Unrealistic timetabling led to the initial failure of Universal Credit, according to the Institute for Government

Job Centre - Image credit: PA images

Over-ambitious timetabling and changes to UK government IT project rules almost put an end to Universal Credit, according to a report from the Institute for Government (IfG).

The Universal Credit programme – pushed through government by the former work and pensions secretary Iain Duncan Smith – aimed to combine six in-work and out-of-work benefits into a single, simpler system.

However, it was dogged with mistakes from the outset and the whole project was reset in 2013.

Now, with the project running five years behind its initial schedule, an in-depth analysis by IfG senior fellow Nicholas Timmins has concluded that something “that is recognisable as Universal Credit” is likely to emerge at the other end.

Timmins' report – which focuses on interviews with civil servants, politicians and project leaders – looks at the reasons for failure, its move to recovery and what it must do next.

In his assessment, Timmins said that the single biggest cause of its initial failure was an unrealistic timetable that led to the Department for Work and Pensions trying to build Universal Credit before it had mapped out the detail of how it should work.

However, he added that “everything that could go wrong did go wrong” over the course of the project.

Issues ranged from resistance from the Treasury, which led to personal confrontations between Duncan Smith and the then-chancellor George Osborne, to five changes in leadership in just a year.

A number of problems related to technology and IT, in part due to changes in the UK government’s attitude to IT projects.

At the same time as Universal Credit was being launched, the Cabinet Office ruled that departments must stop signing long-term, high-cost IT contracts with single suppliers.

In addition, government was pushing for all services to be “digital by default” and wanted IT projects to use agile working, which had been a mainly private sector principle.

Timmins said that, despite the DWP having a better IT record than many other departments, it had no experience of using agile methods, which call for small-scale, iterative testing with fresh code released regularly.

Instead, it tried – unsuccessfully – to merge the existing ‘waterfall’ working – where suppliers design the code, then test it and then implement the whole system in one go – with agile working.

Similarly, Timmins said that the DWP took digital by default to its heart “without, as it turned out, knowing how to do it”.

However this not only posed a security risk – claimants inputting their bank details online were given access to departmental computers – it would also make it hard for those without digital skills or computer access to claim.

The IfG’s briefing paper argues that this failure to engage with users early on was a major stumbling block that was only addressed at the 2013 reset, when the Government switched to “digital as appropriate”.

Taken together, these factors demonstrate that the department had tried to implement “too much innovation”, the paper said.

“With stretched timetables and inadequate resourcing, it can be attractive to replace the old ways of doing things with a new technique, approach or technology,” it said.

“But big, high-profile projects are not necessarily the best places for testing new approaches…Where new approaches or technologies are used, they need to be tightly controlled and managed.”

As part of the 2013 reset, a new digital team made up of staff from the Government Digital Service was set up in London – this was away from the original digital team in Warrington, but closer to the policy team.

Although this is seen as being a major factor in bringing the digital service up to scratch, Timmins’ report reveals some issues with the new team’s approach.

He quotes Lord David Freud – now minister of state for welfare reform – as saying that the GDS team were “very naïve” about how complex it would be to build Universal Credit.

“They were messianic about building the front end, doing it in an agile way, front facing, with their beautiful apps, and they were right about all of that,” Freud is quoted as saying.

“But they had no grasp of how complicated it was to tie the front end to the legacy back-office, these old and creaky legacy systems we have, with which it had to work.”

A further issue Timmins was told about was that the new digital team would not discuss the previous work done by DWP and its IT suppliers – instead attempting to rebuild from scratch.

However the project’s leader at the time, Howard Shiplee, could not understand how the DWP’s IT team (“a very large group of clever people”) had got the whole process wrong.

He commissioned PwC to review the work done by the Warrington team and its suppliers, which found that the work was of good quality.

Given this information, Shiplee told Timmins that the ideal solution would have been to marry the front end apps that the London-based digital team had developed with the work already done.

But, he said, “entrenched attitudes” made that impossible, and so the project was reset with a “twin-track” approach.

This continued the rollout of the ICT developed by the external suppliers, dubbed the “live service”, while simultaneously delivering the “digital” service, which aimed at improving functionality.

The system now has 280,000 people using the “live” version, and around 20,000 the “full” or digital service, Timmins estimated.

He said that there was still “an awful lot” that needed to be done, including a full assessment of the roll-out of the digital service, which is being constantly adjusted.

For instance, the borough that was the first to trial the full service has found “mixed” results, with Timmins’ research indicating that it may increase pressure on local government as councils have to help people who are not able to cope with the online approach to apply for their benefits.

However, he said that the DWP had made steps to build up its in-house IT expertise, and was engaging more closely with the staff and users in its transformation projects.

Nonetheless, Timmins concluded that the system remains “very much a work in progress”. He added: “Recovery is not the same as recovered.”