareness. Prior to that additional set of 4 branches, though, I'll be honest and say that 5 I'd accepted Liz's position, in that S90 was the 6 fix. 7 Q. In any of the conversations you'd had with 8 anyone in Post Office about this, were you aware 9 of anyone in Post Office who was aware of this 10 problem before -- 11 A. No. Personally, no. I find it had to sit here 12 and believe that no one did know but personally, 13 no, I didn't know. 14 Q. What steps were taken in respect of 15 communicating this issue or this bug to teams 16 outside of Problem Management? 17 A. I can't say, I'm afraid, what I can say is 18 I didn't communicate it further than Problem 19 Management. What happened to it beyond then, 20 I'm sorry, I couldn't say, I can't remember. 21 Q. Do you know who was responsible for 22 communicating that outside of Problem 23 Management. 24 A. Someone somewhere in the business at 25 a reasonable level would have had to have made 184 1 a decision as to whether it was communicated in 2 its entirety and a warning provided to branches 3 aligned with that Fujitsu knowledge article, for 4 example, which seems to be sat, as I'm saying, 5 reactively waiting for a storm to occur and then 6 a response. I'm not sure, though, that a branch 7 would have had the ability to have noticed or 8 known what a storm was, or looked like or felt 9 like, so that, you know, may have been the most 10 appropriate response. 11 Q. So you say you pass it on to the knowledge -- 12 the Problem Management Team. 13 A. Mm-hm. 14 Q. Are you aware of this issue being known about by 15 other members of the Post Office outside of that 16 Problem Management Team? 17 A. No, as I say, it was news to me so -- 18 Q. Sorry, after you'd discovered it? 19 A. After I'd discovered it, no, I don't know 20 whether -- I can't remember what or who they 21 communicated with post raising of the problem 22 record but I would suspect, given that clearly 23 there was already awareness of it within 24 Fujitsu, and the fix cued to go out within S90, 25 that the process that's described in there just 185 1 continued up to rollout of S90. 2 Q. Do you think that subpostmasters should have 3 been made aware of this issue? 4 A. Ideally, yes. As I say, I think the trigger is 5 worrying me somewhat because I don't have 6 an understanding of what "event storm" means or 7 whether it would be obvious to somebody at the 8 branch counter end but, ideally, I think, you 9 know, as a basic principle, all branches should 10 have been aware of any particular defects or 11 issues that could have affected accounting. 12 Q. Were you involved with the handling the problem 13 after initially discovering it? 14 A. I don't think I was, actually, no. 15 Q. Did it influence the way you thought about the 16 Horizon IT System thereafter? 17 A. I think not only this but, at that time, others 18 were now already more prevalent, probably for 19 any number of reasons, just, you know, 20 knowledge, understanding, the way that people 21 were then more confident in reporting and 22 reported them through the procedures. There was 23 certainly an upward trend in regards to any 24 number of instances where people believed 25 Horizon may have caused the discrepancy. 186 1 Q. I want to move on to a different issue. Please 2 can we open FUJ00121072. This is an email to 3 you from Gareth Jenkins. Did you work with 4 Gareth Jenkins often? 5 A. No. Gareth was sort of development level from 6 what I remember, deeply technical. It was 7 a rarity for Gareth to reach out to me, 8 particularly if he's copied in Mike Stewart, 9 which would be much more likely my point of 10 contact, occasionally Anne. 11 Q. So the date is in the American form but it is 12 sent on 13 February 2007 -- 13 A. Yeah. 14 Q. -- and it attaches a document called "Rem 15 Misbalance". If we could look at that now it's 16 FUJ00121073. Do you recall receiving this 17 document at the time? 18 A. I don't recall, no. Sorry. 19 Q. It describes: 20 "... a serious bug introduced into Live that 21 can result in accounts misbalancing. 22 "This bug was introduced as part of 23 LFS_COUNTER 35_6 which went to a limited number 24 of branches for a pilot from 4/2/07 to 11/2/07 25 and then to the whole estate on 12/02/07." 187 1 It goes on to describe the history of the 2 issue. If you could go down, please, to "Basic 3 Problem Effects". It says: 4 "The basic problem is when Rem Outs ..." 5 Can you help us with what rem outs are? 6 A. Sending out remittances, probably of cash rather 7 than stock, back to cash centres or back to the 8 Swindon Stock Centre. 9 Q. So it says -- 10 A. Surplus. 11 Q. Oh, sorry? 12 A. Sorry, surplus. 13 Q. "... incorrect transactions are recorded. 14 Specifically some of the Rem transactions are 15 missing. 16 "Take an example of 2 £500 coin bags being 17 Remitted out, then the following was recorded 18 ..." 19 Then if we look at that table, effectively 20 it's showing the cash out on the top line, the 21 rem out cash is 500 when it should be 1,000. 22 A. Yeah. 23 Q. It says below: 24 "The Stock Unit will show a £500 excess cash 25 and a Receipts and Payments mismatch." 188 1 If we can go over the page please: 2 "Some branches have made a further Rem of 3 £500 in an attempt to correct the situation. If 4 this Pouch is subsequently Despatched, then this 5 should result in the branch accounts being 6 correct. 7 "Note that this will result in further 8 incorrect data being sent to [Post Office 9 Limited] FS see discussion in section 3.2. 10 "However, if this Pouch is not Despatched, 11 then the Stock Unit will balance and report 12 correctly but the £500 will remain Stuck in 13 Suspense." 14 It goes on to say: 15 "It is recommended that all Branches are 16 advised to do such a Dummy Rem and to Despatch 17 the pouch to ensure that the Branch Accounts are 18 clear." 19 Can you just in broad terms explain what 20 that advice is, the dummy rem? 21 A. Creation of a false rem, a remittance amount, to 22 trigger the balance accounting being accurate, 23 but it's something that it doesn't exist in 24 terms of, I think, the example £500. 25 Q. So, in effect, redo the transaction -- 189 1 A. Yes. 2 Q. -- without sending the cash so that it balances? 3 A. Correct. 4 Q. It says: 5 "It has been agreed that POA SSC will 6 contact each branch with detailed advice as 7 exactly what to do." 8 Please can we now go to the technical 9 appendix of the Horizon Issues judgment. It's 10 POL00022841, page 52. This is the judgment of 11 Mr Justice Fraser in the Group Litigation Order 12 where the judge was looking at bugs, errors and 13 defects in the Horizon IT System from which this 14 Inquiry is building on the findings. 