Ingredients for a government IT cock-up - lack of vision, digitally-illiterate farmers and...er, GDS?

Profile picture for user slauchlan By Stuart Lauchlan November 30, 2015
Summary:
GDS doesn't usually come in for criticism when it comes to the digital transformation of government, but the National Audit Office has pointed a finger of blame in its direction as part of an overall critique of Defra's CAP payments shortcomings. Lessons need to be learned.

maxresdefault
Muck spreading

Here’s a cautionary tale of a public sector IT digital transformation programme that looks set to land the taxpayer with a 40% increase in its lifetime costs and facing the prospect of paying £180 million a year to the European Commission in fines.

And just to make matters worse, the system in question is one of the Government Digital Service’s (GDS) best practice exemplar projects. That led yesterday to the National Audit Office (NAO) issuing the normally untouchable GDS with a politically incorrect rap across the knuckles for its part in the shambles:

GDS did not provide the support the Department needed.

This is a long story, but an important one, with lessons for government transformation that reach well beyond this particular project. At the heart of all this is the Common Agricultural Policy (CAP) Delivery Programme, the responsibility of the Department for Environment, Food and Rural Affairs (Defra).

As part of the CAP, Defra pays out £1.8 billion to English farmers and landowners each year. The CAP is overhauled every 7 years or so, with the current regulations coming into effect since 2014. These required new systems to be introduced to handle the changed requirements as the existing ones were, according to the NAO:

beset by problems. Under the previous delivery arrangements, farmers had to apply separately for different elements of the CAP, creating a disjointed user experience. The systems used ageing software and were increasingly costly to maintain. Staff had to carry out many processes manually. This increased the risk of errors, which, in many cases, led to the European Commission (EC) imposing penalties.

But the new system didn’t fare much better.

Screen Shot 2015-12-01 at 10.30.15
Applicants were told to use Verify, the government’s identity assurance system, to register for the new service from October 2014. This was a mistake. says the NAO as:

It quickly became apparent that Verify was not sufficiently developed to assure the identity of a significant proportion of farmers, and did not therefore work as the Department expected. Although a small number of farmers were able to register through Verify, the majority of customers registered using the RPA’s existing customer registration process, supported by drop-in centres and RPA’s [Rural Payments Agency] telephone helpline.

This led, in March of this year, to a decision to pull the online applications and replace them with ‘paper-assisted digital’ applications. This meant that RPA staff had to resort to manual data entry and using hurriedly-developed interfaces to the old systems.

At this point, the new systems were not fully integrated, development of links to legacy systems required to make payments had ceased and accounting and payment systems had been de-prioritised and delayed. Meanwhile a new business case has had to be submitted to the Treasury to try to get things back on track. By this time, the cost of the proposed system was £215 million, 40% higher than the original estimate.

GDS criticism

As noted, GDS, brought into play in 2013, also comes under fire:

GDS committed to reducing overall costs, improving delivery confidence and building the Department’s digital capability to support approval of the business case and the adoption of new approaches unfamiliar to the Department.

Although GDS provided resources to the Department, its support was reported to be patchy. GDS provided limited continuity and insufficient insight into how to adopt agile on this scale. It was not able to identify and provide the systems integration skills required.

The NAO criticises the Cabinet Office for increasing the risk levels around the programme with a series of conditions intended to bring it in line with wider government digital strategy which were laid down as a prerequisite for funding:

Each of the changes had potential value in improving digital development across government but, taken together, the seven key changes significantly increased delivery risk.

This programme’s exemplar status became part of the problem, according to evidence from the Rural Payment Agency’s (RPA) chief executive in September with too much focus on:

being first on a number of key development areas, so compounding the difficulties of being a digital exemplar for Digital by Default, being one of the first to use some of the new contracts on government’s G Cloud and being the first to try the identity verification service called Verify in a live environment.

Screen Shot 2015-12-01 at 10.30.06
The NAO cites the use of the G-Cloud procurement framework as a case in point:

G-Cloud was not yet, in 2013, well established as a procurement route for government. Although it accelerated some procurements, there were delays
in establishing a hosting environment as a result of difficulties adapting the standardised contractual terms.

It also questioned the enforcement of a Digital by Default approach when dealing with farmers, as, according to a rather sweeping assessment:

Farmers have a lower average level of digital literacy than the general population, and internet and broadband coverage in rural areas is lower than in urban areas (15% of farm businesses lack internet access).

