Digital Government in the next Parliament

Yesterday I gave a speech to the High Potential Development Scheme cohort of Civil Service Directors and Directors General. I’ve blogged it here in case it is of wider interest.


What does digital mean for you?

Getting a decent laptop for work?

Maybe getting Google alerts of all the relevant news on your brief, straight to your phone?

Maybe a transformation programme for better digital services to meet the needs of our users.

Perhaps, it’s using Twitter to improve your engagement, and strengthen your influence.

Maybe just the wistful memory of a familiar departmental website that you knew your way around, and had some sense of control over?

Whatever digital means, I know that you’ll want to keep it in perspective. My most important asset is not my shiny new laptop, or my knowledge of data science, but time and attention – mine, and that of those around me. The internet in my pocket is useful, but also a liability: an endless source of distraction. It makes time move faster, and disappear more quickly.

I’m doing digital delivery now after a spell at Google, but really I’m a recovering policy wonk who spent 10 years in the Prime Minister’s Strategy Unit, the Treasury, and the Department of Energy and Climate Change working in a way that most of you are familiar with.

Nevertheless, I have drunk the Kool Aid, and I do I think digital will transform the way in which policy is made; the way in which ministers engage; and the way in which Government is done.

In this parliament, digital has largely been in the wings. Noises off. I will argue that it is about to come centre stage. In the next parliament, it will inescapably affect how we play our parts.

Enormous challenges loom over us:

  • Welfare Reform – pensions, troubled families, and social care
  • Britain in the World – from Europe, to terrorism, to Russia & Ukraine
  • Our long term security – our climate, our energy security
  • Our political Union – whether devolution, or disintegration
  • Funding and financing – the costs of the NHS to banking regulation
  • Public trust and engagement in politics and the political process
  • Inequality – wage stagnation, migration, productivity, the squeezed middle
  • Our civil service – competition for talent, capability, the risk of another Snowden

Alongside these challenges we face a fiscal backdrop just as sombre as in the last parliament, with cuts to the budget and staffing of Whitehall set to continue as the new normal.

Digital has potential

From the sidelines, I believe that the Government Digital Service has demonstrated what digital transformation looks like in the context of public service delivery:

  • government wide platforms like GOV.UK, the Performance Platform, and Verify, a new platform to validate users identity when they engage with government. These are high-quality common components for the whole of government to rely on, and they will continue to evolve and improve;
  • rapidly transformed departmental Digital Services like Register to Vote, Lasting Power of Attorney, or Carers Allowance that give citizens and businesses a simpler, clearer faster service, and which have seen delighted users. They’ve even had to install a new button in the Ministry of Justice call centre to register the positive feedback they’ve been getting – they had never had any before.
  • Digital has enabled savings of over £1bn since 2011, according to HM Treasury.

In the three years since GDS was founded, New Zealand and Hawaii have both taken the source code for GOV.UK to run for themselves; we’ve been described as the best startup in Europe by Saul Klein, a leading UK Tech Venture Capitalist, and the ‘Gold standard of global digital government’ by the Washington Post. We beat the Shard and the Olympic Cauldron in a design competition, and the United States government has just created a Digital Service modelled on us.

I’m not here to brag – we’re far from perfect, and it is tough to scale an organisation from a dozen to 500+ people without a few growing pains. I make the point only because it is interesting to contrast the external news of GDS, with the less positive stories you sometimes hear inside Whitehall.

In any case, there is so much more to do:

  • a more concerted effort to replace Government’s worn out tech – like the laptops which take 10 minutes to start, or the email attachments with twelve layers of tracked changes – with the quality of software and hardware that we’re becoming used to using outside of work (and by freeing ourselves from eye-wateringly expensive supplier contracts we would slash the £8bn annual spend on IT in the process);
  • we need systematic use of performance information and data science techniques to understand how we’re doing in real-time, and get better evidence for our operational and policy decisions; and
  • we need more shared cross-government components for service delivery, like licensing, fraud detection, making payments, identity matching, and case management software.

That is what I believe the future of digital transformation looks like in the context of public service delivery.

But what about in the more traditional and strategically central context of policy making?

The we way we’re making policy isn’t working

I think there are five big problems with the classical policy making process.

First, it takes too long.In 2005 I came into the Treasury to work on Productivity. I made the case for an independent review of Intellectual Property policy, which was commissioned, and completed by the end of 2006. Since then there has been a gradual implementation, with thousands of pages of consultation, explanation, and legislation. One of the 54 recommendations made 10 years ago, to legalise copying CDs to your computer, finally comes into effect on the first of October this year. Do people even buy CDs any more?

Second, it has become too rigid in details. The cycle of green paper, then Whitepaper, then draft bill, and a slew of secondary legislation – it quickly leaves the realm of principles, and becomes service design. Fixing operational service design in legislation is a disaster because the best way of meeting the needs of users are hard for anyone to predict from first principles, and they are likely to change over time. But we’re stuck with a statute book that demands wet signatures, which has been a problem for services like Universal Credit and Lasting Power of Attorney, and a paper tax disc for car registration. On top of that, because it is difficult to change, new policy often gets layered on to existing policy, breeding complexity and confusion. Just look at the sedimentary layers in the design of our Energy Market, Tax Code, or policies like Carers Allowance.