15 A. Okay. 16 Q. Section 6 refers to a remming out bug. If we 17 could go over the page to 201, please, it says: 18 "Issue 6(i) arises as follows. What is 19 called 'a remming error' leads to a mismatch 20 between the amounts of cash remmed out to one 21 place and the amounts remmed in from another. 22 The Post Office has submitted that remming 23 errors are a clear violation of Data Entry 24 Accounting and are picked up by Horizon. The 25 two different issues are as follows." 190 1 It says: 2 "As the obverse of the coin of remming in, 3 SPMs rem out pouches of cash to be returned to 4 the Post Office Cash Centre. A single pouch may 5 contain multiple bags of coins or cash and each 6 bag can only hold one denomination, and there is 7 a limit as to how much cash can be placed into 8 a pouch. The cash can be remmed out before it 9 is physically collected. When remmed out, the 10 cash appears in a different line in the branch 11 accounts. On collection, the collection team 12 scan a barcode on the pouch and the cash is 13 removed from the 'cash in pouches' line of 14 accounts." 15 It says: 16 "When remming out, branches should have made 17 one entry for each denomination and value and, 18 if there were multiple bags for a particular 19 denomination, the quantity of bags should have 20 been specified in that single entry ..." 21 So two times £500 of £2 coins. 22 "However, if the SPMs had made multiple 23 entries for each denomination and value (eg one 24 entry for 1 x £500 bag of £2 coins and a second 25 entry for 1 x £500 bag of £2 coins), Horizon 191 1 would only record the first bag as having left 2 the branch's cash holdings, but all of the bags 3 would show on the 'cash in pouches' line. This 4 would have created a discrepancy in the branch 5 accounts because all of the cash would have been 6 collected." 7 Is this the bug that you were dealing with, 8 with Gareth Jenkins in February. 9 A. I can't be absolutely certain but it's certainly 10 reads and appears to be, yes. 11 Q. So, in essence, the problem in lay terms is two 12 pouches, say, of £500 are sent to the cash 13 centre but Horizon only logs that one has been 14 sent? 15 A. Yes. 16 Q. So Horizon records that there is an additional 17 £500, which isn't actually there? 18 A. Correct. 19 Q. Can we please go to FUJ00120587. This is 20 a Known Error Log, "KEL acha508S" and you see 21 it's created on 12 February 2007 or raised then, 22 so at the time this was made this issue was 23 being dealt with, and last updated on 24 15 February 2007. We see from the "Symptoms" 25 that we are referring here to this bug. 192 1 A. Yes. 2 Q. Please can we go down to the bottom section. It 3 says: 4 "Calls about inconsistencies in stock rem 5 outs should be redirected to NBSC. SSC have 6 contacted all branches who have had a problem 7 with cash rem outs, quoting ref PC143435." 8 Is it the position then at this stage, on 9 15 February, the SSC had called the branches or 10 contacted the branches where there was 11 a discrepancy arising from this bug and sought 12 to deal with it. 13 A. That's how I'm reading that, yes, that they've 14 called the ones that have been identified. 15 Q. Please can we turn now to FUJ00121071 and turn 16 to page 3 at the very bottom. We see your name 17 at the bottom. 18 A. Yes. 19 Q. An email on 15 February. If we can go over the 20 page to see the content of that email. The 21 subject is "T30 Release -- Impact on Stock 22 Rems -- Monday 12th February". You say: 23 "Scenario: 24 "Whilst the T30 release was out in the 25 branch network, Monday 12th Feb only for all bar 193 1 120 branches that were incommunicado during the 2 initial regression progress. We have 3 a potential situation where a branch completing 4 the stock rem out [so that's the cash going out 5 to the cash centre] on that day could have 6 a discrepancy due to the fact that not all stock 7 physically returned by the branch may have been 8 deducted from Horizon from the stock on hand 9 table, this despite the fact that the rem slip 10 produced matched the physical stock returned." 11 You then say under "Latest position": 12 "We have a possible 570 branches that were 13 affected by last weekend's T30 release, now 14 regressed. 15 "The pouch IDs have been identified by 16 Fujitsu. 17 "The branches who sent those pouches have 18 also been identified, and the value of the stock 19 returned from each has been established from the 20 POC file." 21 Do you know what the POC file is? 22 A. I'm sorry, I don't remember. 23 Q. "We still don't know which of these branches has 24 actually got a discrepancy due to the problem." 25 So these are potentially affected branches, 194 1 unknown if there's a discrepancy? 2 A. Yes. 3 Q. In terms of "Next Steps" you say: 4 "Identify which of those branches have 5 a discrepancy between Horizon and the rem 6 slip -- Options here are to contact the branches 7 or wait and react to calls made into NBSC, 8 HSD -- I will then arrange a conference call to 9 discuss way forward tomorrow ..." 10 So is it fair to say that these 570 branches 11 are different to the branches the SSC had 12 already contacted in the KEL we'd referred to? 13 A. I can't actually remember but that's the way 14 I am reading this yes, that they're additional. 15 Q. In respect of these branches, there's 16 essentially two options, is it fair: one is 17 a proactive approach of contacting those 18 branches to advise them of the issue -- 19 A. Yes. 20 Q. -- and the second is to wait for calls to come 21 into the NBSC; do you accept that? 22 A. Yes. 23 Q. Thank you. Now, at this stage, the 24 subpostmasters again would have no idea that 25 there was a bug in the system, would they? 195 1 A. Correct. 2 Q. A subpostmaster would only be aware of the issue 3 in balancing when they came to balance the stock 4 unit? 5 A. Correct. 6 Q. Again, this could be some time after remming 7 out? 8 A. It could. 9 Q. So, again, it's another example of when 10 a subpostmaster may have thought that 11 discrepancy on balancing was caused by 12 a mistake? 13 A. Yes. 14 Q. They had no reason to believe, if they're not 15 told, that it was due to a bug, error or defect 16 in the Horizon IT System? 17 A. That's also correct. 18 Q. Please can we go to page 3. It's quite 19 confusing the way this email is set out, but you 20 see halfway down there's an email from Dave 21 Hulbert, 5 February 2007 -- 22 A. Yeah. 23 Q. -- forwarding on your email to Andy McLean. 24 Just starting out with some identities, who was 25 Dave Hulbert? 196 1 A. My line manager at the time. 2 Q. And Andy McLean? 3 A. His line manager. 