Universal Credit: How did it go so wrong?

Early days

The year was 2010, and in the early days of the coalition government, plans were hatched in Whitehall for a £2.2bn programme for “Universal Credit: welfare that works”. The aim, according to the 2010 whitepaper, was to “make extensive use of online technology to allow people to better manage their claim and understand the benefits of entering paid work”.  

This was the DWP’s first major IT project using an agile methodology, relying on a more iterative method, in line with the then Cabinet Office minister Francis Maude’s “digital by default” initiative, led by the Government Digital Service (GDS), to move away from large, outsourced IT programmes. 

At the time, the plan was for the DWP to begin taking claims for UC from October 2013. But it soon became clear that this was a very ambitious (and unrealistic) target. The UC IT system was to be built in an agile fashion, which would set the project apart from previous government IT failures. Things didn’t go completely to plan, however.

In September 2012, Labour called for a halt to the programme because of unresolved IT problems, which the DWP claimed was nonsense. In November that year, the then programme director, Malcolm Whitehouse, decided to step down – the first of many changes to the programme’s leadership.

Documents – not released until 2016 as part of the long-standing FoI battle – show that DWP officials knew of problems with the programme in 2012, while senior figures continued to state publicly that the project was on time and on budget. By the end of 2012, the government had spent £638m on the IT for Universal Credit, with £441m going on design and development.

Unlucky (20)13

In early 2013, the warnings began to escalate. Computer Weekly headlines at the time read “Universal Credit has the hallmarks of another IT failure”, and “£200m of Universal Credit IT could be scrapped, MPs told”.  

At that time, not much detail about the project was publicly known, but the newly released internal project assessment reviews (PARs), undertaken by the Major Projects Authority (MPA), shed some light on what was going on behind the scenes.

The PAR undertaken in February 2013 stated that the roll-out of UC was expected to be complete by the end of 2017, but the document added: “The delivery confidence for the programme is Red.”

In fact, the review team was so concerned about the programme that it recommended “a complete rethink of the delivery approach, together with streamlining of potentially overly elaborate solutions”.

“We recommend that subsequent work should not proceed until the complete rethink of the delivery approach is concluded,” the February 2013 document said. “Historically, the programme had been purely IT-driven with little, if any, interface with the business and a reputation for impermeability.”

The review also criticised the DWP’s work on agile, saying the IT methodology lacked rigour and discipline, leading to an ineffective programme.

However, a key issue – and one that does not necessarily apply just to UC, but many government IT programmes – was that the review team noted that the approach “does not recognise key aspects of what is required to turn around and deliver a business transformation on this scale”.  

No blueprint

It also found there was “no blueprint or business architecture” in place, which it said was a “root cause of other critical problems”.  

The review also noted concerns about the stability and viability of budgets, particularly capital IT spend.

In June 2013, the MPA undertook another review. At that time, the programme had managed to deliver a blueprint, and the review team was criticised by the senior responsible officer (SRO) of the programme, Howard Shiplee, for the review’s timing, saying a “more useful timing” would have been beneficial.  

The June 2013 review was not as dire as the previous one, and seemed positive about the progress the programme had made, but there were concerns over commercial and supplier management.

“A clear understanding of the agile approach being used, where appropriate, by the programme must be there on both client and supplier sides,” the report said, adding that there was also a risk of cost estimates being wrong “due to the high-level nature of requirements at the tender stage”, but that these could be mitigated through proper contractual arrangements.

It is worth noting that, at this point, the planned £2.2bn cost figure for the programme had risen to £12.8bn. 

Suppliers did indeed become a bit of a problem for UC. As previously reported by Computer Weekly, the DWP was “seriously considering” suing its suppliers over their involvement. However, Accenture, HP and IBM had all previously written formally to the department detailing their concerns about the programme’s feasibility. 

A month after the June review, programme SRO Shiplee told a work and pensions select committee that “all is well” and that he had no worries about the success of UC. Then came news of a major reset. 

By midway through 2013, the DWP had already been urged to “reset” the programme, and work was ongoing. However, as the review documents show, no formal decision was made until December that year.

