A panoramic photograph of Malham Cove in North Yorkshire.

Politics, trust, and GDS.

Tom Forth,

I have a love/hate relationship with the UK’s Government Digital Services team, GDS. I love some of the improvements that they’re making to central government services, I love the way that their methods and guidance are clear, open, and consistent across different departments, and I love that they’ve been able to drag many of the UK civil service’s dinosaurs into the present.

But I dislike their obsession with user-centred design, their tendency to do so much in London, and their willingness to lecture others with a tiny fraction of their budget on the best way to do things. And let’s not even talk about data.gov.uk

I’ve written before about how the GDSification of paying taxes for my business shows that GDS is going in the right direction. They’re at the Windows 8 stage of Microsoft’s transition from Windows 7 to Windows 10. They’ve made great progress, but they’re not there yet.

Here I want to talk about online voter registration and the challenges for GDS that it made obvious.

Before I start, I want to sneak in a quick disclaimer. Some of the details in this piece are probably wrong. I’ve spoken to people, but I’d like to speak to more. Leave a comment and I’ll correct things.

The history of online voter registration

In 2014, voter registration in Great Britain switched from being based around households to being based around individuals. I’m not convinced that that was a good decision, but it happened.

So GDS stepped up and built a fantastic tool for individuals to register to vote. It works on mobile, it’s accessible, it’s fast, it’s simple. Importantly there’s now just one website and process for the whole country, instead of hundreds of different methods for each local authority. The tool had a few problems early on, but they were quickly fixed.

So far so good.

But I’m here to tell you why the individual voter registration tool is a failure of user-centred design and centralised thinking. Buckle up and prepare your rage-replies.

The foundation of democratic trust is local

Let me introduce Jo Miller, the CEO of Doncaster City Council. I’ve never met her, I only know her through twitter, and I suspect that she votes Labour. But the word suspect is crucial here. I have no idea who Jo votes for, and I doubt I ever will. Her discretion is a big part of why she and the many other council Chief Execs and returning officers in the UK deserve the large salaries that they earn. How much would you need paying to dedicate your whole life to public service and then give up your opinion on politics? Jo and others do that.

But Jo, and others like Joanne Roney in Manchester, Kersten England in Bradford, and Tom Riordan in Leeds, matter to elections in an even more important way. They organise and take responsibility for the teams that unlock polling stations, sharpen pencils, check registers, and count votes. My faith in our democracy is only as strong as my faith in Jo and her team. It's the faith that if the vote was tied, at 6am, two days since they last slept properly, and they saw a ballot paper with a cross next to a party they despised that had slipped down a vent in a freezing cold sports hall in Doncaster, that they’d spend half an hour fishing it out to count it.

I wouldn’t trust a politician, or a lawyer, or a doctor, or a teacher, or a nurse to do that. Not for sure. I certainly wouldn’t trust a national civil servant. But having met a few of Jo’s type, I’d trust them and their teams for sure. And survey data suggests that my trust in local over national on these issues is well-founded.

This faith is why it’s right that electoral rolls are still held and maintained locally. And it’s why further centralisation of voter registration is not the right way to fix the mess that we’re in.

Centralising money and control while decentralising costs and responsibility

You can probably see the problem by now. Individual voter registration took both money and control from local authorities and centralised it in London. We got a great new tool at the centre that fails the most basic test, it cannot tell you if you’re already registered. Electoral rolls are maintained by local authorities who receive frequent summaries of registration from the national tool. They then manually check whether people are already registered, and add those that are not.

This is what happens.

A change in registration system, and the recent memory of having to re-register even if you hadn’t moved home, means that no-one is sure whether they’re registered or not. I suspect that many young Britons think that they need to register for each election. And so every election we urge them to register again, just to make sure, and our fancy new national tool can’t tell them that they’re already registered. And all of this wastes local money, as well as citizens’ patience to engage, at a time when both are in short supply.

Okay smart-arse, what should we have done?

There was probably no way around this failure once politicians had pushed through a new voting system and demanded that the civil service make it work. Cuts to local government have been so severe in the past five years that capabilities and budgets to work with the centre as equals just don’t exist. And GDS isn’t set up to work across the whole of the UK. It was designed to drag central government departments in London (or that could be summoned to London) into modernity, not to empower local government to deliver better digital services.

We need a new system. It’s one that I could build at ODILeeds if someone wants me to.

First up, we need a decent chunk of GDS’ budget for at least two years, considerable freedom to spend it how we like, and a list of problems to solve in local government. I’d base my team at ODILeeds and start with five challenges,

  1. Write an open source voter registration system for local government.
  2. Commission MySociety to update and improve their FixMyStreet app and use it as the basis for a wider standard local government app.
  3. Write an open source toolkit of common local government processes, such as collecting business rates and council tax, and applying for a school place.
  4. Work with GDS’ Verify team to incorporate individual verification into local government services. There are already attempts to do this, but I want to be clear; GDS on a train from London working with local government will not be as good as us in Leeds working with local government. Structure matters.
  5. Use the experience of implementing all of the above to define and publish common data standards for local government.

Crucially, for each service we build we’d buy local authority involvement with cold hard transparent cash, not experts on a train. Rewards for joining in, not the current patronising mix of hand-holding and fines.

We’ve already built a lot of the structures that we’d need to run this thing – we have a large space in central Leeds, fantastic relationships across the UK, a brilliant network to hire talent and buy solutions from, a brand that demands open solutions and promotes open data, and a proven record of delivering excellent products on very low budgets.

We’ll talk less about user-centred design (though we’ll still do it), and we’ll be much less visible in Whitehall, London’s tech scene, and the national media than you’re used to. And if a Minister or a senior civil servant wants to check how we’re doing, they’ll have to come to us, not the other way round.

But we’ll get the job done. If you want.

blog comments powered by Disqus