4 Q. It says: 5 "Andy 6 "See update below. 7 "Not much progress today. The dilemma for 8 Gary is approaching branches is proactive but 9 opens the risk of litigation in future ie we're 10 telling 570 branches that Horizon may have 11 caused a discrepancy -- low risk but a risk -- 12 being reactive doesn't feel right as we've 13 caused the problem for branches but this may be 14 the right option in this situation." 15 To what extent was the risk of litigation 16 taken into account at the Post Office at this 17 time when deciding how to handle a known bug, 18 error or defect? 19 A. It's possibly more of a question for David and 20 Andy. My reason for escalating this, obviously 21 beyond the seriousness of the situation, is just 22 a growing awareness of the volume of defect or 23 bug related calls and enquiries that we were 24 undertaking. But also very aware that the 25 mantra and message was still "There are no 197 1 Horizon integrity issues and there are no 2 systemic issues that cause problems", and yet 3 here's clearly an example, whether driven by 4 human error or not, that was creating 5 discrepancies in branch. 6 Q. But in this case -- it's right to say in this 7 case there was a known error -- 8 A. There was, in this instance a known error, yes. 9 Q. -- that caused discrepancies -- 10 A. Yes. 11 Q. -- and your line manager is suggesting here of 12 not taking a proactive approach, and one of the 13 reasons given is the risk of litigation? 14 A. Yes. 15 Q. The risk of that litigation presumably is that 16 someone will turn round and say, "Horizon is 17 causing discrepancies" -- 18 A. Absolutely. 19 Q. -- which, in fact, it was in this case? 20 A. It was. 21 Q. What did you think to this email when you saw 22 it? 23 A. It's different because now I'm in a position 24 many years later where I'm reflecting on it, of 25 course with the benefit of hindsight. But it's 198 1 uncomfortable. It's uncomfortable reading today 2 and it was uncomfortable reading then. 3 Q. Was this reflective of a broader culture in the 4 Post Office at the time, that Horizon shouldn't 5 have been challenged? 6 A. I don't want to repeat going back to the 7 messages but, absolutely, on the basis that 8 there was no accepted agreement, that there was 9 Horizon integrity issues, that was the general 10 culture and that is what drove the behaviour. 11 Q. Please can we turn to page 2 of this email. At 12 the bottom you email -- I think it looks like 13 you email Dave Hulbert and Andy McLean? 14 A. Yes. 15 Q. You say: 16 "Bad news. 17 "On further investigation today we have 18 established that the POC file actually matches 19 POL-FS." 20 What does that mean? 21 A. It meant that we still didn't and couldn't 22 identify the values or the branches involved in 23 this particular issue. 24 Q. It says: 25 "This means that we still don't know what 199 1 value stock has been returned by these 570 2 branches. Furthermore the only way we could 3 find out would be for someone at Swindon to find 4 the specific pouches and physically open each 5 one and count the stock. 6 "This rules out a proactive recovery plan." 7 Do you still agree with that? 8 A. Roll it out in the sense that what would we have 9 told the branch? We could have told them there 10 was a problem but nothing more. Couldn't give 11 them any indication of -- to the extent of 12 impact that it had had on them. But, actually, 13 with hindsight, there would have been 14 an opportunity to, I think, have -- well, I can 15 see that throughout the email chain that Product 16 and Branch Accounting have been included within 17 this because they would have been seeing as 18 well, the outcome and the impact of this. 19 There was clearly an opportunity to have 20 reached out to the 570 branches, even in the 21 absence of detail and values, and inform them so 22 that they were aware so that we could have done 23 something more collaborative with them to 24 resolve the issue. 25 Q. You go on to say that: 200 1 "I have ensured that both NBSC and HSD have 2 scripts to deal with any queries relating to 3 stock from these branches." 4 So that's if someone rings in. 5 A. That's that reactive process again and that 6 script again, to make sure that they catch it 7 and the branch is recompensed and the correction 8 made. 9 Q. The Inquiry hasn't seen or had sight of that 10 script. Do you know where those scripts would 11 usually be held or stored? Was it on the Remedy 12 system as well? 13 A. Given the temporary nature, I'm not sure whether 14 they were uploaded digitally or not, actually. 15 But if they were anywhere, it would be Network 16 Business Support Centre environment. 17 Q. If we can move on to another issue, please, and 18 that's a PEAK that you refer to in your witness 19 statement. It's POL0001313. Before the Inquiry 20 sent you a document like this, had you seen one 21 of these before, a PEAK? 22 A. Yes, occasionally. I can't remember what the 23 PEAK acronym sort of means but I think it 24 relates to Fujitsu's internal tool or system for 25 recording bugs and defects that needed to be 201 1 queued in a backlog then for fix. If there was 2 any reason for somebody like Anne Chambers -- 3 I think it is Anne in this instance, isn't it -- 4 to make me aware of that, I suspect it was 5 probably attached to the email explaining to me, 6 hopefully in simple terms, why I was being 7 informed. 8 Q. This says, issue at Branch 106129, which has 9 a non-zero trading position. The entry on 10 18 February at 1500 hours, just go down 11 slightly: 12 "Branch 10629 appears to have been affected 13 by a known software problem ... which causes 14 a non-zero trading position, a receipts and 15 payments mismatch and an incorrect discrepancy." 16 Further down on 25 February, we see that you 17 were notified of this problem. Do you have any 18 independent recollection of that? 19 A. I'm afraid not, no. 20 Q. Over the page, again on 25 February, it says: 21 "Thanks, Anne. Final BIMS issued to POL 22 including 'Information for BIMS' text. 23 Returning call to EDSC for closure." 24 Can you assist with what that means in terms 25 of the "final BIMS issue to POL"? 202 1 A. Not an expert, I'm afraid, on the issues 2 management process but I think that this would 3 have been communicated to Product and Branch 4 Accounting. It's not something that they'd have 5 been looking for support from incident 6 management with and I'm afraid I have no 7 recollection at all what the acronym "EDSC" 8 means. 9 Q. We'll leave that there then and move on to 10 another matter. POL00023765, please. This is 11 a PEAK PC0152014, branch 183227. It refers to 12 an incomplete summaries report. Can we go to 13 the bottom of page 1, please. We see an entry 14 on 10 December, right at the bottom. 