There was too high a turnover of Senior Responsible Owners (SRO) on the programme, four in succession, each coming from different departments - Defra, the Major Projects Authority, the RPA and GDS - and:

Not all had previous experience of leading complex government major projects on this scale.

They also seem to have had ‘issues’ with in-fighting becoming a major concern:

There were deep and persistent personal rifts at senior levels and at times these led to counter-productive behaviour by the Programme’s leaders. The differences in strategic vision in the design phase were not resolved. Senior people told us that they found it almost impossible to work together at times. Interviewees reported confrontational behaviour between senior Programme staff at the RPA and GDS.

Rifts between senior Programme officials went beyond the creative tension that is to be expected in a multi-organisational programme, and impacted on implementation and delivery as well as staff morale and stress, especially from mid-2014. The dysfunction and inappropriate behaviour at the top was very apparent to Programme staff at this time, and created a frustrating working environment for them, preventing the culture of trust and collaboration needed to deliver a large and complex programme.

Amyas Morse, head of the NAO, sums up:

The Department [for Environment, Food & Rural Affairs], the Rural Payments Agency and Government Digital Service have not worked together effectively to deliver the Common Agricultural Policy Delivery Programme. There are serious lessons in this episode for all three.

Lessons

So what are those lessons to be learned? Well, for a start, the mistakes demonstrate once again the need to understand a common set of priorities to deliver on a consistent vision. The NAO says the original vision and design here was narrow, focusing too much on procuring IT and not enough on wider transformation considerations.

Another aspect of the vision problem was that all the stakeholders involved had their own version! The NAO notes:

The RPA’s priority was to pay farmers accurately and on time, and minimise the risk of disallowance penalties, while the focus for Natural England and the Forestry Commission was to achieve environmental outcomes and build relationships with their customers. The Cabinet Office aimed to encourage innovation, reduce costs and develop learning across government as part of the government’s digital strategy to build digital services based on user needs.

There was also an element of ‘head in the sand’ that meant that emerging problems were not acknowledged or tackled:

The Department did not consistently and transparently acknowledge and respond to findings from external assurance. The pervasive good news culture meant that early warnings of possible failure were unvoiced or not adequately addressed, nor appropriate contingency plans developed.

Screen Shot 2015-12-01 at 10.30.24
The end result of all this? Higher levels of disallowance penalties, increased programme costs, poorer customer experience and difficulties paying farmers accurately at the earliest opportunity.

For its part, Defra is in defensive mode:

The new Common Agricultural Policy (CAP) is widely acknowledged as the most complex ever and a new IT system was needed to handle this additional complexity and to make claiming as simple as possible for farmers. While there was a problem with one part of the online interface that enabled farmers to put data directly into Rural Payments, the system has always worked and has successfully started making accurate payments to thousands of farmers on the very first day of the payment window. The Rural Payments system will be further improved next year to make it easier for farmers to apply for their CAP payments.

As for GDS, a Cabinet Office statement talks of “leading a monumental change to public services” and saving taxpayers £3.5 billion over the past three years. Beyond the mission statement and the reflection on past glories, it says of the CAP chaos:

The CAP programme was one of the earliest and most ambitious of these [public] services. We build services in a way that enables us to test and improve the system based on user feedback before it goes live, and gets away from the old days when problems were often not discovered until it was too late. We've learnt a number of lessons during the development process for this programme, which we take with us for similar programmes.

For its part, the NAO says that the lesson GDS needs to take on board are:

  • provide stronger written guidance and capability building for departments on agile management and governance for major projects and how it fits with traditional governance structures.
  • take into account departments’ capability and capacity to apply new methods and ensure sufficient flexibility and support in their adoption;
    adopt an approach that is more sensitive to the risks and culture of the organisations and teams involved, in order to ensure that the best use is made of GDS’s limited resources and that cost savings are delivered.
  • support departments in acquiring the management and technical skills required to apply agile at scale.
  • ensure it has appropriate accountability for programme delivery when significant changes to programmes are enforced
  • step in to resolve conflicts quickly when departments are not able to do so themselves.

My take

An old-school government IT cock-up but with digital at the heart of it. Plus ca change? Let's hope not.  This sort of thing isn’t supposed to happen anymore.

The finger of blame seems to point in a lot of directions here, but to me the most telling point seems to be the cultural clash between old school IT and the new digital way of doing. Resistance to change on the one side, an over-zealous evangelism on the other.

As is traditional, everyone solemnly swears to learn the lessons from the latest faux pas. For the sake of other major transformation programmes - yes, Universal Credit, that’s you we’re looking at! - let’s hope this time they do.

-