Written evidence submitted by the Thorney Island Research (DCG0006)

The Digital Centre of Government

Evidence to the Science, Innovation and Technology Committee

1. This submission addresses the following points in the call for evidence:

• Are there any case studies that the committee should consider as part of its inquiry?
• What lessons do previous and contemporary digital transformation initiatives offer for the digital centre?
• How should the digital centre be delivered in a way that ensures equitable access to public services?

2. It uses as a case study the UK government’s organisation of its digital centre during the period 1994 to 2024. There is good documentation available covering this history,[1] and numerous reports from the National Audit Office and select committees on specific aspects. Two problematic projects are highlighted as part of the case study: Universal Credit and the Rural Payments Agency’s farm subsidy payments system.
3. It briefly summarises the many organisational formations and changes at the centre of the UK government over the last 30 years tasked with delivering the “digital transformation of public services”. Functions have been distributed across these structures in various ways and problems have arisen from some of these combinations. The current proposals raise red flags, based on harms to citizens that have previously occurred, leading to our recommendations.
4. Thorney Isle Research is a research and advisory body specialising in the interactions between technology, public policy, governance, and public administration. The author of this evidence is a former civil servant with over 20 years’ experience in central government information technology (IT) units, including some discussed in this submission. Thus, it draws on personal familiarity and that of former colleagues as well as independent sources. Note that the term “e-Government” was the predecessor to “digital government”.

Summary of recommendations

5. Based on the case study analysis of 30 years of the organisation of digital functions in the UK government, provided below, we recommend the following:

1. Clarify what the definition is of “public service”, what the objective is in relation to the “transformation of public services”, and how progress will be measured. These clarifications should be precise and placed in the context and language of government i.e. policy design, development, implementation and outcomes; policy instrument selection; legislation; and the operation of administrative processes.
2. Maintain a robust separation of the objectives, management, operations and ministerial oversight between those central digital teams that develop and support technical products and methods and those that set strategy, policy, standards, and mandates, or exercise financial control — even if, or especially if, they are in the same department.
3. Challenge techno-solutionism i.e. searching for or promoting uses for a particular tool instead of seeking the best solution to a real policy or operational problem.
4. Clarify what is mandated to departments, how that is enforced and by whom, and what the sanctions are.
5. Recognise and resolve the conflict of accountabilities that arise when central units make decisions affecting departmental projects or expenditure, or provide products.
6. Combine all business case elements including technology acquisition into a holistic case for a policy or programme, for a single process for scrutiny and approval (presumably by HM Treasury).
7. Maintain a separation between project and programme assurance and digital financial, technical and procurement functions.
8. Ensure public policy work maintains a focus on the long-term best result for the country, undistracted by internal government tactical influences. Ensure sufficient ministerial bandwidth to give sufficient attention to both.
9. Reframe the thinking about digital government from its default assumption of transactions and information on the web to thinking in terms of policy implementation that is inclusive while making the best use of technology.
10. As a corollary to the above, ensure technical teams understand the core business of government and its legal and constitutional context, and that it is completely different to commercial service management.
11. Challenge or prohibit the adoption of exclusionary digital designs.

Case Study: A history and analysis of 30 years of the organisation of digital functions in the UK government

Central “digital” functions

6. Looking across the last three decades, there have been several distinct functions relating to “digital government”, i.e. the application of digital technology across government departments and agencies. They have typically been found in central departments (in Cabinet Office, Treasury, or attached to No 10).

• Strategy, common requirements specification, standards, operational guidance, rule-setting
• Common product development, support and maintenance (including government web sites, verification, AI)
• Cross-government data strategy, data sharing policy
• Procurement
• Business case and expenditure scrutiny and approval
• Project assurance
• External relations (local government, national bilateral, EU, UN, OECD etc)
• Head of Profession (IT/Digital Data & Technology), professional development planning and cadre management

“Digital” public policy areas

7. As well as the above, there are many topics of policy and legislation (such as data protection) relating to technology that have generally been housed in policy departments such as business, culture, innovation etc, and now DSIT. Sometimes however, aspects have been placed in central departments — usually the Cabinet Office — alongside the above. These include, for example:

• Broadband supply
• e-Democracy
• Digital exclusion/inclusion
• Digital products and services regulation
• Identity services

8. Some public policies such as data protection (including automated decision making) and digital exclusion have an effect within as well as outside government.