15 A. Yes. 16 Q. If we could go over the page, it says: 17 "This is due to a single SC line written for 18 $1,000 (£484) with no settlement in the middle 19 of two RISP transactions." 20 It says: 21 "On call PC0151718 the harvester exception 22 was corrected and now the transaction for the 23 day don't zero, hence this issue with the 24 incomplete summaries [form]." 25 It goes on to requesting the message form. 203 1 Do you understand broadly what this problem was? 2 A. I don't. It's too technical for me but, 3 fundamentally, it's describing another software 4 related, application related issue that's 5 created a discrepancy in the branch. 6 Q. It refers to OCP17510, which has been raised. 7 Do you remember what an OCP was? 8 A. Operational Change Process or Procedure or 9 something like that, I think. It was the audit 10 trail that was created to match with whatever 11 action was being taken, so that that 17510 12 should match whatever correction activities, so 13 whether that's message store or otherwise, was 14 done to make this right. 15 Q. So when you say "message store", do you mean 16 where Fujitsu were making changes -- 17 A. The way Fujitsu went into the system to correct 18 the issue. 19 Q. The OCP was the -- 20 A. And the OCP was the audit trail for that 21 activity, yeah. 22 Q. Do you recall, from your recollection, what were 23 the controls, the security controls, on the use 24 of OCP? 25 A. I don't. I mean, I would have seen plenty in my 204 1 time for both awareness and, on occasions, 2 approval but I don't really remember, I'm 3 afraid, the process, quite how we raised them, 4 how they were triggered. I think we just had 5 an approval or sight of within my team. 6 Q. Can we look at the OCP. It's FUJ00087194. So 7 we see "Write corrective bureau message" for the 8 branch. It says: 9 "A single SC message ... was written in 10 error on 26th November ... selling 1,000 US 11 dollars, with no corresponding settlement line. 12 To remove the effects of this message at both 13 the branch and on POL-FS, we will insert a new 14 message to negate the effects of the original 15 message." 16 So is your understanding of this that what 17 Fujitsu would do is insert, essentially, 18 a transaction into the branch accounts? 19 A. To balance the books, yes, and an accepted 20 process as well. I know this is just one 21 example but that would be what I'd expect to 22 happen in these circumstances. 23 Q. When were you first aware of Fujitsu's ability 24 to -- 25 A. Oh. 205 1 Q. -- insert transactions like this? 2 A. I don't know but I'm -- I'm going to say 3 probably HNG-X rather than Legacy Horizon, so 4 much later on. 5 Q. Well, we're here in 2007. 6 A. Oh, right, okay. 7 Q. Horizon Online is 2010 onwards. 8 A. Yes, so -- but I still -- I'm afraid I don't 9 know. It's not something that you were 10 particularly conscious of at the time or that it 11 maybe particularly concerned you or felt wrong 12 in any kind of way, if I can -- I can't find 13 a way to articulate it. It felt part of a BAU 14 process from an internal operational 15 perspective. 16 Q. Were you aware of any ability of Fujitsu to 17 insert, edit or delete transactions from the 18 branch accounts outside of the OCP procedure? 19 A. Oh, no. I'd have always expected there to have 20 been communication with Post Office and, if it 21 was making a correction, OCP was -- I was going 22 to say the only -- there may have been another 23 but that was a process for audit trailing 24 amendments that were being made. 25 Q. If we can go down to the bottom of this OCP, 206 1 sorry, not that far. Sorry, just a little 2 further up. Thank you. We refer to extra 3 detail and it gives the original message and the 4 new message attributes. Do those -- I mean, 5 from your position, you say you're 6 non-technical, do those words mean anything to 7 you, insofar as could you challenge them or 8 check they're accurate? 9 A. No. You know, it's -- I would always work, my 10 team would always work, based upon the business 11 outcome element of the -- any given situation. 12 So for me, really simply, this is about 13 rectifying correctly for a branch a situation 14 which is not of their making. 15 Q. So it says at the bottom there: 16 "Gary Blackburn (POL) is already aware of 17 this issue." 18 A. Yes. 19 Q. If we go down, there's an email, which you 20 aren't in copy for, but it says: 21 "Hi Gaby, 22 "POL approve this change. 23 "As soon as I saw the branch name, 24 I realised that this is the one that Gary spoke 25 to Anne Chambers about earlier." 207 1 A. Yes. 2 Q. So, at this time, you were relying on Fujitsu to 3 implement the change and no one at Post Office 4 was checking it from a technical perspective? 5 A. Not from my perspective and, as I've said, I and 6 my team, of which Julie Edgley was one at the 7 time, would not have had that ability, so it's 8 going on the advice, the guidance given. 9 Q. Would you have told or arranged for the 10 subpostmaster to be told of this change? 11 A. I don't remember doing so. I really would like 12 to sit here today and be able to say to you that 13 we did, it is possible that we did. We spoke 14 to -- and I, through all my years, spoke to many 15 subpostmasters but I really cannot say that 16 I did on this occasion. 17 Q. In general terms, when you had these types of 18 transactions, would you -- 19 A. In general terms, yes. Naturally, we would want 20 to ensure that the postmaster was aware because 21 there was an issue of awareness and timing of, 22 because it would become apparent to them because 23 an entry was being made. I think there might be 24 a different example within the pack where we 25 have actually recorded an email that we've spoke 208 1 to a particular postmaster but I can't remember 2 this particular example. 3 Q. If we could go back to the PEAK we looked at 4 before, it's POL00023765, and to page 3. 5 14 December, the entry at 15.37 says: 6 "Email to Gary Blackburn." 7 The Inquiry doesn't have or hasn't had sight 8 of that email but the following entry says this: 9 "The counter problem which caused the first 10 issue has been corrected by inserting a message 11 into the message store, for equal but opposite 12 values/quantities, as agreed with POL (OCP 13 17510)." 14 Which was the document we just went to. 15 A. Yes. 16 Q. "As a result of this corrective action, the net 17 effect on POL-FS is zero, and the POL-FS figures 18 are in line with the branch. POL-MIS received 19 both the original message and the corrective 20 message. 