xxreportnamexx Contents

1Performance, costs and benefits

1.On the basis of a report by the Comptroller and Auditor General, we took evidence from the Cabinet Office and the Government Digital Service (GDS) about accessing public services through the Government’s Verify digital system.1

2.GDS created Verify with the aim of providing a single identity assurance service across government. It is designed to allow people to prove their identities so they can securely access a range of government services online (for example, claiming tax back or receiving benefit payments). Through Verify’s predecessor, the Identity Assurance Programme, GDS developed an identity assurance framework in 2012. It then developed the Verify platform, which went live in May 2016.2

3.The programme contracts out verification services to five ‘identity providers’, all of which are private sector companies. An individual wanting to access government services through Verify signs up with one of these providers to create a Verify account and log-in. The provider checks identifying evidence and documents such as passport and driving licence before verifying that individual’s identity. Providers are paid based on the number of people they sign up as Verify users.3

4.GDS spent £154 million on Verify and its predecessor programme from April 2011 to September 2018. In October 2018, the Cabinet Office announced that government funding would stop in March 2020. It has capped the amount it will spend on Verify between October 2018 and March 2020 at £21.5 million. GDS intends that responsibility for Verify will then pass to the private sector, including responsibility for investment to ensure Verify’s future delivery. Revised 18-month contracts were signed with five providers in October 2018 to cover the transition period to March 2020.4

Performance and departmental buy-in

5.GDS forecast in 2016 that 25 million people would use Verify by 2020 and that at least 46 government services would adopt Verify by March 2018.5 However, Verify has underperformed against both of these expectations. At March 2019, 3.9 million users were signed up to Verify and 19 government services were using it (11 of these services also provide alternative online options for people to access their services).6 GDS acknowledged that its original projections were “hopelessly optimistic” and said it had reduced targets for user volumes as part of its 2018 business case. It told us it was on track to meet these revised targets.7

6.GDS believes the reason take-up has been so low is because departments did not develop their online services as quickly as expected and therefore did not need Verify. The Cabinet Office also recognised that the right incentives did not exist for departments to adopt Verify, largely because it was expensive for departments to implement.8 GDS noted it was easier for new services to incorporate Verify into their systems. There were compatibility problems when a pre-existing system, such as a tax platform or Universal Credit, tried to adopt Verify, which led to further work and costs for departments.9

7.Some departments have continued to use and develop alternative ways for people to access their services online. For example, HMRC did not adopt Verify for all its services, in part because Verify could not deal with business customers or agents acting on behalf of others. Instead, HMRC continued to develop and enhance its own system, Government Gateway. Only around 4% of HMRC customers now use Verify to access its services.10

8.From 2016, GDS pursued several attempts to boost user volumes, such as offering Verify to local authorities and the NHS. However, no local authority currently uses Verify.11 GDS also introduced a new category of verifications that allowed people’s identities to be verified to a lower level of assurance (level of assurance 1 or LOA1, rather than Verify’s standard level of assurance, LOA2), but this too failed to increase user numbers significantly.12

9.The Cabinet Office acknowledged that it and GDS could not ‘bring the government market to the table’ and had gone on for too long trying to do this.13 The problems that GDS and the Cabinet Office faced in getting departments to adopt Verify share many similarities with other central initiatives we have seen, notably shared services. This Committee’s 2016 report on shared service centres found that — as with Verify — the Cabinet Office did not secure sufficient buy-in from departments, did not create an environment of collaboration among departments, and did not intervene effectively when things began to go wrong (in part because it could not instruct departments to keep to the migration timetable).14

Costs and benefits

10.GDS spent a total of £154 million on Verify and its predecessor programme between April 2011 and September 2018. This was less than GDS had anticipated, and is because the lower than expected volume of Verify users reduced the amounts payable to providers.15 However, GDS has not attempted to quantify the overall costs of Verify to government as a whole, and £154 million is likely to underestimate total costs across government.16 Departments have borne the costs of integrating Verify into their online services and incurred other costs resulting from problems with the Verify system. For example, large numbers of people applying for Universal Credit have struggled to successfully use Verify. Consequently the Department for Work & Pensions expects to spend around £40 million over 10 years on processing applications for Universal Credit manually.17

11.The Cabinet Office and GDS told us they regard Verify as successful because claimed benefits are higher than the costs of the programme. However, they acknowledged that Verify did not achieve the level of benefits forecast in its original business case.18 In 2016, Verify’s benefits were estimated to be £2.5 billion over the 10 years from 2016–17 to 2026–27. GDS now estimates that £366 million of benefits will be achieved between 2012–13 and 2019–20.19

