Thorny government IT issues

March 2014

ThornsThe faltering Universal Credit follows in Operational Strategy’s footsteps, says David Bicknell, with the government’s high profile benefits grand design mirroring another similarly challenged IT project from 30 years ago.

Those with an interest in the future of public sector IT - either those delivering it or, as a citizen, using it - will have noticed that the thorny problem of government IT projects has hit the headlines again.

In the last parliament it was the National Programme for IT (NPfIT) in the health service. In the last few weeks, the real truth behind Universal Credit has started to emerge.

It is nearly three years since Work and Pensions Secretary Iain Duncan Smith kicked off Universal Credit, a new single payment to replace a range of working-age benefits at the Conservative Party Conference. The target for the project and the IT behind it was always ambitious, and from its earliest days there were queries whether the project’s timescale was too ambitious.

However, ministers and civil servant kept on insisting that the criticism was wrong and that the project was on track and would be delivered on time and on budget.

Four senior responsible owners for Universal Credit came and went. The project lost its Chief Information Officer, Philip Langsdale, who passed away last Christmas after only three months in the role. The latest project leader, Howard Shiplee, has been recruited largely on the back of his Olympics success as construction director.

Shiplee joined in May 2013, not long after the Major Projects Authority, the Cabinet Office’s project troubleshooter, expressed serious concerns in February 2013 about programme implementation and the department’s lack of a detailed blueprint and transition plan for Universal Credit. In response to these concerns, the head of the Major Projects Authority, David Pitchford, asked to conduct a 13-week ‘reset’.

A stark picture

That the project should be in the headlines now is largely due to the publication of a critical National Audit Office (NAO) report which lifted the lid on the project’s woes.

The report paints a stark picture of weak programme management, over-optimistic timescales and a lack of openness about progress, as well as Whitehall friction between DWP and the Cabinet Office over security and the use of an agile development methodology.

The NAO report says that, despite commissioning specialist advice on agile development methods, the department struggled to incorporate an agile approach into its existing contracts, governance and assurance structures and ‘repeatedly redefined its approach’.

In particular, the report says, ‘the Cabinet Office does not consider that the Department has at any point prior to the reset appropriately adopted an agile approach to managing the Universal Credit programme.’

The NAO explains that the government had never tried to use agile methods, whereby an iterative and collaborate approach is used for IT development, on a major programme before.

However, it adds, the DWP recognised the associated risks. For example, the department was managing a project which had more than 1,000 people working on it, using an approach usually used in small collaborative teams, and had not defined how it would monitor progress or document decisions.

The report adds that the DWP was confined by needing to integrate Universal Credit with existing systems, which use a more traditional ‘waterfall’ method to managing changes, and was working within existing contract, governance and approval structures.

When the programme was reset, the Government Digital Service (GDS) was brought in to try to redesign the system and processes. And since the reset a cross-government ministerial oversight group, including ministers from the Cabinet Office and HM Treasury, has taken greater control of major decisions.

Agile 2.0

The NAO says the DWP decided to pursue an agile development path in December 2010, before moving to what it termed ‘Agile 2.0’ in January 2012, as a measure to try to integrate agile development with its legacy updates, which use a ‘waterfall’ method.

Agile methods offer a fundamentally different approach to tackling business problems compared to traditional methods and tools, the report explains. Agile processes mean technical work can start on programmes before requirements have been finalised, and so can be suitable for projects with shorter timescales.

According to the 2001 ‘Agile Manifesto’, published by a group of software developers, the core philosophies behind agile development are individuals and interactions over processes and tools, working software over comprehensive documentation, customer collaboration over contract negotiation, and responding to change over following a plan.

Another NAO criticism is that throughout its development of Universal Credit, DWP has lacked a detailed view of how the project is meant to work. The NAO suggests that the department was warned repeatedly about its lack of a detailed ‘blueprint’, ‘architecture’, or ‘target operating model’ for Universal Credit and although throughout 2011 and the first half of 2012 it made some progress, the concerns were not addressed as expected.

By mid-2012, that meant that DWP could not agree what security would be needed to protect claimant transactions and was unclear about how Universal Credit would integrate with other programmes.

Another problem was the project’s culture, which according to the department’s permanent secretary - essentially the department’s chief executive - was different to that on all other projects within DWP which is one of Whitehall’s biggest departments.

A fortress mentality

It is easy to imagine from that development how the Universal Credit team acquired what the NAO describes as a ‘fortress’ mentality within the programme, and a ‘good news’ reporting culture that limited open discussion of risks and stifled challenge. DWP had ring-fenced the Universal Credit team and allowed it to work with a large degree of independence.

As well as a lack of transparency and challenge, the Universal Credit team also had inadequate financial control over supplier spending - there was limited understanding of how spending related to progress and insufficient review of contract performance - and ineffective departmental oversight, which meant that DWP has never been able to measure its progress effectively against what it is trying to achieve.

