Government Digital Services (GDS) was established in April 2011 to implement a 'Digital by Default' strategy for the UK Government. It is the ‘expert centre for the delivery of digital, technology and data services’. GDS collaborates with departments across central government to build platforms, define standards, and operate digital services.
GDS is widely considered to be one of the leading public sector digital transformation agencies in the world, its approaches being adopted as a template by others. The UK was ranked at the top of the United Nations E-Government Survey 2016, up from its tenth position in 2008.
Stephen Foreshew-Cain, Former Executive Director (September 2015 to August 2016) and Chief Operating Officer (April 2014 – September 2015) at GDS UK discusses the milestones from his time at GDS. He tells us about the evolution of strategy and its implementation and building services in such a way users would opt in.
Developing and implementing Government-as-a-Platform (GaaP)
Inception of the idea
Government in the UK, like many governments, is a very siloed organisation with departmental services, defined more often by what differentiates them than by what is common. The tax department is separate from the benefits and pensions department, which is separate from the environment department. There is a lot of duplication and replication across government and new services are coming online all the time.
Mr. Foreshew-Cain said, “We recognised early on that if we try and transform each individual service one by one, we would never be done. So, in 2015 we re-focused our strategy on building common platforms, build them to a very high standard and make them available to all parts of government to consume. These platforms provide important services but they are not necessarily big or complicated. For example, we would have a payment platform that would be available for the whole of government to use. That was a transitional pivot point for GDS.”
Understanding needs and building the case
"One of my proudest moments at GDS was finding support across government to buy into the platform strategy as a way of accelerating digital transformation. For departments to support it and, most importantly, to see HM Treasury fund it in the 2015 Spending Review and Autumn Statement, was a major step forward for digital government in the UK," Mr. Foreshew-Cain said while discussing the process of getting the platform strategy off the ground.
Early on in GDS’ evolution a leadership forum was established, where digital, technology and data leaders from departments come together and share what it is that they needed.
This wasn’t GDS at the centre alone deciding to build a notifications platform (GOV.UK Notify) or payment platform (GOV.UK Pay). The departments were telling GDS what services they operate, what services they intend to build in the future and working together to find the common components. GDS explored if they could be made available centrally at a lower cost. The mantra was ‘Build Once, Build Well, Use by All’.
The initial list included around 40 or so potential platforms. It included payment, notification, identity verification, publishing content (some of which GDS has already built) and potential platforms such as a platform for managing biometric data, which is going to be a growing need in the future. Licensing was another very common component.
Describing duplication and redundancy, Mr. Foreshew-Cain said, “With something like payment, there were very few variations around how that could or should be done. And yet government approached it as a unique problem each time a service needed to take payment. And really, if you are doing the same thing more than once, there has to be a very good reason for that.”
The entire government-as-a-platform business case is predicated on delivering savings by moving to shared platforms. Not necessarily from individual transaction costs (although there were savings to be delivered there too) but the avoidance of duplication and building replicated systems and then having to operate, change and maintain them.
If a provider makes a change to their backend system, you might need to dedicate staff up to enable your services to adjust to those changes. If you go to a centralised platform, that is accessed on an trust and consent basis (via open APIs) it should be much easier to make changes. The cost to department is reduced to onboarding to the platform as opposed to cost of designing, developing and operating common technologies.
However, there were challenges and concerns raised relating to the platform approach which we worked through as we sought HM Treasury approval.
“There is no value for me now”: Some of the large departments said we don’t need these platforms. We have funding, we already have integrated our systems with payment providers. You are not saving us any money. That’s true perhaps in the early evolution of these platforms. Also, large departments have agreements with platform providers, which are cost prohibitive to move off. They get locked into one platform provider. So, there might not have been immediate value for them but there was a long argument for the whole of government over the long-term value. This is at its heart the siloed departmental view versus a connected cross-government view.
Who’s setting the agenda: There was a natural concern expressed by departments as to who is going to set the agenda for the platforms, as we go forward. It is important that departments deliver and operate digital services, not the centre. And at times, it may be that a specific departmental need is not as important as the collective need of the other departments. But that doesn’t make it any less important for the department. There was a lot of discussion and negotiation to come to a consensus about which platforms we build and in what order, and how to accommodate those departmental specific requirements which a platform may not immediately support.
Setting spend controls
In addition to working closely with departments, GDS continues to administer the Technology Spend Controls and apply the various service standards that had been established from the very beginning. GDS, as the expert centre, has a responsibility that government is spending public money well. If a department wants to spend significant money on a developing a digital service or implementing technical infrastructure, they continue to seek approval from GDS to do so.
GDS is also the owner and administrator of the digital service standard, which is the shared definition of what “good” looks like when it comes to government digital services. No service could be made available to the public without passing the service standard. The service standard defines principles which inform practice. Whilst there is no explicit platform standard, the service standard continues to evolve to ensure that GDS are building platforms that meet user needs and not for government convenience.
Development and Adoption
GOV.UK Pay and GOV.UK Notify were delivered in a relatively short period. Those two platforms were just getting started a year ago. They are now in Live Beta and are being used by live services across government. An identity assurance platform for the individual, called GOV.UK Verify, was brought to a live status.
GOV.UK Pay and GOV.UK Notify are both in Beta. They are in a live public test, which means that they are being used by services which helps provide feedback for building better platforms. (According to the GOV.UK Pay website, Pay is currently being tested in partnership with Companies House, the Environment Agency, the Home Office and the Ministry of Justice).
GOV.UK Notify is a platform which provides for communicating status to users. It covers SMS, email as well as physical post. Now it has savings targets associated with the platform and it has a strong and healthy pipeline of government services coming online.