12.The Cabinet Office and GDS believe that the benefit-cost ratio of the programme is 2.4:1.20 Its calculation includes supposed benefits up to 2019–20, but costs only up to September 2018. It does not factor in the additional costs that will be incurred until government funding stops in March 2020. These additional costs could be up to £21.5 million, the amount at which government has capped expenditure on Verify from October 2018 to March 2020.21

13.The National Audit Office was not able to replicate the claimed level of benefits for the Verify programme from the information that GDS provided.22 For example, the largest category of GDS’s estimated benefits is avoided building costs, or avoided spending by departments on building alternative identity verification systems. GDS reported validated savings of £111 million from 2012–13 to 2014–15 for avoided building costs relating to Universal Credit, an average of £37 million a year. It has continued to claim the same yearly amount — £37 million — for avoided building costs from 2015–16 to 2019–20, although this figure has not been independently audited or validated.23 GDS recognised that such benefits were difficult to measure, but said they adhered to the normal methodology approved by the Cabinet Office and HM Treasury.24

User experience

14.We understand that people have often found Verify “clunky” to use and many have faced problems even signing up with it in the first place.25 GDS assesses the programme’s performance at verifying people’s identities through a measure called the verification success rate. This rate measures the proportion of people who can sign up for a Verify user account, having had their identities successfully confirmed by a provider.26 Only 50% of people attempting to sign up were able to successfully create a Verify account.27 The Cabinet Office acknowledged that Verify’s identity checking was perhaps “not to the same standard” as other checks that take place in government.28

15.Government services using Verify each have different verification success rates. This is in part because different services require different levels of assurance, with higher levels of assurance entailing more checks. For example, GDS told us that for the health service pension, which requires a relatively low level of assurance (LOA1), 95% of users were able to sign up to Verify on a first attempt. For Verify’s best performing LOA2 services (which involve a higher level of assurance), successful verifications are much lower — the success rate for these services are in the range of 70% to 80%.29

16.The success rate for a service also depends on the demographic characteristics of those using the service. For example, Universal Credit is Verify’s largest customer, but only 38% of Universal Credit claimants can successfully create a Verify user account. GDS attributed this low success rate to people on Universal Credit not having strong digital footprints.30 GDS told us it had recently employed two people from the Department for Work & Pensions’ Universal Credit team to work with it to improve the verification success rate for Universal Credit.31

17.Even if people do manage to sign up for a Verify account, this does not mean they will be able to access the government services they actually want to use. This is because government services will compare an individual’s Verify details with any pre-existing data they have on that individual.32 For instance, if someone’s name details do not match, that person will not be able to access the relevant government service. GDS’s verification success rate does not measure whether people can get to the government service they want in the end; it measures only those entering the Verify service and leaving it. GDS told us it limits its success measure to the parts of the process it has control over to protect users’ privacy.33

1 C&AG’s Report, Investigation into Verify, Session 2017–19, HC 1926, 5 March 2019

2 C&AG’s Report, paras 1.1–1.2

3 C&AG’s Report, paras 1.3–1.4

4 C&AG’s Report, paras 2.2, 3.7

5 C&AG’s Report, para 1.7

6 Q 87; Government Digital Service, GOV.UK Verify Dashboard webpage, ‘GOV.UK Verify’ for week ending 31 March 2019, accessed 4 April 2019. Available at: www.gov.uk/performance/govuk-verify

7 Qq 27–30

8 Qq 31, 35, 48, 102

9 Q 95

10 Q 95; C&AG’s Report, paras 1.13, 1.18

11 Qq 57–59

12 Qq 70, 103; C&AG’s Report, para 1.10

13 Q 150

14 Qq 36, 102; Committee of Public Accounts, Shared service centres, Twentieth Report of Session 2016–17, HC 297, October 2016, Conclusions and recommendations, para 3

15 Q 30

16 Qq 38–40; C&AG’s Report, para 2.2

17 C&AG’s Report, 1.14

18 Q 30

19 Qq 29, 37; C&AG’s Report, para 1.18

20 Qq 29, 37

21 Q 105; C&AG’s Report, para 3.7

22 Q106; C&AG’s Report, para 1.19

23 C&AG’s Report, para 1.20

24 Q 106

25 Qq 155–157; Mr P.G. Slater (GVS0001)

26 C&AG’s Report, para 1.7

27 Government Digital Service, GOV.UK Verify Dashboard webpage, ‘GOV.UK Verify’ for week ending 31 March 2019, accessed 4 April 2019. Available at: www.gov.uk/performance/govuk-verify

28 Q 159

29 Q 95. For more details on the requirements for different levels of assurance, see C&AG’s Report, Appendix Two.

30 Q 95

31 Qq 153–54

32 C&AG’s Report, para 1.17

33 Q 93

Published: 8 May 2019