Independent report

Additional information to Kroll report

Updated 18 October 2024

Restricted use warning

This report is an addendum to a report prepared by Kroll Associates UK Limited (Kroll) who were appointed by the Department for Business and Trade (DBT) to undertake an independent investigation into Capture (hereafter the ‘Public Report’).

Kroll has consented to a request from the Client that this Public Report may be published on website(s) indicated by the Client, where it can be accessed in the public domain and has been prepared for this purpose only. This Public Report may not be re-published or used in any other manner without Kroll’s prior written consent. Other than as set out above, any communication, publication, disclosure, dissemination or reproduction of this Public Report or any portion of its contents to third parties without the prior written consent of Kroll is not authorised.

Kroll assumes no direct, indirect or consequential liability to any third party or any other person who is not the intended addressee of this Public Report for the information contained herein, its interpretation or applications, or for omissions, or for reliance by any such third party or other person thereon. To the extent information provided in this Public Report is based on a review of publicly available records, such information, as presented, relies upon the accuracy and completeness of those records, which have not been corroborated by Kroll.

Statements herein concerning financial, regulatory or legal matters should be understood to be general observations based solely on Kroll’s experience as risk consultants and may not be relied upon as financial, regulatory or legal advice, which Kroll is not authorised to provide. All such matters should be reviewed with appropriately qualified advisors in these areas.

This Public Report does not constitute a recommendation, endorsement, opinion or approval of any kind with respect to any transaction, decision or evaluation, and should not be relied upon or disclosed as such under any circumstances.

1. Introduction and methodology

1.1 Introduction

This document is an addendum (‘the Addendum’) to the report entitled Independent Investigation into Capture Accounting Software (hereafter referred to as ‘the Report’) which was prepared by Kroll Associates U.K. Limited (‘Kroll’, ‘We’, ‘Us’ or ‘Our’) for DBT and published on 30 September 2024. The Addendum should be read alongside the Report, and, unless stated otherwise, the limitations and caveats set out in the Report apply to the Addendum.

This document contains conclusions following Kroll’s review of additional datasets (see section 1.2) that, given the scope and timeline of review which led up to the Report, were considered relevant but not a priority source for the Report. Although these additional datasets were not collected specifically for the purposes of our review, there was a possibility that they contained relevant information and were reviewed for completeness.

On 9 September 2024, Kroll signed a contract variation with DBT, to enable Kroll to conduct a targeted review of these additional datasets, and this document sets out our conclusions following that review in relation to 3 thematic areas (‘the Themes’) which stemmed from the Report. The Themes are introduced below, and our methodology and approach for the review is set out in section 1.2.

1.1.1 Theme 1: internal communications responding to issues with the Capture software

The Report comments that limited information around internal communications and reporting/escalation of issues related to the Capture software was identified. Analysis of this thematic area seeks to further the understanding of how Post Office responded internally to the issues identified with the Capture software between 1991 and 1999.

Searches related to this theme were targeted towards the following:

• communications from Post Office Limited (either the Capture team or ‘management’) with Retail Network Managers and Area Managers (who held more operational positions) on Capture, including any feedback on the software
• ‘board level’ or senior management discussions involving Capture, such as board meeting minutes or papers

1.1.2 Theme 2: communications between the Capture development team and Post Office Limited control functions

The Report sets out that there was limited connection between the Capture development team and various Post Office Limited control functions (including Post Office Limited Finance Department (PFD), Legal, Audit and Investigations). The Report sets out further that it is unclear whether / how Capture system bugs/errors that could have resulted in calculated shortfalls were considered when audits or investigations were initiated.

Searches related to this theme were targeted towards the following:

• discussions between legal, audit, investigations and the Capture team surrounding bugs and/or errors and impact on respective workstreams
• any regular reporting to the control functions, specifically Audit and Post Office Limited Investigations Department (POID) after the release of circulars or new versions of Capture
• correspondence between POID and the Capture team on specific cases

1.1.3 Theme 3: Capture specific information about known sub-postmaster cases

As set out in the Report, Kroll interviewed a number of sub-postmasters and reviewed documentation relating to their cases that was made available to us. The information reviewed included available documentation provided by Post Office Limited (for example, from branch files), and, in some cases, information provided by the sub-postmasters themselves.