Third, it learns too slowly from experience. Aside from professional lobbying groups, people are generally not good at giving feedback on abstract propositions. We are much better at giving feedback on our experience of a service. And as civil servants, I’d argue that we’re also much better at responding to this kind of feedback. Testing a small prototype service with real users gives invaluable insight, and also forces us to think about the service interface early rather than at the last minute. Services like the Green Deal that launch with a ‘big bang’ run enormous implementation risks.

Fourth, fear of technology removes a vital design tool. The benefits of losing the paper counterpart to driving licence have been manifest for a decade, but fear of the IT change required put policy colleagues off. We need a better education as to what is hard, and what isn’t.

Lastly, policy possibilities are limited by a technical archipelago. There has long been a policy desire to be able to set up a business with ‘one click’? Policy colleagues across departments are able to agree that this would be a boon for business. But technically, this is genuinely hard because these systems stationed in different departments and agencies haven’t been designed to talk to one another. Have you heard of ‘Tell us Once’? The ambition is laudable – but this is a policy pipedream until we interconnect our systems properly.

Where big data systems have grown up in one department as a national asset – like insurance numbers, cars, or patients, legally and technically they tend to serve the needs of that one department first, with others struggling to get access. I saw this recently in the Illegal Working review, where the Home Office was requesting access to national insurance numbers from HMRC and DWP, to creating a service for employers to check on the status of potential employees.

Digital will be needed to solve the upcoming big policy challenges

The way we solve problems needs to change.

Classically, I’ve tended to understand policy as flowing from strategy. But I think I’ve got things upside down. I’ve begun to think that policy should flow bottom up, rather than top-down: from a service, and engagement with users, not from a strategy.

Digital services are quick to build, easy to iterate, a constant source of rich & objective feedback. They also open new opportunities for policy that may be able to break open the stale debates that have stymied solutions to problems as diverse as road charging, energy efficiency, media regulation, and child protection.

My default mode of operation, was to take a problem and analytically describe a solution, with as many supporting stakeholders, case studies, and cross-national parallels as I could muster. Our bosses, ministers, and the Centre prod and poke it, and if it stands to reason, is affordable, and has politics on its side, then it has a good chance of being accepted and announced.

We don’t know if it will work, and it usually takes a couple of years between announcement and implementation to find out.

But the public policy problems we have now are too difficult to solve once. We need adaptive solutions that can be experimentally applied, extended or cancelled, without loss of career, or years of paperwork.

The tools with which we solve problems may need to change as well. As a policy analyst, I was taught that there were always basically three options. Regulation. Taxation. Spending.

Service delivery came nowhere.

What if the operational design of a service was our first concern rather than our last?

Lets build something, and then test our assumptions for real. Lobby for ministerial announcements that describe delivery, rather than just intent.

The sums of money involved are relatively small to build prototypes and alphas – often smaller than the amount we would spend on policy development. The Treasury has even adapted their business case methods in the Green Book to enable this to happen.

We need more Engineers, not more lawyers. We have a dearth of experimentation, not a shortage of rational process.

Digital is, of course, not the only answer – the themes of design thinking, open policy making, What Works centres, pilots and RCTs, behavioural insights and cognitive biases, all of them link back to experimentation, and digital is a golden thread that links them together.

It is now just a few weeks after the latest government IT disaster – the Home Office’s eBorders programme. So I understand if it feels premature to sound the trumpets for Digital. But as someone once said, the Future is already here, its just unevenly distributed.

The mechanics of Government in other countries are becoming Digital

This is not a story that is native to Britain. Other countries are fellow-travellers.

Estonia – started from scratch in the aftermath of their independence. They have passed a law allowing any citizen to refuse to provide government with a piece of information if they have already provided it. This requires radical interconnection between their agencies and departments. They don’t need many people to run their digital services. They said that seven people run their benefits service.

China, Mexico, Singapore, Finland – all these countries are taking a fresh approach to public policy, by bringing digital delivery to the centre, and breaking the old habits of government.

After the trauma of, the US Government is also paying much more attention to service delivery.

You will be central to this – what do you need to know to succeed?

There are a bunch of things might help you understand digital skills in more detail that I will circulate later today.

But as well as skills, I would emphasise Culture. Our design principles are a straightforward list, but to to make them happen requires a big change in culture.

As senior leaders, this is particularly a job for you. You’re the guardians of culture for the civil service, and you’ll have a big say in what the future looks like.

Change doesn’t come easily. In the 16th Century, Japanese Samurai didn’t readily embrace guns or other firearms, because they distrusted the new technology, and because the sword was deeply embedded as a part of their art, culture, and honour. They went to extraordinary lengths to avoid using guns, but eventually they did, because they found that without it they were losing.

I would pick three of those design principles as particular cultural challenges:

Make things Open – despite Snowden, despite civil service salary transparency – making things open makes them better.

Do less – in particular, write less: on the tour, did you pass the publication ticker for government. It is astonishing to watch the volume and variety of words the government publishes.

Understand context – get know your users – visit the user testing lab downstairs, or sit with your departmental digital team doing some guerilla user testing.

And without wanting to sounds like a counsellor, you may want to have a conversation with someone who can advise you on the particular challenges you’re facing, and the ideas you’ve got for developing some of these themes. Perhaps the digital leader in your department, or the Chief Digital Officer, if there is one. If you’re stuck, then come back to us and we’ll be able to help or put you in touch with someone who can.

And now, after so many words, surely it’s time for us to go and build something.