Assessing the UK’s Government Digital Service
Despite numerous troubles, Osbo gives it the green light
The Year in Review For the Government Digital Service, 2015 has been a year of challenge and uncertainty. Many issues have dogged the body, but some of the most high profile included the £215m IT fiasco for farmers' rural payment allowances and the departure of Mike Bracken, the man whose name has become synonymous with the body, amid rumours that the GDS was for the chop.
Yet, despite widespread expectations the Cabinet Office unit would be dramatically scaled down, the body emerged from George Osborne’s Spending Review at the end of November with a surprise £450m endorsement for the next four years.
It remains to be seen how much new money that figure amounts to in real terms, and a new digital strategy won’t be out until early next year.
But it was still a shock announcement to many who believed the mass exodus of GDS top brass was essentially its death knell.
In order to understand what the body badly needs to get right over the next four years, it’s worth reviewing where it's come up short over the last 12 months - and further back.
This is of interest because a) it's our money and ought to be subject to critical scrutiny, and b) the government still spends an outrageous amount of cash on ageing, expensive IT, so any attempt to push through changes to unlock that needs to be effective.
New IT methods, old mistakes
The year got off to a particularly bad start with the collapse of the digital interface of the Common Agricultural Policy (CAP) IT system in March, which was supposed to be one of its key exemplars for GDS in showcasing its "agile" approach to delivering digital technology.
Later in the year, the National Audit Office revealed that the IT fiasco has meant the system will be 40 per cent over budget at £215m and has led to disallowance penalties to the EU of up to £180m per year due to mis-payments caused by the system.
Of course, it’s not the first time the government, or even that particular project, has encountered major problems. But the incident left a considerable amount of egg on GDS’s face, as it had boasted the system would be a complete break from how IT had been done in the past.
Former GDS head Mike Bracken had said of the project: "It's really, really exciting. It was created so quickly, such a small amount of money compared with some of the big IT programmes that have preceded it.”
Fast-forward two years and the NAO and the Public Accounts Committee's autopsy of the project described deep rifts between GDS and the Rural Payments Agency, with their respective leaders displaying "deeply dysfunctional" and "childlike" behaviour. One MP described the role of GDS as having "all the power and none of the responsibility".
The programme's failure also raises some serious questions about what the future of large IT projects will look like. Currently, there doesn’t appear to be an "agile" approach to programmes that is working at scale.
That will be even more of a challenge as a number of large government IT contracts are supposed to be expiring (although we understand many have already been extended) and a new model is required if the government is serious about kicking its habit of costly contracts with system integrators.
However, early signs suggest the Department for Transport may be making progress by moving staff in-house and stepping away from big contracts. So perhaps the lesson for GDS is that it needs to work with departments to push for change, rather than take an "arrogant and commanding approach" or giving "drive by advice" – an attitude multiple sources have previously accused it of having.
Transactions yet to happen
The extent to which its 25 "exemplar" services, of which CAP was one, have achieved real digital transformation remains highly questionable. Not to mention the fact the programme took more than two years, and with the exception of the CAP system many of those services are very basic indeed.
Certainly the National Audit Office is of the opinion that there is as yet little evidence of cost savings due to digital transformation. The cost savings claims released by the Cabinet Office certainly look highly dubious.
Underpinning any serious move to digital transactions is the necessity of getting its online identity assurance system – Verify – off the ground. Both farmers and recipients of UK.gov's flagship marriage tax allowances were left unable to identify themselves online.
Once again, the GDS missed its sign up targets for Verify last month, which has now been four years in the making. It had hoped to get 500,000 tax self assessment users onto the system.
One of the main problems with the system is it only appears to work for textbook users, with an established online footprint. That could also be a major stumbling block as the system develops, as HMRC has committed to making its tax systems fully digital for users by 2020.
The Register also revealed that serious security considerations had been raised by an academic over the in-house-built Verify hub, an issue it will have to address if it is to gain acceptance from users.
Some sources and commentators have suggested that one of the biggest problems GDS has created for itself is adopting a "build everything in-house" approach, and effectively biting off more than it could chew by over-engineering rather than buying commodity off-the-shelf IT.
The successes of GDS – such as putting in spend controls and pushing open standards – certainly don’t have much to do with building systems.
There are many people who desperately want government IT to be better. But there has also been a perception that to criticise GDS somehow equates to being on the side of the outsourcers whom it has rightly been railing against.
One source who is close to GDS and sympathetic toward the intention of changing government IT remarked: "They had the right idea, at the right time, but there's been poor execution and a lack of humility in their approach."
If the next incarnation of GDS leadership starts 2016 with nothing more than the aim of under-promising and over-delivering, that could be a good sign for the next four years. ®
Sponsored: Customer Identity and Access Management