Searches related to this theme were targeted towards identifying any further documentation or correspondence related to communications on Capture software issues and their impact between Post Office Limited and the sub-postmasters and internally between Post Office Limited’s teams.

Review relating to themes 1 and 2 was prioritised.

1.2 Methodology

1.2.1 Electronic documents held on behalf of Post Office Limited

During the course of our investigation, Kroll was made aware of various additional datasets containing many millions of electronic documents held on behalf of Post Office Limited, which were collected for a variety of purposes unrelated to our review.

Peters & Peters, who are appointed as Post Office Limited’s external counsel, acted as an interlocutor for Kroll, specifically in explaining the available datasets, in providing guidance on methods for searching the various datasets, and in coordinating with KPMG, who host the data within these datasets on behalf of Post Office Limited.

With the exception of the casework spreadsheets (see below), the datasets are managed by KPMG. The datasets themselves are also not ‘static’, and process new documentation on an ongoing basis owing to their specific objectives.

Whilst the information contained within these datasets has largely been gathered for alternative purposes, we deemed it possible that they may contain relevant information to our Terms of Reference, and thus have conducted a targeted review of these documents (see section 1.2.4).

The datasets accessible include the following:

• Dataset 1^{[footnote 1]}
• Dataset 2^{[footnote 2]}
• documentation gathered for the purposes of Bates and others v Post Office Ltd
• Dataset 3 - we were advised that the vast majority of this dataset relates to the period post-2011, and as such searches were not performed against it
• back-up tapes and servers stored in Post Office Limited’s Chesterfield offices
• casework spreadsheets and summaries of documentation gathered by Peters & Peters focused on criminal prosecutions sought by Post Office Limited since the mid-1980s. Information from this source was provided by Peters & Peters on 27 August 2024

1.2.2 Requests for documentation made to Post Office Limited and Peters & Peters

Following guidance provided by Peters & Peters, Kroll requested that Post Office Limited, as the owner of these documents, instruct KPMG to extract a subset of potentially relevant documents for our review, with reference to 2 different sets of search terms:

• a list of generic Capture related search terms (‘Generic Capture Terms’)
• specific search terms related to a number of sub-postmasters known to Kroll who provided consent (‘Sub-postmaster Specific Terms’)

These search terms were applied to documents within the date range 1991 to 2006, a time period agreed based upon the operating period of Capture (1992 to 1999) as well as the availability of data generally across all the datasets, which was summarised in heatmaps provided to Kroll by Post Office Limited on 24 July 2024.

Following the initial submission of the Generic Capture Terms list on 17 July 2024, Kroll received the first Search Term Report ^{[footnote 3]}, from Post Office Limited on 1 August 2024. During the intervening period, Kroll was working alongside Post Office Limited and Peters & Peters, in respect to data protection concerns and Data Sharing Agreement provisions, which are outlined in the limitations section of the Report. These were resolved by Post Office Limited at the end of July, with the Search Term Report sent shortly thereafter.

In addition, Kroll sought the consent of sub-postmasters that had been interviewed, in order to run searches across the datasets (the Sub-postmaster Specific Terms). The objective was to identify further information specific to their own individual experiences of Capture, as well as what electronic data may have been held on these files.

Sixteen sub-postmasters provided consent for search terms such as their names and certain identifiable branch details, to be run against the documents. We shared the search terms for these subjects on 7 August 2024 and 13 August 2024. We were provided with 2 Search Term Reports by Post Office Limited on 19 August 2024 that included the number of documents matching each of the Generic Capture Terms (second report) and the Sub-postmaster Specific Terms (first report) across the datasets.