21 "Once the problem was corrected, there 22 should have been no impact on the branch. 23 However it has been noted that the stock unit 24 BDC had a loss of $1,000, which was generated 25 after the correction was made. We have already 209 1 notified Gary Blackburn at POL (email attached). 2 This appears to be a genuine loss at the branch, 3 not a consequence of the problem or correction." 4 Do you recall being made aware of this? 5 A. I don't. But clearly I was. 6 Q. Is it a fair summary to say what Anne Chambers 7 is effectively saying here is there was 8 initially a problem which caused a $1,000 9 discrepancy? 10 A. Yeah. 11 Q. They inserted a transaction to try to correct 12 that, so a $1,000 transaction the other way, and 13 now there was a $1,000 loss to the branch, which 14 they were saying was unrelated to the use of the 15 remote access. Does that strike you as odd? 16 A. Yes. 17 Q. There's no record in this PEAK of this being 18 challenged at the time. Do you think you would 19 have challenged it? 20 A. I'd like to think so but, again, you know, 21 passage of time, I honestly -- it isn't 22 something that I remember, but, you know, 23 clearly $1,000 is a trigger there to say too 24 much of a coincidence. 25 Q. It's correct, isn't it, that under the way Post 210 1 Office interpreted its contract with 2 subpostmasters at the time, that a discrepancy 3 for which there wasn't a system explanation, the 4 subpostmaster would have been expected to make 5 good that discrepancy? 6 A. That would be my understanding, yes. 7 Q. So as a matter of basic fairness, do you accept 8 they should have been told about, the fact that 9 a transaction -- I won't use that term, sorry -- 10 the fact that remote access had been used to 11 insert a transaction? 12 A. Yes. 13 Q. But, again, there's no record, in this PEAK, of 14 the subpostmaster being told? 15 A. And I can't remember, I'm afraid, whether it was 16 or it wasn't. 17 Q. Before moving on, can we turn to page 12, 18 please, of your witness statement, paragraph 23. 19 You deal here with remote access and you refer 20 to the OCP procedure. You say: 21 "I also knew that they ['they' being 22 Fujitsu] could see what had happened in branch 23 down to the detailed sequence of individual 24 keystrokes." 25 So is it your evidence that Fujitsu held 211 1 audit data that showed what keystrokes had been 2 held in branch? 3 A. Yes, that was my understanding yes. 4 Q. How often did you request such data or see such 5 data? 6 A. Possibly once or twice with post offices that 7 were in this kind of scenario. I'm not saying 8 the same particular trigger that were disputing 9 discrepancies. I do remember it was actually 10 through the BIM process, through Nick Crow, that 11 it had been requested to try to prove or 12 disprove a sequence of events that might have 13 explained whether it was system driven or user 14 driven. 15 Q. So you say you use it two to three -- 16 A. I think a couple of occasions, only a couple of 17 occasions where there was really quite difficult 18 protracted enquiries and investigations. I'm 19 sorry, I can't remember the names of the post 20 offices. I remember one was in Exeter is all 21 can say. 22 Q. Was it difficult to obtain that information? 23 A. Yes. 24 Q. Why? 25 A. It wasn't -- it's something I had to -- you 212 1 know, you'd have to ask for, and it may have 2 been on occasions that, you know, it wasn't 3 available for whatever reason, it's a question 4 for Fujitsu to understand how they audited and 5 retained and held that information. But, no, it 6 wasn't something that I could just freely 7 obtain. 8 Q. Did you have to get internal approval to seek 9 such information? 10 A. I don't remember doing so, I remember it just 11 feeling that if it was the right one -- one of 12 the right questions to ask in a circumstance 13 that we would ask. 14 Q. Were you aware of any costs implications of 15 obtaining such data? 16 A. I wasn't no. 17 Q. Are you aware of any whistleblowing policies in 18 Post Office at the time, relating to issues such 19 as remote access or bugs, errors and defects? 20 A. No, I'm sorry. 21 MR STEVENS: Sir, I'm just looking at the time. 22 I've still got a few topics to cover and there 23 are questions from Core Participants as well. 24 I understand Mr Blackburn is available tomorrow 25 to give evidence. 213 1 A. I have come prepared, yes. 2 MR STEVENS: I'm happy to carry on but -- 3 SIR WYN WILLIAMS: Are you saying collectively that 4 we would go substantially beyond 4.30 if we 5 tried to finish him today? 6 MR STEVENS: I'll just check. How long do people 7 think they'll be with questions? 8 MR JACOBS: Sir, as matters stand, I don't have 9 anything at the moment. 10 MS PAGE: I've only got short questions. 11 MR STEVENS: In which case, I'd understood there 12 would be many more questions and so I think 13 I can fit it in, in that time. 14 SIR WYN WILLIAMS: Yes, I am reasonably happy to go 15 until about 4.45. Thereafter, I have to say my 16 concentration levels will start to waiver but 17 I'm sure that Mr Blackburn would prefer to 18 finish if he could tonight. 19 A. I would. Thank you. 20 MR STEVENS: I have been asked if we can take 21 a five-minute break for the purposes of the 22 transcriber. 23 SIR WYN WILLIAMS: I tell you what, let's take five 24 minutes and then everybody can sharpen 25 themselves to try to finish by 4.45. 214 1 MR STEVENS: Thank you, sir. 2 (4.11 pm) 3 (A short break) 4 (4.15 pm) 5 MR STEVENS: Sir, can you see and hear me? 6 SIR WYN WILLIAMS: I can. 7 MR STEVENS: Good thank you. I'll carry on. 8 Mr Blackburn, before moving on to the next 9 topical I failed to put something to you 10 earlier, which I should do now, so apologies for 11 treading over ground we've covered. Can we go 12 back to FUJ00121071. This was about the remming 13 out bug we discussed earlier. At page 3, we 14 discussed about the ruling out of proactive 15 recovery plan and also a -- so relying on 16 a reactive one. I feel I should put it to you 17 squarely, from this, is it fair to say that 18 a reactive recovery plan was, in fact, followed 19 in this case? 20 A. Yes. 21 Q. Thank you. If we could move, then, to cover 22 a dispute arising from the Hogsthorpe Post 23 Office, and that is, if we could turn to 24 POL00021163, please, and page 4. Thank you. So 25 at the bottom, please, we see that this refers 215 1 to the Hogsthorpe Post Office, who at the time 2 was operated by David Hedges who liked to be 3 known as Tom Hedges. Mr Hedges is a Core 4 Participant in this Inquiry who has provided 5 a witness statement, and this is an email from 6 Karen Arnold; do you recall who she was? 7 A. I'm going to say that generic term again, Retail 8 Network Management arena. Yeah. 9 Q. It says: 10 "Further to our conversation last week 11 regarding the losses at Hogsthorpe, the SPMR 12 David Hedges (who liked to be known as Tom) has 13 contacted the NBSC to establish what the 14 BAU ..." 