The history of central “digital” organisational formations

9. This is a brief and approximate (full detail would be too lengthy[2]) overview of the configuration of central digital units since 1990.
10. Pre-1996. The Central Computing and Telecommunications Agency (CCTA, an executive agency of HM Treasury) evolved from origins in the 1950s to become a centre of technical and procurement expertise for government computer and telephony systems. It provided support to departments on IT systems, strategies and security. It was the principal source of advice to ministers and officials on technical matters and managed external relations. It launched the first government web site in 1994 as a source of government information.
11. 1996 – 2000. The Central IT Unit (CITU) was created in 1996 in the Cabinet Office (CO) and took over some of CCTA’s policy, strategic, external relations and standards functions to focus on a strategy for government use of the internet and other emerging technologies. The Performance and Innovation Unit (PIU, attached to No 10) produced reports and recommendations on e-Commerce and e-Government.
12. 2000 – 2004. CITU was merged into the new Office of the e-Envoy (OeE), which took on policy roles for e-Commerce, broadband, e-Democracy and data sharing. It had internal groups for e-Government (CITU’s functions), information systems security, developing and managing a new government web site (UKOnline), and the e-Delivery Team (eDT) that built the Government Gateway authentication system (still in use today). Based in the CO it reported jointly to CO ministers and business department ministers. CCTA merged with estate management functions into the Office of Government Commerce (OGC) which dealt with procurement and project and programme management guidance, and assurance for major IT and construction projects.
13. 2004 – 2012. OeE functions were split. Public policy issues were dispersed to various deptartments. eDT work went to DWP/HMRC. Directgov replaced UKOnline as the government’s web site, moving in 2006 to the Central Office of Information (COI), an executive agency of the CO; in 2008 to DWP; then in 2010 back to the CO to become part of the Government Digital Service (GDS) in 2011. HMRC managed the BusinessLink website aimed at business users.
14. The e-Government Unit (eGU) continued in the CO with the digital government work (formulating IT strategy and policy, developing additional common components for use across government, promoting best practices across government, and security). It established the Government IT Profession in 2005. Later, the Head of Profession role moved from CO to DWP, and then to the Ministry of Justice (MoJ).
15. OGC functions were distributed in 2010, with procurement moving to the Efficiency and Reform Group (ERG) in the CO and renamed the Government Procurement Service (GPS) in 2011. Infrastructure commercial matters went to Infrastructure (UK) and project support and assurance to the Major Projects Authority (MPA), both in the Treasury. Digital inclusion policy was located in the Department for Communities and Local Government (DCLG) which produced a major strategy in 2008.
16. 2012 – 2021. The Government Digital Service (GDS, in the CO) was formed to “transform the provision of online public services”. Its gov.uk web site replaced both Directgov and BusinessLink in 2012. While initially alongside eGU, it took over all its functions and IT spend controls whereby departments had to get its approval to buy technical products and services. It promoted the overarching strategy of “digital by default”. The Government IT Profession was abolished in 2013 and later replaced by the Digital, Data and Technology Profession, run from GDS. In 2014 the GPS merged with departments’ procurement functions to form the Crown Commercial Service (CCS) as an executive agency of the CO.
17. 2021 – 2024. The Central Digital and Data Office was set up in the CO alongside GDS to take on strategy, policy, guidance and standards setting for the use of data and digital technology across government.[3] GDS’ other functions continued. The Incubator for AI (i.AI) was set up in 2023 in the CO and No 10 to “provide an incubation function to quickly test and evaluate ideas for applications of AI to improve the function of government and the delivery of public services”.[4]
18. The Centre for Data Ethics and Innovation (CDEI) was set up to develop tools and techniques that supported responsible adoption of AI in the private and public sectors, and renamed the Responsible Technology Adoption Unit (RTA) in 2024.[5] It sat in DSIT alongside digital inclusion and AI policy areas.
19. In 2024 after the General Election, GDS, CDDO & i.AI moved to DSIT.
20. 2025. GDS, CDDO and i.AI, the Geospatial data team and the RTA merged to become the new Government Digital Service in DSIT. Procurement remained in the CO.