Kroll’s review of the electronic documents was deprioritised at the time we issued a draft of the Report to DBT on 21 August 2024, owing to the remaining time available for review by Kroll within the original contract period, which expired on 11 September 2024. Despite this, there remained the possibility that the datasets contained relevant information that could inform observations outlined in the Report and we continued to discuss this risk with DBT as per section 1.2.3 below.

1.2.3 Search Term Reports

Following the contract variation agreed on 9 September 2024, Kroll re-engaged with relevant parties to enable the review of documents that had been made accessible to Kroll and had been accounted for in the latest Search Term Reports of 19 August 2024. We had identified minor variations in the number of documents with hits in the Generic Capture Terms Search Term Reports dated 1 August and 19 August 2024, and therefore sought an explanation from Post Office Limited and KPMG.

Following explanations provided to us by KPMG, the second Search Term Report contained a revised total number of documents with hits for the Back-up tapes dataset. A third Search Term Report was provided by KPMG on 26 September 2024, showing a revised number of documents with hits owing to work conducted by KPMG to remove duplicate documents. The resultant hits from the 26 September 2024 Search Term Report are set out below, which indicated that there were 178,571 hits for Generic Capture Terms.^{[footnote 4]}

Table 1: Search Term Report, 26 September 2024 – Generic Capture Terms

Dataset	Number of documents with hits [footnote 5]	Number of documents with hits plus families [footnote 6]
Dataset 1	60,410	967,579
Dataset 2	14,779	277,064
Group Litigation (Bates and others v Post Office Limited)	12,882	279,170
Back-up tapes	90,500	231,801
Total	178,571	1,755,614

As referenced above, we also received Search Term Reports for the Sub-postmaster Specific Terms on 19 August and 26 September 2024. We identified no variance between these Search Term Reports.

Table 2: Search Term Report, 26 September 2024 – Sub-postmaster Specific Terms

Dataset	Number of documents with hits	Number of documents with hits plus families
Dataset 1	7,839	244,663
Dataset 2	5,498	91,857
Group Litigation (Bates and others v Post Office Limited)	1,442	55,156
Back-up tapes	7,923	28,325
Total	22,702	420,001

We were informed that the results of the searches set out above reflect the sum of documents with hits across each individual dataset. This means that there are potential duplicate documents counted across the various datasets. These datasets were further de-duplicated, in the process outlined below.

1.2.4 Documents made accessible to Kroll

On 21 August 2024, Kroll was provided with access to an e-discovery platform to enable our review of available documentation. The platform is managed by KPMG on behalf of Post Office Limited. We understand that documents contained in this e-discovery platform (set out in Table 1 and Table 2) were de-duplicated across the datasets resulting in 1,102,093 unique documents.

1.2.4.1 Indexes for e-discovery platforms

In general, indexes are created on e-discovery platforms to support, amongst other capabilities, efficiently running search terms against the documents hosted on these platforms. Indexes are typically created using a combination of data fields such as file name, extracted text, author and available date fields (e.g. date created, date sent).

We understand that the datasets accessible to KPMG (section 1.2.1) each have their own index constructed, and in some datasets the index was built prior to KPMG’s engagement with Post Office Limited. We understand each index in the datasets accessible to KPMG would have been constructed differently both in terms of the data fields used and the order in which the fields were added to the index.

It was not feasible to replicate these indexes on the 1,102,193 documents accessible to Kroll, given the multiple different indexing methodologies already implemented across the various datasets. Kroll therefore built an index on the documents accessible, which would likely differ from the indexes built on the datasets accessible to KPMG.

1.2.4.2 Application of the Generic Capture Terms and the Sub-postmaster Specific Terms against the documents available to Kroll

We ran the Generic Capture Terms and the Sub-postmaster Specific Terms provided to Post Office Limited (see section 1.2.2) against the documents made available to us in the e-discovery platform. This resulted in 155,196 documents with hits against these search terms, and a total of 1,097,545 documents including each document’s ‘family’ (e.g. associated files such as email attachments).