15 Is that business as usual? 16 A. Yes. 17 Q. "... BAU/correct process is for suspending 18 a session of Smartpost." 19 Can you assist us with what Smartpost is? 20 A. Only it was a transaction at the time. I can't 21 remember what, specifically, but it was 22 a posting transaction. 23 Q. It says: 24 "Tom tells me that the NBSC said it was okay 25 to use either of the methods he describes. As 216 1 a reminder, I have copied information below in 2 respect of what described to me last week." 3 That is over the page. We don't need to 4 look at that in any detail. 5 If we could have on the screen at the same 6 time, please, paragraph 21 of your witness 7 statement on page 11, and on the actual 8 document, POL00021163. If we could go to 9 page 5, please. If you go below that for the 10 time being. Thank you. 11 Your response on 7 July was to say that: 12 "Fujitsu would not check a replace processor 13 automatically, but I don't believe that would 14 add any value in this instance. As we discussed 15 last week the most likely explanation was/is 16 user error, but given the always into NBSC and 17 HSD, we should assume that this is not the root 18 cause at this time." 19 In your witness statement at paragraph 21, 20 you say: 21 "When Karen Arnold first contacted me, my 22 initial response comes from my preempting the 23 most likely root cause of the problem being 24 within branch, which was purely based upon the 25 fact that at the time I had no understanding, or 217 1 even belief, that the Horizon application could 2 or did generate erroneous discrepancies linked 3 to this transaction type." 4 Now when you say that, are you saying 5 transaction type is Smartpost? 6 A. Yes. 7 Q. So you were aware of the ability of Horizon to 8 create discrepancies -- 9 A. The general was, we've discussed, yes. I was 10 particularly -- hence I chose my words quite 11 carefully there. It was Smartpost, 12 specifically. 13 Q. Thank you. We can take down the witness 14 statements, then, for the time being. Why did 15 you think user error was the most likely 16 explanation, even though you had seen evidence 17 of bugs, errors and defects in Horizon in other 18 areas? 19 A. It's I think an example, looking back, of just 20 how we all thought, behaved, believed, worked 21 within the processes. And that's despite -- and 22 I know that sounds incredibly silly to say that 23 at the end of this review, but that is how we 24 believed, if there had to be evidence to support 25 Horizon creating discrepancies. So in the 218 1 absence of the general stance, rightly or 2 wrongly, was that it was more likely to be user 3 error interacting with the service in completing 4 the transaction incorrectly. 5 Q. Please can we bring that document back up. So 6 it's POL00021163, and I think it was page 4. 7 Yes. The bottom paragraph of your email. You 8 say: 9 "If Tom has specific information such as 10 transaction time and values, please send this 11 across and I will get Fujitsu to investigate 12 immediately. If has no evidence then I'm afraid 13 there is nothing for Fujitsu to investigate." 14 What evidence could Mr Hedges have provided 15 which Fujitsu or the Post Office didn't have 16 access to? 17 A. Yes, I think ultimately it would have been the 18 accounting, the transaction logs, the trial 19 and/or final balances. But it's right that that 20 would have all been available to Fujitsu, 21 correct. 22 Q. What was the reason for asking Mr Hedges for 23 this? 24 A. I can't recollect, but I think it comes from my 25 previous statement about in the absence of -- 219 1 you know, it wasn't automatically taken forward. 2 Q. Please could we look at page 3 now. So Karen 3 Arnold replies, saying: 4 "I'm not sure why Fujitsu would be changing 5 the processor if they didn't think there was 6 a problem. 7 "Having spoken to Tom today, once the new 8 processor is install he is going to do a BP 9 rollover and then keep a tally manually of every 10 Smartpost item to check against Horizon. This 11 however won't help anything that has gone [on] 12 previously." 13 Pausing there, if Mr Hedges receives a new 14 processor unit, approaches his transactions in 15 the same way, and he stops getting the 16 discrepancies, that suggests that the processor 17 is at fault. 18 A. It certainly could have been an explanation, 19 yes. 20 Q. If he still had discrepancies, that would 21 suggest that the processor itself wasn't at 22 fault, it may have been something else -- 23 A. Yes. 24 Q. -- it may have been the Horizon System, but not 25 the processor. 220 1 So changing the processor would give you 2 some evidence relevant to determining the cause 3 of the problem; do you accept that? 4 A. Yes, it could have done, yes. 5 Q. Can we turn to page 2, please. You reply on 6 2 July. You say: 7 "Karen, 8 "Fujitsu have always had a preventative 9 maintenance policy and therefore sometimes will 10 swap out kit without actually finding a fault. 11 Also it generally helps with customer perception 12 of the service they have received. I accept in 13 this instance that this policy could work 14 against us, but are you suggesting that if after 15 swapping the processor and all discrepancies 16 cease, that Tom will claim this is clear proof 17 of Horizon creating discrepancies? I strongly 18 suggest that Tom obtains the necessary evidence 19 now, if it is available." 20 Pausing there, can we at the same time, 21 please, show your witness statement page 11, 22 paragraph 21. At the bottom, it's three 23 quarters of the way down, you say: 24 "I went on to offer advice which was to 25 obtain evidence prior to requesting that Fujitsu 221 1 attend site. Fujitsu had, at the time, a basic 2 maintenance policy which was to replace hardware 3 (if they couldn't find an obvious onsite fix or 4 explanation). This policy was also likely 5 driven by the need to fix and close calls as 6 quickly as possible, driven by service level 7 targets. In this instance that would have led 8 to a loss of information contained in that 9 processor that potentially supported the 10 postmaster at Hogsworth [sic]. Hence my 11 reference to 'this policy could work against us' 12 that this cause of action could lead to an end 13 of investigative options and establishing root 14 cause." 15 Insofar as Fujitsu were planning to change 16 the original processor, there was no indication 17 they were going to simply destroy it, was there? 