Hitting the reset button

On 5 December 2013, the secretary of state for work and pensions, Ian Duncan Smith, finally admitted that UC would miss its 2017 deadline. 

The reset of UC meant writing off part of the existing spending and adopting a twin-track approach, whereby a portion of the already created IT system would be used alongside the future digital service to support the nationwide roll-out of a fully digital UC system.

In March 2014, the MPA undertook another review. This time, the delivery confidence was signalled as Amber/Red. Despite seeing “evidence of improvement”, with the UC project now seen as a “transformation programme and not an IT programme”, the review team was still concerned.

“We have not found a single coherent integrated plan or clear target operating model,” the review said. It added that the period between the February 2013 review and the December 2013 reset had been “very challenging for all concerned”.

“There is evidence of poor behaviours and morale, and progress suffered,” it said. Although it applauded the programme for making progress with the implementation of the IT system, it said there were still serious issues that had not been addressed.

The review team said it did not see a “single, coherent, holistic plan across the UC programme”, adding: “There is no agreed contingency plan should the digital end state fail to deliver. We saw no detail on migration plans. Even the timing and scale of migration was unclear.”

It also said there was a lack of clarity around the build and re-use of components from the existing IT and that the affordability of the digital solution remained to be proved.

The DWP continued with the roll-out of its controversial welfare reform behind what critics called “a veil of secrecy”. For a long time, not much was known about what was really going on. In November 2014, the department began testing its digital service, while the National Audit Office was still unsure whether it would ever deliver value for money. 

Ploughing on

Then came the 2015 general election, and with it further secrecy and further criticism, when it was made public that the government would not reveal the full cost of UC until after the election.

The election came and went, and with a Conservative government now in place, the roll-out continued – as did the secrecy. In February 2016, the Public Accounts Committee criticised the programme for a lack of transparency, saying the lack of clear milestones for the roll-out of the UC digital service created uncertainty. 

At that point, the roll-out of the digital service was due to begin in May 2016, aiming for completion in March 2021. No problem at all, if we were to believe UC bosses at the time.

But this may not have been entirely true, if you believed the October 2015 MPA review. Although the review team was feeling better about the programme, at that point, having a “single strategic direction”, it was still concerned “that the scale of people change may have been underestimated”.  

“DWP is successful at rolling out new policies. But UC is different – it is transformational and will require fundamentally different ways of working,” the review said.

“There is a question about whether sufficient attention is being given to the overall customer experience/customer journey for claimants making the transition from tax credits to UC, which is important given that tax credit claimants form such a large percentage of the UC total.”

This issue continues today. To illustrate the point, back in 2010, when the welfare reform was first announced, charity Citizens Advice said: “We broadly welcome the direction of welfare reform proposed by the coalition government and we support the intention to make the benefit system simpler and clearer for recipients, and to make work pay.”

Fast-forward seven years, and the charity’s 2017 report called for the roll-out to be halted, as it had found serious problems with the UC digital service, including people lacking the digital skills to use the system and struggling to verify their identity.

The latter issue, in particular, is an interesting one. The UC digital service aims to use the government’s flagship identity platform Gov.uk Verify, but Computer Weekly revealed in January 2018 that currently only 35% of UC users could set up a Verify account online, 30% were not able to, and the remaining 35% could use Verify but did not.

The MPA was already worried in 2015 that Verify would not be able to deliver. “The review team heard feedback that confidence in this solution [Verify] being available is very low,” said the project review.

“The programme is therefore making a more cautious assumption of 40% online verifications. And there is work ongoing on an alternative solution. This might take [about] two years, but the timeline aligns with volume requirements.” 

Lessons learned?

These were all valid concerns at the time, but the public was none the wiser. If the public had been, would it have made a difference? Would the reviews have been more optimistic?

There are common denominators with most failed government IT projects – rapid staff turnover, particularly in senior leadership, failure to see the full picture, lack of skills, cultural and cost issues. Another thing they all have in common is the failure to admit the problems.

The DWP is not alone in constantly claiming there are no issues, when the truth is the complete opposite. Will Universal Credit and its digital service ever succeed? Only time will tell, but one thing is for sure – this wasn’t The Lego Movie, and not everything was awesome.