Universal Credit is one of the government’s biggest IT-based projects and the biggest test for agile in the public sector. It is due to start rolling out in April 2013.
This is what the Starting Gate report says on agile aspects of the project.
Agile
The challenging timetable for delivery of UC meant that DWP elected to use an Agile approach to the delivery. There is no evidence of such a methodology being used on a public sector programme of such scale and during the course of the review it was evident that there had been some initial scepticism to the use of such a methodology with a programme of this scale.
However, during the review there was overwhelming evidence of buy-in to the methodology at all levels up to and including the highest levels. DWP have set about thoroughly educating all involved on what can be expected from them and there was clear evidence within the interviews that this is being taken up enthusiastically.
There was a view that policy decisions being made later in the programme would pose a problem for delivery. This was countered by the view that the methodology should allow decisions to be made when they need to be made, which is in contrast to fixing requirements early in more traditional (‘waterfall’) methodologies.
On balance, the review team found that the use of the chosen methodology here was judged by interviewees to provide greater assurance of delivery in such an environment. The review team agrees with this finding.
In terms of the use of Agile within Government, DWP also have the best current experience via their Automated Service Delivery (ASD) Programme, which used a slightly less ‘lean’ version of the methodology based on an Accenture interpretation.
However, there are still valuable lessons that can be transferred from this programme and there exists experience that is being directly deployed on UC. The review team felt that whilst effectively piloting this methodology on a programme such as UC did pose a risk, this was acceptable in view of the risk of delivery out of line with expectations, for example in terms of timing or quality of service to the public.
Accenture remain involved in UC, although DWP have brought in consultants (Emergn) to provide an independent methodology not based on any ‘out of the box’ methodologies, but rather one that Emergn have tailored.
New contracts supporting this development are due to be awarded in June 2011 and DWP state that their use of this independent methodology will serve to remove any supplier advantage.
There was evidence that DWP have understood the need for decision-making delegated to the level at which the expertise exists, with the appropriate empowerment supported within the planned governance re-design.
There was also an acknowledgement that the right domain/business knowledge needs to be made available at the workshops that will drive the detailed design processes. It was also accepted that there is a continuing need for this knowledge to be made available and also that it will need to keep pace with the changing policy.
One key risk identified by DWP is how an Agile methodology will interact successfully with the various approvals processes that will come into play across the programme – most especially the ICT Spend Approval process (formally known as the ICT Moratorium Exception process).
Engagement has begun already with the Major Projects Authority (MPA) on designing the Integrated Assurance and Approval Plan (IAAP) that will ensure the correct internal and external assurance is brought to bear for the identified approval points. The production of this plan is seen by the review team as a key mitigating factor for the risk identified and it is recommended that this is produced, with MPA guidance, by the end of March 2011 at the latest. This may need fine-tuning as approval points are finally agreed.
Recommendation:
DWP, with guidance and assistance from the MPA, produces an Integrated Assurance and Approvals Plan (IAAP) by the end of March 2011.
As noted earlier, there are contracts that are relevant to this development that are being re-competed at this time, with a wish to award in June 2011. There was some evidence that the design of contracts to deliver in an Agile environment will require a different design in order to draw out supplier behaviour in line with an accelerated delivery environment.
There is a always the risk that any development methodology will fail to deliver and whilst this methodology itself provides early warning of failure, there is recognition that in such a circumstance the prioritisation of customer journeys with high-value returns would be needed.
There was much evidence of the reliance of UC on successful delivery of the HMRC PAYE Real-Time Information (RTI) programme. There was also recognition that whilst ‘just-in-time’ decisions as a consequence of policy development could be made within UC, the RTI requirement would need to be more rigidly fixed as the traditional ‘waterfall’ development methodology in use cannot so easily absorb such changes without consequence.
There was some concern that fraud would remain a major issue for UC and appropriate Information Assurance should be built into the requirement from the outset – rather than being a ‘bolt-on’. Also, as UC and its interface with PAYERTI [PAYE RTI] will become part of the UK Critical National Infrastructure, appropriate discussions should be maintained. There was evidence that DWP have gripped these requirements.
Overall, the use of an Agile methodology remains unproven at this scale and within UK Government; however, the challenging timescale does present DWP with few choices for delivery of such a radical programme.
That said, there has been evidence of strong support at all levels and DWP do have some expertise within their own organisation that they can call upon from the outset. The review team not only felt that an Agile development is an appropriate choice given the constraints, they also believe that DWP are well placed with their level of support, knowledge and enthusiasm to act as a pilot for its use at such a scale.
Compound failure
DWP has made a strong start in identifying risks to delivery. This could be developed further by thinking through the likelihood and impact of a number of risks being realised simultaneously (eg lack of synchronisation between reduced income and UC top-up, plus wrong employer data plus labour market downturn).and what the responses might be.
The programme could extend its preparedness by drawing on a wider range of experience the elements of recovery and their prioritisation; and test their robustness in advance, including an early warning system for Ministers.