18 A. No, it went into -- so processes were recycled. 19 The Post Office only had so many, you know, 20 sufficient to support the branch network with 21 the spares to keep the whole cycle going. So 22 without any intervention, that processor would 23 have just simply been swapped, gone back into 24 the recycling process and wiped, in effect, in 25 terms of any information contained on it. 222 1 Q. But the processor could have been replaced 2 without losing the information and continued the 3 investigation? 4 A. It could have been, I believe, yes. 5 Q. When you're saying "this policy could work 6 against us" here, are you referring to the loss 7 of investigative options or are you referring to 8 the fact that it may have shown a problem with 9 the Horizon System? 10 A. No, it's the first thing. I was -- again, 11 rightly or wrongly at the time, I was really 12 wanting the postmaster to try and gather the 13 evidence so that we could take forward 14 an investigation. The option of simply having 15 the processor replaced, for me, was one that 16 I would rather have avoided at the time. 17 I don't know what followed this particular stage 18 of the investigation, because I don't believe it 19 was the end of it, but this was only a step in 20 my belief, my mindset at the time. Let's have 21 one more go at trying to obtain the evidence. 22 Is there anything there that allows me to take 23 that forward in the way we discussed earlier 24 with Fujitsu, rather than have it removed and 25 potentially lost? 223 1 Despite what you just said a few minutes 2 ago, Sam, if we'd have replaced the processor 3 and the new processor didn't display the same 4 symptoms and errors, I believe that there's 5 a strong possibility that would equally have led 6 to an end to the investigation and there would 7 have been an assumption made that the 8 discrepancies were created in branch, rather 9 than the technology. 10 Q. Was this not a type of case where keystroke data 11 may have been of assistance? 12 A. Yes, I think it probably would have done, on 13 reflection, but I don't remember, and clearly 14 haven't mentioned it at this point in time, and 15 I don't remember it being raising subsequently 16 either. 17 Q. Thank you. If we could take down the witness 18 statement, please, but keep up the document. If 19 we go up to the next page where it's John 20 Breeden's email. If we could go to the next 21 page up, please. Thank you. 22 This is an email from John Breeden on 23 3 July. Do you recall who John Breeden was? 24 A. Again, Retail Network. Possibly Karen Arnold's 25 line management chain. More senior. 224 1 Q. It says: 2 "I have read the recent emails on the above 3 and considered the information. I am concerned 4 if we swap the processor now and the errors stop 5 this could lead to (i) a claim that Horizon has 6 problems in its accuracy and fuel some of the 7 recent press articles and (ii) the SPMR will 8 claim that all previous errors are down to 9 Horizon and we have no way to disprove this if 10 everything is resolved when the new processor is 11 installed." 12 Over the page, it goes on to say: 13 "I think we need to think this one through 14 carefully and the SPMR should be providing 15 evidence to support his claims which can be 16 investigated before we change pieces of 17 equipment." 18 So at this stage, was there a general 19 concern in Post Office affecting investigations 20 about the risk of litigation relating to 21 Horizon? 22 A. Well, the retail network managers would be 23 better placed to answer that, but as maybe 24 I mentioned earlier, I believe, you know, there 25 was an upward trend in the volume of this type 225 1 of enquiry. So that would make perfect sense if 2 there was. 3 Q. If we look at that email, I think if we go up, 4 in August 2009 -- sorry, July 2009. Were you 5 aware of a Computer Weekly article in May 2009 6 regarding the robustness of Horizon? 7 A. Actually, now you mention it, I think that was 8 one of the things that we were all informed 9 about, yes. So I think it was brought to our 10 attention as staff members. 11 Q. What was said to you, in respect of that 12 article? 13 A. I think it was -- and again, I'm sorry, because 14 this is just, as you can tell, how I believed it 15 was at the time and how we were, and how we 16 operated, but I believed, again, we were 17 reassured that Horizon was fit for purpose. 18 Q. Shaun Turner today gave evidence that 19 communications on that issue came down from what 20 he thought were board level, regarding the 21 robustness of Horizon in response. Would you 22 agree with that? 23 A. Yes. I mean, the part of the business I worked 24 in is where the message would have come from, 25 ultimately, I would assume. It's come from 226 1 higher, but the IT directorate and function 2 would have been the ones communicating to the 3 likes of myself. 4 Q. Do you think the points mentioned by Mr Breeden 5 in points (i) and (ii) here are in any way 6 appropriate considerations to take into account 7 when deciding how to investigate potential 8 discrepancies? 9 A. Well, you know, again, it's passage of time, 10 isn't it? But it's interesting that John seems 11 to be coming to the same conclusion as me in 12 terms of the process we're following, but for 13 a slightly different reason. My entire job was 14 about trying to resolve issues and fix issues on 15 behalf of the Post Office and the postmasters, 16 not to be covering up anything or being 17 negative. So for me personally, no, it was not 18 a criteria that was necessary in determining 19 what actions we were going to take. 20 Q. Can we quickly bring up, please, POL00012547 at 21 page 3. Under the Wednesday entry -- sorry, 22 this is a note from Karen Arnold. It's on this 23 issue. On the Wednesday entry, it says: 24 "Checked with Paul Kellett and no losses 25 settled centrally on 05.08.09. Contacted Tom 227 1 who confirmed he had been short by approximately 2 £40 on his last BTS, and had made this good. 3 Advised that Fujitsu confirmed that they have 4 not found any system errors which would have 5 caused the discrepancies and concluded that 6 there was nothing wrong with the processor. 7 Advised Tom that he needs to provide evidence to 8 support any claims that the problems with losses 9 were as a result of Horizon and that he is 10 responsible for making the losses good." 11 So these discrepancies, you accept, are 12 something for which Mr Hedges could have -- 13 well, would have been asked to make them good 14 and settled them himself? 15 A. It certainly looks that way. That's the way 16 I interpreted that, yes. 17 Q. If he didn't, he could face suspension? 