We observed that this total is lower than the total number of documents accessible in the e-discovery platform (1,102,193), a difference of 4,355 documents (the ‘Unmatched Documents’) which is attributed to Kroll and KPMG relying on different indexes to facilitate their searches of documents.

1.2.4.3 Application of search terms to identify documents relevant to Themes 1, 2 and 3

For the purpose of identifying documents relevant to Themes 1, 2 and 3 (section 1.1), we applied further search terms specific to those themes.

These search terms were based on the Generic Capture Terms and the Sub-postmaster Specific Terms (section 1.2.2), but were further refined as follows:

• restriction to documents dating from the period 1991 to 1999, the period largely relevant to Themes 1, 2 and 3
• refining the search term lists to be targeted towards identifying documents relevant to Themes 1, 2 and 3 set out in section 1.1. For example, this included the removal of terms searching for specific suppliers, as our Report concluded that these entities were not connected to the development of the Capture software, and the addition of certain relevant search terms, for example, a combination of Capture and terms representative of board / governance activities (minutes, board, agenda, etc.)
• as the 4,355 Unmatched Documents were not responsive to the Generic Capture Terms and the Sub-postmaster Specific Terms we ran, including the date filter (documents from the period 1991 to 1999), we also applied the term ‘Capture’ against these Unmatched Documents to identify further relevant documents, which were subsequently reviewed
• following an initial review, exclusion of documents referring to other Post Office Limited IT projects which were not relevant to Themes 1, 2 and 3, with Horizon as an example

The application of these amended search terms resulted in a set of 5,756 documents to review.

We thank Post Office Limited, Peters & Peters and KPMG for their assistance and cooperation in the production of the Addendum, and do not consider that we have been subject to any unreasonable delays or obfuscations in our requests for information.

2. Observations

Kroll’s observations based on the targeted review of the additional datasets set out in the methodology do not change the conclusions contained in the Report. No significant or material changes have come about as a result of this review and as such, this addendum does not set out any significant revisions to the Report, we have set out where additional context or explanation has been provided to our understanding of the Themes set out in section 1.1 based on the documents reviewed.

2.1 Theme 1: internal communications responding to issues with Capture software

Kroll’s searches did not identify any evidence of systematic governance and reporting around the Capture software or issues related to it, particularly to individuals in senior management at Post Office Limited. However, we have summarised below some points of additional clarification and context that were identified from our review.

2.1.1 Governance

We have not identified any evidence that references a Capture specific governance body. We have identified reference to an Automation Transformation Steering Group (the ‘Steering Group’)^{[footnote 7]} around June 1998 which we understand was responsible for overseeing various Post Office Limited IT projects, particularly the roll out of Horizon. We have not identified any reference to Capture related issues in the context of the Steering Group.

2.1.2 Report from RNM referring to Capture errors

A report dated 13 December 1994 from a Retail Network Manager (RNM) from Post Office Limited’s North East region to the region’s Management Information team setting out the effect of errors in the Capture system^{[footnote 8]}.

The report stated that 3 sub-postmasters in the RNM’s area had experienced cash account errors related to the update of the software from version C40 to C45. The report further set out that the release notes accompanying this update were not clearly worded – i.e. that the release notes included a list of steps to be taken when applying the update, but did not clearly indicate that additional steps were required to be taken (merely including a reference to additional documentation previously published rather than setting out in detail the steps that were required to be taken).

We did not identify any further correspondence related to this report, and are therefore not able to comment on further steps taken by Post Office Limited in relation to this report.

We note that the Report makes reference to coverage of the C45 update in Focus, where the Capture team reported it had received feedback that the upgrade was considered a ‘huge improvement’ on C40^{[footnote 9]}.

2.1.3 Additional information on the timeline and rollout of Capture

Confirmation that the Capture helpdesk was run by Post Office Limited staff: We have identified a Service Level Agreement (SLA) contract between ‘Agency Development Centre’ and ‘Outlet Systems and Equipment’, an internal team that was responsible for staffing and operating the Capture helpdesk^{[footnote 10]}. The SLA makes reference to the Capture Helpdesk’s responsibility including the logging of faults, their progression and escalation.