Discussion and key lessons

Separation of functions

21. Almost always, central digital functions have been placed in the main central departments (Cabinet Office and Treasury) or under the umbrella of No 10. This is logical as these departments have the processes, networks and levers to work across departments for co-ordination, support, scrutiny or approval actions. These are not typically present in mainstream policy departments such as DSIT or DWP: they have the processes and networks to manage public policy and legislation, not typically found in the Cabinet Office. As well as these practical matters, without clear separation, some public policy may inappropriately be influenced by, or influence, internal government digital aims (such as regulating AI, or data protection). Consequently, combining those types of functions, as for example in the Office of the e-Envoy, has not proved durable.
22. There are important lessons to derive from the long history of separation or combination of the central functions relating to digital government.
23. The creation in 1996 of CITU separately from CCTA in order to develop a strategy for the government’s use of the internet illustrates a desire for innovative thinking and developing new ways of working unencumbered by existing practices. This worked well. The creation of CDDO alongside GDS, taking on some strategy, standards and policy functions, was similar. The distinct role of CDEI/RTA is another example of a unit set up to stand back from current activity and look beyond government.
24. Maintaining a separation of project scrutiny and audit from all other functions is important and has largely been maintained.
25. Support for the technology profession within government has generally been centrally located. In common with some other government professions, the Head of Profession role has sometimes been held by a senior official in a major department, to ensure the profession aligns with departmental needs. Again, separation has benefits.
26. However, a central unit that develops common digital products or services (e.g. gov.uk, UK Verify, AI tools), advocates and supports particular methods (e.g. agile development, the GDS Service Manual with design standards), or provides other resources to departments (e.g. technical staff), can create conflicts of interest if that is combined with strategy development, power to create and mandate standards, scrutiny, or power to approve or block expenditure or procurement. A closed system emerges that restricts innovation. Consequently, the absorption of CDDO and RTA, and possibly spend controls, into the new GDS raises concerns.
27. The Government Digital Service from 2012 to 2015 is illustrative, notably in the development of Universal Credit (UC) and the Rural Payments Agency’s farm subsidy payments system (RPA). In both these cases, harm accrued to both the public and to the government (as extensively documented by the NAO and select committees). The issues observed apply to the current DSIT proposals.
28. In essence, there was evidence of conflicting priorities and motivations, and more than a hint of the self-serving use of blocking or political power to drive the adoption of chosen approaches or products, or to prevent the use of unfavoured products or suppliers. This included for example the requirement for UC and RPA to adopt an interactive web-based solution built with agile methods and deploying (the non-functional) UK Verify authentication process when these were inappropriate. More widely, friction between departments and GDS was evident, and clear accountability for projects was muddied.[6] The following quotations tell part of the story.

It had become increasingly clear over the summer [of 2012] that the Cabinet Office was anything but on side [with UC]. Wedded as it was to its agile and its ‘digital by default’ approach, distrustful of the major suppliers, and deeply worried by the view of the Communications-Electronics Security Group (CESG) – the information security arm of Government Communications Headquarters (GCHQ) – that the system was insecure, in October it used its power to approve contracts to block the next big step in the IT development – a £55m infrastructure and network contract, which included security.[7]

But the development of the ‘agile’ [UC] digital service was immediately hampered by a spat over how quickly staff from the GDS were to be withdrawn. According to the leaked minutes, ‘GDS wished to accelerate their withdrawal from the design-and-build team to allow DWP to take ownership. However, as GDS have supplied most of the expertise and resource to date, and a recruitment exercise needs to be undertaken to fill the technical vacancies, there is therefore the likelihood of some delay.’ Some in the DWP saw the withdrawal of staff referred to in the minutes as almost an act of pique on [Minister for the Cabinet Office]’s behalf.[8]

A report published [in December 2015] by the National Audit Office (NAO) spending watchdog found that “personal rifts”, frequent leadership changes, and a failure to coordinate the work of the RPA – an agency of the environment department (Defra) – with the Government Digital Service (GDS) team sent in by the Cabinet Office had led to serious problems with the programme, which was overseen by four different Senior Responsible Owners in just three years.[9]

Asked by [a Conservative PAC member] whether the GDS focus on a digital-by-default system had been a distraction from the RPA’s “important emphasis” on ensuring payments were made on time, [the RPA Chief executive] replied: “There is that possibility.”[10] [Civil-service speak for “Yes”.]

