[Software]

[Aquamarine]

The full extent of how far short the UK government's expensive, digital-first identity assurance scheme has fallen was revealed today by a brutal report on Verify by the country's spending watchdog.Among other figures in the damning 32-pager was the number of people and government services using the service: currently at less than 20 per cent of early targets.

The latest assessment of what was intended to be the Government Digital Service's flagship programme comes from the National Audit Office, and will make for difficult reading in GDS as it lays out the huge margins by which it missed aims.

The abysmal performance includes missed targets on user numbers, uptake by government services, success rates and cost savings. It has rendered the scheme so unattractive that only the UK's Department for Work and Pensions (DWP) remains committed to it – largely because one-benefit-to-rule-them-all Universal Credit relies on it.

However, it isn't clear what will happen to DWP's use of the system after April 2020, which is when the government will farm Verify out to the private sector, having decided last year to pull the plug on public sector funding.

The NAO also questioned how departments would be charged to use Verify after April, and warned that the government might end up paying for 380,000 users to be "re-verified" now that two private providers have dropped out of the scheme.

The report is the latest in a long line of problematic assessments of Verify. Both the service and its predecessor have been through 20 internal and external reviews since 2011, which have repeatedly revised targets downwards.

The aim was for Verify to become the default way for people to prove their identity when interacting with the government. Those wanting to sign up are asked to pick from a list of identity providers – such as Barclays, Experian or the Post Office – which check documents like a passport or against data they hold, like credit histories. The person then gets a login for different services.

It was expected to cost some £212m, and bring in benefits of £873m between 2016-17 and 2019-20, but the latter figure was lowered to £217m in 2019, largely due to poor take-up.

In both the 2015 and 2016 business cases, GDS said 25 million people would use Verify by 2020. However, as the NAO said, the reality is "significantly lower" than these predictions.

As of February 2019 there were just 3.6 million people verified, just 2.3 million more than figures released in March 2017. Based on current trends, the NAO said sign-ups would reach 5.4 million by April 2020 – 20 per cent of the target.

The number of government services signing up to Verify is also much lower than expected. In 2014, this was estimated at 100, a figure that was revised to 50 in 2015 and 46 in 2015.

But in February 2019, just 19 government services were using it. Moreover, only eight of these have Verify as the only route, with HMRC's services all offering its own Government Gateway, on the grounds it didn't trust Verify.

This was a major issue for GDS's Verify: not only did HMRC have its own, established system that it carried on maintaining, it even developed a replacement, which was due to take over from Government Gateway at the end of last month.

HMRC said it didn't adopt Verify because the system couldn't deal with business customers or people acting on behalf of someone else – the taxman also felt the early iterations of Verify wouldn't encourage people to use digital services, which it was desperate to do.

The NAO noted that this echoed problems with the Cabinet Office's shared services programme, which faced implementation difficulties as departments said a centralised solution didn't meet their specific needs.

This had a huge impact on user numbers: GDS had thought it would capture all annual PAYE users – in fact, just 4 per cent of HMRC's customers use Verify.

There have also been major issues with success rates: in 2015, GDS said 90 per cent of people signing up should have their identities successfully confirmed by Verify's providers.

But overall, the success rate is at 48 per cent – and this doesn't include people who signed up but can't use the service they want to because the data they had verified doesn't match that on the government's systems.

For instance, 2018 stats from the Driver and Vehicle Licensing Agency (DVLA) show 38 per cent of people trying to access its services through Verify were able to set up an account – but 8.3 per cent of these gave info that didn't match DVLA-held data.

Problems with verification have also caused knock-on problems for departments: because only 38 per cent of Universal Credit claimants trying to use the service managed to do so, DWP's operational costs will rise by £40m over a decade.

Meanwhile, GDS's plans for the service to be self-funding by March 2018 also fell flat. To date, it has put in about £154m – although the NAO said the real figure was likely higher – and, a year past the deadline, Verify is still centrally funded.

To add insult to injury, GDS has failed to get departments to pay what they owe, as departments and agencies were supposed to pay £1.20 a year for each user who accessed their services through Verify.

Only one department, HMRC, has paid – despite invoices being raised for others. The taxman has handed over £6.7m for its use of Verify between 2015-16 and 2017-18, while a total of £2m remains outstanding from DWP, the Department for Environment, Food and Rural Affairs, DVLA, the Driver and Vehicle Standards Agency (DVSA), the Department for Business, Energy and Industrial Strategy and the Home Office.

However, it seems some of this is down to user error: the NAO reported that the DVSA's invoice was initially sent to the wrong address.

About that private sector plan
In the face of poor take-up and obvious problems with throwing money at the proverbial dead horse, the government said in October that it will only spend £21.5m more on it and intends to cut off all funding in April 2020.

However, the NAO said that in order for this to happen, the cost of verifying identities would need to fall by 95 per cent.

At the moment, the government hands over cash to providers for each successful identity verified. Although the aim was for these costs to decrease as users increased, this hasn't been the case, with costs remaining at more than £20 per identity verified between 2013-14 and 2018-19.

The new contracts drawn up in October – when the government announced plans to shift Verify onto the private sector – do offer providers less for each sign-up, with reductions as volumes increase.

Two providers out of the initial seven – Royal Mail and Citizen Safe – decided not to sign up to the new terms, and the NAO said it was likely that the 380,000 people would have to re-verify with another provider, which would mean the government paying twice for those users.

Ahead of the October 2018 decision, the government set itself two tests: first to ask whether introducing a lower level of assurance would boost take-up – it didn't – and second, to see whether DWP, HMRC and the DVLA would continue their commitment – and only DWP did.

After April 2020 there is even less certainty for these departments, because it isn't clear what the model for use will look like when it is led by the private sector providers.

"One possibility is that departments would procure and pay for Verify's services directly from providers," the NAO said. "It is not clear what price would be charged for verifying identities, as this would be set by the new market."

The NAO said the Infrastructure and Projects Authority plans to conduct a final review of Verify as it comes to the end of its time as a government project. That will assess GDS's assumptions that moving Verify into the private sector is a viable option, and what impact it will have on departmental Verify users, especially smaller ones that don't have alternative identity verification options.