18 A. Err, I'd like to think not, but that was outside 19 of my remit. It was retail network management 20 process. 21 Q. In this case Mr Hedges, in due course, was 22 convicted of a criminal offence arising from 23 issues relating to discrepancies for which his 24 conviction has been overturned. Were you 25 involved in that element of his case? 228 1 A. That's news to me. I didn't know that. 2 Q. In this case, the fact that Fujitsu said they'd 3 found no problems with the system, we discussed 4 in your evidence earlier that there was no -- or 5 you didn't have sufficient support internally 6 with IT to check what they were saying, and you 7 relied on their expertise? 8 A. (The witness nodded) 9 Q. Was this a case where you would have gone to 10 someone within IT to check on it? 11 A. I would have liked to have had the opportunity. 12 I really think that that wasn't the case. 13 I don't think that this investigation went -- 14 I mean, naturally emails don't record 15 everything, there were lots of conversation, but 16 I think that this pretty much summarises the 17 investigation. 18 Q. I'm going to simply ask you now to turn to 19 page 15 of your statement, please. In it, you 20 say that: 21 "[You] have chosen to remain with the IT 22 service industry, and if I've learnt one thing 23 during that time, it is that IT can and does 24 fail and that people are fallible and make 25 mistakes. It is how you respond to those 229 1 circumstances that matter. Openness and honesty 2 is, and always will be, the best policy. 3 Clearly there were occasions when the benefit of 4 doubt could and should have been given. I'm 5 sickened by what I've read post-my time at Post 6 Office and have experienced a wave of emotions 7 from sadness, shame and anger." 8 Do you think the Post Office should have 9 been more open and honest? 10 A. With the benefit of hindsight, in some respects, 11 but yes, clearly. 12 Q. If so, how? In what ways? 13 A. I think I can only try and convey what I thought 14 my role was, and it wasn't to cover up any 15 issues with technology. It was to ensure that 16 we were providing a quality service to the 17 British public through the branch network and 18 that includes the independent network. 19 Everything I ever tried to do, and my team tried 20 to do, was for that end and that benefit. 21 To watch the programme, as I did two years 22 ago on TV, and see people rather than FAD codes 23 was ... was difficult. Apologies. 24 Q. There is no need to apologise. If you wish to 25 take a moment, please do so. 230 1 A. It's hard for me to reconcile, years down the 2 line, in my mind, given all the evidence that 3 we've gone through today with me, let alone what 4 I'm sure you've gone through throughout the 5 Inquiry, how the HR and the human element was so 6 misaligned. It almost feels to me that there 7 was some disconnect between the reality of what 8 we were all trying to manage in the right way, 9 for the right reasons, and the impact and the 10 outcome. 11 I just cannot understand, probably because 12 I wasn't part of that part of the business, 13 I didn't see what retail network managers had to 14 go through, the policies that they had. I'm not 15 even familiar with a subpostmaster's contract, 16 for example. But still find it very hard for 17 you to tell me that the gentleman at Hogsworth 18 (sic), you know, suffered in that way, off the 19 back of something where there was clearly 20 an element of doubt. 21 And I think that's the thing for me. There 22 was doubt. Despite the messages, despite 23 Horizon has no integrity issues, messages 24 constantly coming down. Clearly there's 25 evidence to say that there were some problems, 231 1 that there were risks involved. And they may 2 have been small in the greater context, as in 3 the volume of -- I appreciate the impact has 4 been huge on individuals, but I find that -- 5 I simply can't reconcile it, in my own mind now, 6 today, I'm afraid. Hence the emotional words 7 I've used there, and that's where the anger 8 comes from as well. I worked very hard. 9 MR STEVENS: Mr Blackburn, I don't have any further 10 questions but is there anything further you 11 would like to say before I ask if the Core 12 Participants ... 13 A. No, I just hope that it's of use. 14 SIR WYN WILLIAMS: Are there any other questions? 15 MS PAGE: Sir, I just would like to follow on with 16 the rest of what is actually in the same 17 paragraph that Mr Stevens has just referred to, 18 if I may. 19 SIR WYN WILLIAMS: All right. 20 MS PAGE: It's very brief. 21 SIR WYN WILLIAMS: Yes. 22 Questioned by MS PAGE 23 MS PAGE: At the end of that same paragraph, you 24 said that: 25 "If [you'd] had more visibility of the 232 1 action that was taken against subpostmasters, 2 you hope that you would have found the courage 3 to challenge, sound a note of caution, and 4 promote the communication of outcomes more 5 vigorously and robustly than I perhaps did." 6 A. Yes. 7 Q. My question is simply: what do you think the 8 reaction of those above you would have been if 9 you had? 10 A. Well, I think the fact that I didn't probably 11 gives you your answer. I don't think -- 12 a highly politicised organisation, very 13 hierarchical, I'd have been seen as stepping out 14 of line with the message. I can't imagine that 15 that would have been good for my career, so I'm 16 sure at that point in time -- and this is 17 obviously a hindsight reflection -- I obviously, 18 on occasions, chose to unconsciously protect 19 myself. 20 MS PAGE: Thank you. That's all I wanted to ask. 21 SIR WYN WILLIAMS: Thank you, Ms Page. 22 Thank you, Mr Blackburn, for firstly giving 23 a witness statement and, secondly, answering 24 a good deal of questions this afternoon and I'm 25 grateful to you for the frankness with which you 233 1 have expressed some of your emotions. 2 So I'm glad we were able to complete today. 3 I'm sure that's of great relief to you. I'll 4 see everybody else at 10.00 tomorrow morning. 5 MR STEVENS: Thank you, sir, and we're hearing from 6 Anne Allaker and Gayle Peacock. 7 SIR WYN WILLIAMS: Yes, thank you. 8 MR STEVENS: Thank you, sir. 9 (4.45 pm) 10 (The hearing adjourned until 10.00 am 11 the following day) 12 13 14 15 16 17 18 19 20 21 22 23 24 25 234 I N D E X SHAUN TURNER (affirmed) ..............................2 Questioned by MR BLAKE ........................2 Questioned by MR STEIN ......................104 Questioned by MS PAGE .......................116 GARY DAVID BLACKBURN (affirmed) ....................120 Questioned by MR STEVENS ....................120 Questioned by MS PAGE .......................232 235