29. In addition to the issues described above, there is little logic in separating spending controls for IT from overall business case and expenditure approval for the project or programme to which it contributes. The business case and associated financing is for the output or outcome of a policy implementation, given a necessary and sufficient set of inputs. Separate authority over one input— like having a special process for approving the cost and source of the concrete to build a building — can only sub-optimise the case and project timetable. In central government, it disturbs the basic accountability roles of ministers and departmental Accounting Officers.

… good governance requires authority and accountability to be aligned: that those charged with responsibilities should have sufficient authority to be able to discharge those responsibilities   [11]

Terminology

30. Throughout the period covered, there was no clear definition of what “digital transformation of public services” actually meant, and therefore no clear objectives or measures for it. The most specific statement of a target came in the period 1997 to 2000, in the form of a proportion of government transactions with citizens and business being online. Perhaps arising from this, the implicit assumption throughout this period was that “digital transformation” equated to putting transactional forms, processes and information on a web site. Indeed, the clearest evidence of this was provided by GDS itself when in November 2024 it claimed[12] that

There are over 10,000 document-based forms published on GOV.UK.

31. As commentators have observed, using the phrase “putting lipstick on a pig”, this did not change the underlying administrative functions or effectiveness of government and thus scarcely warranted the label “transformation”.[13] Further, while there are very many such transactions, public-facing transactional instruments are but one aspect of the administration of government’s policies so focussing on them is a distraction from seeing the whole picture of policy development and implementation.
32. Related to this is the commonplace use of the terms “service” and “user”. “Service” has been used loosely[14] to refer to an organisation (e.g. HMRC, DWP, DVLA), a policy domain (e.g. taxation, welfare, vehicle registration), a policy implementation (e.g. Making Tax Digital, Universal Credit, State Pension, Vehicle Excise Duty), specific public-facing functions (e.g. tax collection, claims processing) or back-office functions (as in shared accounting or HR support), or transactions within functions (e.g. online VAT or VED payment or benefits claims). In relation to digital government, its interpretation tends towards the last of these, in contrast for example to its use in Probation Service, Prison Service, etc. “User” derives directly from “user of a computer system”, leading again to an image of the online transaction model.
33. This all has led to a lack of clarity about what “digital transformation of public services” means, and a focus of effort on making web sites the medium for policy administration regardless of whether that is optimal or gives equitable access to those who need it.

Digital exclusion

34. Access online to relevant administrative information (e.g. what licences a business needs, which benefits someone can claim) is hugely convenient, as is making applications, receiving licences or paying fees through online transactions. On the public sector side, automating many transactions and speeding up and simplifying the publication of information, indeed spurring the streamlining of some processes, has improved the operations of many public bodies.
35. However, all this requires that the population of individuals and businesses affected by any policy implementation are competent users of the internet. The Ombudsman is clear on what constitutes good public administration. The Rule of Law implies equality before the law. Both say that the design of public administrative processes must ensure full access by all those who may need them.
36. This is not the case for all web-based processes. This begs the question “Is it ethical/lawful for a public authority to expect or require people to fulfil a statutory obligation or claim a statutory entitlement by a means they cannot use?”
37. Throughout the period discussed, there have been several government strategies for digital inclusion, and initiatives such as “Assisted Digital”[15] to help people access online government processes via third parties. None has solved the challenge that there is a big overlap between those people who depend on public sector support and those who do not use the internet, nor altered the web-site focus of digital government.

Consistently across the different measures, a range between 15% and 20% of UK citizens, children or households are affected by limited digital access.[16]

Overall

38. Despite all the rotations of functions, the same strategy — in terms of statements of aims and practical actions — has been repeated many times over 30 years, all landing in the same place of a web site as the basis of “digital public services”. [17]

Yours faithfully

Paul Waller

Research Principal

21 February 2025