Mr. Foreshew-Cain talked about the opt-in from the users, “We didn’t mandate use of the platforms. The entire take-up is people opting in. I think that’s the big success of the platforms.”
Delivery teams on the ground recognised that having to build these things for themselves is difficult and risky. GDS took away some of that risk by having a proven, implemented platform service that they can simply integrate with. They don’t have to build it or buy it anymore. There will only be an knowable onboarding cost for coming to the platform, and the more services they bring on, the lower that onboarding cost as it shared across government.
A key measure is how many people can interact with government simply, easily and fast. With digital services, you can easily see how many people go from point A to point B. Who’s dropping out, why are they dropping out.
During the first three years of GDS, GDS saved around 3.56 billion Pounds for the government.
“Building simpler, better, faster services for users also delivers significant cost-reduction in servicing that demand”, Mr. Foreshew-Cain said.
Last year in the 2015 Spending Review and Autumn Statement, GDS was allocated 450 million pounds to continue the work of transforming government. Mr. Foreshew-Cain saw it as an affirmation of the GaaP approach.
When things go wrong
Mr. Foreshew-Cain said that extensive testing, with alternatives available, is the way GDS deals with this.
"That’s why we have alphas, that’s why we have prototypes. When we are in Beta, we always have a contingency . If something goes wrong, there is somewhere else to go.
“The electoral register went down 2 hours before the close of registration for the Brexit referendum. That was highly visible and very painful. There was an unexpected and truly unprecedented surge of demand. But when something like that happens, we are open and transparent about what went wrong and what we are going to do to fix it.
We work with an assumption that something at some point is going to go wrong. We prepare our ministers, senior civil service leaders and our teams to view failure as learning and ensure we don’t fail when it matters – when users are relying on us for critical services. It can be managed, as long as that is not the only option, and we have other ways of servicing demand”, Mr. Foreshew-Cain said.
This is a part of the culture of GDS, and it’s in the design principles – make things open, it makes things better. Historically, we might expect that government would be slow to react. Recently during a recent DDoS attack on Dyn (an organisations which provides DNS services for many large websites) GOV.UK was unavailable.
GDS restored the service by removing Dyn from the infrastructure for 'www.gov.uk' and moving to an alternate provider. GDS one of the first to switch. But GDS was able to act quickly, because they were prepared for it. GDS blog and publish whenever there are outages on GOV.UK. That engenders public trust; that when things go wrong the team behind GOV.UK are on it.
Attracting the right talent
During last year, GDS took a serious look at the approach to attract, develop, and retain digital talent within the Civil Service.
GDS had done some work in our first few years in advising departments and agencies on the skills and capabilities they needed to develop. That was stepped up to talk about the professional structures and career models required to make government a desirable destination for people with the highly-competed digital skills.
The first step was understanding the capabilities needed. A lot of these people had been hired, but the structures weren’t there within government for them to continue to develop and grow and thrive in government.
Most governments have limited funding to direct to digital skills. In addition, there are restrictive covenants around civil service and public sector pay.
Mr. Foreshew-Cain contrasted the generalist vs specialist models, “A lot of the conventional wisdom regarding civil service careers was targeted at what we would call a generalist, somebody who gathers a general and broad toolkit of skills to be applied in different ways across government. A generalist could assume an administrative role at one time, might do some policy development at another. In the digital, data and technology professions, people develop very specific skills. The model for reward that we had for general civil service just doesn’t incentivise that, or respond to a competitive market.”
At the point Mr. Foreshew-Cain left government, GDS had started a dialogue within government on developing a grading and reward structure.
Private sector vendors must focus on Government’s needs
During last year, GDS defined supplier standards. Principles were set out by suppliers in the private sector would be expected to work with government. This was done in consultation with the private sector suppliers on how this would work.
Mr. Foreshew-Cain explained the primary change, “Private sector suppliers to government need to value the things that government values. GDS pivoted government to consider user needs first, not government needs. A similar transition needs to happen in the relationship between government and its private sector suppliers. We expect you as a supplier to put government’s needs first, in the same way that we in government have learnt to put citizens’ needs first.”
There were other things that materially shifted the relationship. For example, considering data to be a public asset and not something that the supplier gets to own. The services should be built on open standards. And if government uses components a certain vendor, they should have the opportunity to reuse them across government, if they see a need to do so.
And most importantly that vendors engage with GDS in being transparent, moving to transparent contracting and sharing of learning.
Defining Digital Transformation – Are our institutions designed for the 21st century?
We asked Mr. Foreshew-Cain what digital transformation means to him. In his words:
“Digital is about applying the culture, processes, practices and technology of the internet era to respond to people’s raised expectations. But we will get it wrong if people focus on the digital part of it and not the ‘transformation’ part.
Most importantly this is not about technology. We are talking about institutional reform. Is Uber just a taxi company with better tech? No – it’s a whole new business model that works fundamentally differently. The institutions that we have in the UK were designed around an administrative bureaucracy in the mid-19th century, not the networked 21st centrury.
For me digital transformation starts with a question. Are our institutions designed to operate in the 21st century? Taking a hypothetical, if I were to set the Driver and Vehicle Licensing Agency (DVLA) today, would it look anything like the administrative bureaucracy that we have evolved to today? The answer, again, is no. It would most probably have more in common with the platform-focused digital start-ups.
In the private sector the cycle time between strategy and implementation has been reduced dramatically. Government’s not there yet. We still have policy professionals that exist in a separate world, thinking about what are the choices we need to make to implement the political agenda of the government of the day.
Policy design should not be a big upfront intellectual exercise. Policy needs to adopt the design practices of the digital economy. Build something, test and learn over short cycle times. Test policy because the cost of change now is so low. That will drive a different shape of government.”