That has led to continual problems with governance, changed governance structures and, during the ‘resetting’ of Universal Credit in early 2013, the complete suspension of the programme board.

What all this means for the Universal Credit programme, the NAO says, is that DWP will have to scale back its original delivery ambition and reassess what it must do to roll out Universal Credit to claimants. That means revising the programme’s timing and scope, particularly around online transactions and automation. That in turn means that Universal Credit will be more expensive and complex to administer than originally intended, while delays to rollout are likely to reduce the expected benefits of reform.

The Universal Credit team will be working together with GDS to ‘take the best of the existing system and make improvements’, though as the NAO suggests, DWP does not yet know to what extent its new IT systems will support national rollout. £34m worth of new IT assets - amounting to 17 per cent - have already been written off, and current ’pathfinder systems’ have limited function and do not allow claimants to change details of their circumstances online as originally intended.

The good news is that the problems of Universal Credit are belatedly beginning to come to light, which means they can start to be solved, though ultimately the project’s scope and timetable may have to be changed, to project managers’ and politicians’ embarrassment. A change in timetable and scope?

The project is due to be delivered in 2017. And although Duncan Smith, Devereux, Shiplee and the Major Projects Authority all believe this target to be achievable - indeed Duncan Smith still insists that the project will be on time and on budget - rarely do government projects achieve their targets.

Perhaps significantly, Prime Minister David Cameron recently told the House of Commons Liaison Committee that he was not religious about the 2017 date in response to a question from Work and Pensions Select Committee chairman Dame Anne Begg.

Cameron said, ‘The Secretary of State was questioned very closely in the House of Commons. That is the department’s position - they are shooting for 2017.

‘But the key thing is getting the early part of the introduction right. The more you can test out and hold pathfinders and get people on to universal budget and then start to take existing benefit recipients on to universal credit - the more you can get that right in the early years, the more chance you have of hitting your target of total rollout.

‘My view is this is a good reform that will make work pay, that is widely supported across politics and other sectors. So we need to get it right. But we shouldn’t be religious about timings. We should be religious as it were about the overall concept of what we are trying to do.’

That sounds as if the ground is already being prepared for Universal Credit to take longer than expected to be delivered. If so, it will largely have followed the same path as Operational Strategy, another benefits computerisation project in the 1980s, which also failed to meet expectations, and which was also the subject of repeated claims of success.

DWP however, is pressing on regardless. Its optimistic statement following the publication of the NAO report says, ‘We are committed to delivering it [Universal Credit] on time by 2017 and within budget. ‘Under this new leadership we are making real progress and we have a plan in place that is achievable and safe. The NAO itself concludes that Universal Credit can go on to achieve considerable benefits for society.’ Let’s hope they’re right.

David Bicknell is the Editor of Government Computing.

Universal credit: some key facts

  • £2.4bn - expected cost of implementing Universal Credit, to 2023
  • £425m - spending to April 2013
  • £38bn - Department for Work and Pensions estimate of the net benefit of Universal Credit between 2010-11 to 2022-23 onwards (taken from its December 2012 business case)
  • 184,000 - projected number of Universal Credit claimants by April 2014
  • £396m - planned IT investment in the current spending review period, from the May 2011 business case
  • £637m - planned IT investment in the current spending review period, from the December 2012 business case
  • £303m - IT investment to April 2013

Image:iStock/154997194

Comments (3)

Leave Comment
  • 1
    Bob Shallcross wrote on 12th Mar 2014

    What are the parallels with Operational Strategy? What are the lessons that have been ignored? This article is all pastry & no filling.

    Report Comment

  • 2
    Paul Higgins wrote on 12th Mar 2014

    The Operational Strategy was delivered and was a success. The Strategy included systems that supported Income Support, the Social Fund, State Pension, Disability Living Allowance, Incapacity Benefit, Jobseekers Allowance and War Pensions to name the more well known. Additionally there were linking systems to provide audit, common customer data, and fraud detection. The majority are still running, many since 1989 or the early 90s. How can they not be seen as successful?

    Report Comment

  • 3
    Geoff Sharman wrote on 13th Mar 2014

    Sadly, UK governments of both left and right have a poor track record on large projects. I was tangentially involved as a consultant on the Child Benefit project under Patricia Hewitt's reign. It would take too long to list the errors committed but the killer was the design of a particular transaction which achieved only 10% of its target interactive performance, due to very high levels of database access during the online transaction. Presumably, this arose because of inflated requirements.The project was quietly shelved a year after an announcement that "we are taking time to get this right".

    The combination of over-ambitious requirements, deadlines set by election timetables, and the reported lack of transparency seems fatal. A culture change is definitely needed, starting with a review of whether existing systems can be adapted to new requirements.

    Report Comment

Post a comment