The Government Digital Service's (GDS) current monopoly position on providing Whitehall IT places it at high risk of repeating the same costly IT disasters of the past.
"With a monopoly position and a client base compelled to turn to GDS for advice, there is a risk that they could become an inefficient organisation removed from the efficiency drivers of the market," said a report by consultancy firm BDO.
The report [PDF] added the body's role in government "blurs the accountability of the responsible department", leading to risk. It also questioned the need for having 425 civil servants and 210 interim staff.
In an accompanying blog post, Jack Perschke, director of central government at BDO, said: "Can a monolithic GDS of 600 people, with all the vested interests and perverse incentives that brings, really sit at the heart of a digital revolution?"
He suggested that instead of trying to build systems to process data, GDS should be using its energy to allow government entities to buy systems.
Eddie Copeland, head of technology at centre-right think tank Policy Exchange, said there was a contradiction in the GDS "build not buy" approach:
[They intend to] create commodity components that can be used repeatedly and reassembled locally, but they’re building it with bespoke code. It means we’re left with dozens of components that do not exist anywhere else in the market, and which the government will forever be responsible for updating and maintaining.
Copeland added: "I fear people are also being too dogmatic about using open source. Open source has many benefits, but we should also be prepared to work with good proprietary systems as long as they have open APIs. It’s open standards in data that count."
He said the government currently has 770 transactions, and the task of digitising those is huge. So far, GDS claims to have just made 15 digital services live – the vast majority of which lack any real transactional element.
One government insider told us that GDS has become an in-house big consultancy firm itself, building everything bespoke, making it "special", unjoined up and delivering late.
Interestingly, GDS appears to have removed details from its service design manual regarding a buy approach.
The original document about "government as a platform" is still available on Github. It says: "A move to platforms does not mean that government has to develop everything in-house: many of the government's needs can be met by existing cost-efficient utility services."
No mention of a buy approach remains on the "government as a platform" page. This suggests GDS is currently intent on sticking to a "build everything in-house" approach.
Due to the upcoming general election in the UK, the civil service is in purdah, meaning it cannot comment at this time. ®