Additional data points on the timeline and rollout of Capture: Post Office Limited discontinued supplying hardware alongside the Capture software in July 1995^{[footnote 11]}. According to an edition of Counter News, the Capture sales desk was withdrawn in December 1999, but the Capture Helpdesk continued to be available^{[footnote 12]}.

Capture implementation a condition in offer documentation: The Report included evidence of 2 instances where the installation of Capture software was a contractual requirement by the mid-1990s. We have identified 3 further instances where the offer letters to new sub-postmasters included the conditions that the software is installed within a given time period, or that ‘serious consideration’ be given to installing the software^{[footnote 13][footnote 14][footnote 15]}. We have not seen sufficient documentation to comment on whether this was standard practice across all branches, there is some evidence to corroborate the assertion of some sub-postmasters that the installation of Capture was a requirement (or at least may have been perceived by them as a requirement).

2.2 Theme 2: communication between the Capture development team and Post Office Limited control functions

We have not identified any evidence of proactive communications from the Capture development team to any Post Office Limited control functions informing them of potential issues related to the Capture software. However, we have identified evidence in one case of RNMs or other control functions (e.g. Investigations / Audit teams) requesting the Capture development team to test / comment on a specific case where a sub-postmaster suspected that their cash accounts may have been affected by a ‘Beryllium’ computer virus ^{[footnote 16]}.

The Outlet Systems Group within Post Office Limited provided a report in March 1999 commenting that the virus would not cause any accounting errors within the Capture program and asserted that, in accounting errors and incorrect figures on the cash account would be resolved by other controls (in particular error notices published by PFD).

Capture issues were raised by sub-postmasters during audits / investigations. Specifically, we have observed 2 instances of correspondence between RNM and the Post Office Limited Investigations team relating to the possibility of Capture-caused processing errors in Pensions and Allowances between 1997 and 1999^{[footnote 17][footnote 18]}. The correspondence between the Capture development team and the control functions reviewed by Kroll was limited to the specific cases, and we did not identify further correspondence related to these cases. As such, it is not clear how Post Office Limited ultimately responded to these cases.

2.3 Theme 3: Capture specific information about known sub-postmaster cases

Kroll’s searches did not reveal any additional information or correspondence relating to known sub-postmaster cases that warrant inclusion in this addendum.

1. Anonymised in order to protect identity of third-party system. ↩

2. Anonymised in order to protect identity of third-party system. ↩

3. A report summarising the count of the documents matching the search terms ran across each dataset. ↩

4. Kroll removed 2 individual terms from the search terms owing to a disproportionate set of results. This decreased the documents with hits from 268,705 to the current amount, and documents with hits plus families from 2.5 million to the current amount. ↩

5. This column sets out the number of documents that were responsive to the search terms within each dataset. ↩

6. This column sets out the number of documents with hits plus related documents (for example, an email plus all attachments). ↩

7. Automation Transformation Steering Group meeting minutes for 23 June 1998. ↩

8. Report from RNM in NE Region to NE Management Information team, 13 December 1994. The document does not describe the function of this team. ↩

9. Focus, December 1994. ↩

10. Capture Service Level Agreement, signed 29 March 1994. ↩

11. Bringing Technology to Post Offices and Benefit Payments, Question & Answer Brief, 27 November 1996. ↩

12. Counter News, 16 to 22 December 1999. ↩

13. Letter from RNM to Sub-postmaster confirming appointment, December 1996. ↩

14. Letter from RNM to Sub-postmaster confirming appointment, September 1993. ↩

15. Letter from RNM to Sub-postmaster confirming appointment, May 1995. ↩

16. Letter and report sent by Outlet Systems Group to Sub-postmaster, March 1999. ↩

17. Letter from RNM to NE region investigation team, March 1996. ↩

18. Report on suspension of Sub-postmaster by investigation team, May 1998. ↩