GDS' government-as-a-platform ambitions are misunderstood – it's not about control

The summer period is dubbed 'silly season' in the world of business journalism because everyone is supposed to be on holiday and nothing is really meant to happen for a few months. A lovely respite, supposedly.

Not if you're covering the Government Digital Service (GDS) in the UK. The past two weeks have been filled with a number of shock announcements, which were triggered by director Mike Bracken announcing that he will be stepping down from his position at the end of September.

This raised a number questions about the future of the digital organisation at the centre of Whitehall, resulted in a number of other senior figures announcing their departure and led to diginomica declaring that it has no confidence in Civil Service CEO John Manzoni to undertake the institutional reform required to carry out true digital transformation.

However, the biggest worry concerns the future of the government-as-a-platform (GaaP) agenda. It seems there is not the required backing from those at the top of the Civil Service, as well as the political ladder, to try GaaP.

We've described GaaP previously as:

GaaP aims to introduce common systems for re-use across government, which will inevitably cut down on costs, introduce efficiencies and ‘standardise’ public services.

Part of the problem is that those operating in Whitehall don't have the motivation to change the system that they currently operate in, as it suits too many people and businesses to keep it that way. However, with estimates that GaaP could not only save billions of pounds every year for Whitehall but also improve every citizen's interactions with government, it's worth keeping the conversation around its benefits alive.

But part of the problem, I believe, is that people don't fully understand the end-game for what GDS was trying to do with GaaP. Over the past couple of weeks, I've heard comments, from some journalists, analysts, and stakeholders that GDS lost its way. Collectively, they think it either got too big, wants to control everything or wants to build everything itself.

These comments come from people that I respect and are a lot smarter than I am, but I think it's worth having on the record that I fundamentally disagree with this portrayal of GDS. That's not to say that I think GDS is perfect. We have criticised them for areas where we believe they have gone wrong. However, I think GDS' long-term vision around GaaP is an interesting proposition, and one that not many people fully understand.

Commoditisation is the name of the game

One of the biggest arguments against GDS at the moment is that it wants to build everything and is becoming a software shop at the centre of Whitehall. The problem with this, if true, is that it will handle a hell of a lot of code. Can GDS build and operate all the commoditised building blocks for GaaP?

No. It cannot. But that's not what it is doing.

From my understanding GaaP has three distinct features – 1) data; 2) a marketplace that relies on open source code and; 3) competition. These are three separate features, but all link to the success of GaaP.

Data, for example, is currently a mess in the public sector. Everyone has their system of record and citizens have to tell each department the same information time and time again.

The plan was for the government to create a variety of canonical sources of data and registers, which could be used across the public sector. For example, a register for addresses, for doctors, for companies, for car ownership, for land, etc. These would not be stored in a database, they'd be hashed and use blockhain-style encryption to ensure security. This effectively means that the data is as secure as it can be, given that each change is logged.

Simply put, there would be a data registry for all the core types of citizen and/or public sector information that will be required.

From the conversations I've had, the plan was to put control of this data in the hands of the citizen. So, for example, if the Department for Work & Pensions wants access to information about a welfare payment, access permission is given with which to retrieve that data from a number of registers. Just that one time. In that context, data sharing becomes irrelevant.

The second and perhaps more interesting point relates to GDS creating an open source marketplace for commoditised services. When people say GDS wants to build everything, they're mistaken. Consider Verify, one of the core components of GDS' vision for GaaP. Has it built any identity authentication systems?

No, it hasn't. It has built an open-standards based wrap around for a number of payment platforms already in the market, which buyers (departments) can then plug into.

What I don't think many people have understood is that the open standards nature of the wrap around, allows GDS to commoditize those services. There is no lock in and it has the buying power of government at it's fingertips. Want to charge a premium for a payment mechanism that should be a commodity? Good luck when you've got a number of other players using the same standards and when GDS can run a reverse auction for that service.

Consider how this could work for hosting. A department wants to build a service for some new policy it's testing out. It could go to GDS and ask for hosting, a payments platform, a licencing platform, etc. Would it care where these came from if it was secure and commoditized? I doubt it. Using those standard building blocks I referred to above, the department could then tap into the data registers for the information it needs, but only pull that information when authorised by the citizen.

Finally, the whole point of a platforms-based organisation is that control isn't kept at the centre. There has to be competition across the organisation. GDS could become the doorway to a marketplace, but there's no reason that it couldn't be a supporter or founder of the open source community that IS the marketplace.

From there on, departments thinking about new policies and a different way of approaching government interaction with citizens wouldn't need to invest everything in one delivery model. They could have three or four competing services for the same policy, where the best one wins out, using the building blocks available to those services. Competition is important. Universal Credit would have been better first time around if approached this way.

This means control isn't with GDS. It's a just a gateway to those building blocks. The digital capability HAS to lie with the departments. From my understanding, based upon numerous conversations with officials, GDS can't and doesn't want to build it all.

For an example of what GaaP looks like in the real world, take a look at this presentation posted by outgoing deputy-director at GDS, Tom Loosemore.

My take

It's a vision that hasn't been tried anywhere else in the world. My hope is that it is one the current teams at GDS pursue. If we really want to save money and deliver better services to people, this is a real opportunity.