
Last year I was interviewed by the mercurial Kai Brach for his magazine Offscreen, which seeks to highlight the human side of technology, and does so in style. You should subscribe. Quite understandably, Kai doesn’t normally allow the contents of Offscreen to be republished online, but wanted our conversation to get a wider audience, so granted me permission as a one-off. Thank you, Kai.
[Offscreen] Back in the ’90s you were studying towards an engineering degree when a cover of Wired magazine changed your career plans pretty dramatically. What was so impactful about that cover?
It was 1994 to be exact. The cover of the European edition of Wired prominently featured the quote, ‘We have it in our power to begin the world again. – Thomas Paine, Digital Revolutionary.’ That cover stopped me in my tracks.
By then I had been messing around with the internet for a few years. It was pure magic. In my gut I knew this internet thing was going to change everything, but I didn’t dare to share my excitement with anyone else because it felt sort of ludicrous. I mean, at the time, it was pretty geeky, clunky computer stuff, mostly running off a command line.
Seeing that cover crystallised what I had been feeling for a while. The ability to copy and share at zero marginal cost with zero friction globally was an incredibly powerful new concept. It meant that as a society we had it in our power to reinvent a lot of things – for good or for ill. I remember feeling a sense of responsibility, even then. That’s when I started putting my thoughts into writing and ended up publishing articles about the web in various newspapers in the UK.
After graduating I scored a job at Ove Arup, a respected engineering company, though I didn’t last very long there. In the first week I got wind of an internship opening up at Wired and there was nothing that could stop me. A month later I had a full-time job at Wired as ‘Internet Editor’.
[Offscreen] What did you take away from working at Wired?
Getting a glimpse into the world of the pre-dot-com bubble was incredible. I was lucky to work with and learn from amazingly creative people. I also got to know the early internet community that was forming in London and San Francisco. Most importantly, it taught me to see through the hype and realise that the digital revolution was initially going to be one driven by utility – by making people’s lives easier. That was reflected in my writing, in which I often made the case for putting train timetables on the web, allowing people to bank online, and so on.
Eventually, the UK edition of Wired shut down and I was out of a job. In retrospect, this was a good thing because I realised that I didn’t want to be a journalist. It was time for me to stop writing about web stuff and start doing web stuff. I got a job at Capital Radio, the UK’s biggest commercial radio group, where I helped roll out their first website. I then moved on to work for the BBC and lead the creation of their website for the 1998 FIFA World Cup – my first large-scale product development project. It really opened my eyes to what a small team with tight deadlines and complementary skills is capable of.
After doing my own thing for a while – including launching UpMyStreet.com, a local information website – I went back to the BBC as a senior manager in 2001. With an incredibly talented team, we pushed many ideas that were ahead of their time – digital identity, social networking, instant messaging, online streaming. Sadly, few of them saw the light of day. The headwind coming from the BBC bureaucracy was strong. They were just too scared to really go all-in.
[Offscreen] Is that where you first learned how to work around the bureaucracy of large organisations?
It was at the BBC that I learned how to hack corporate politics. First, though, I tried to play the game and do things through the right channels. I tried to not upset anybody and get consent every step of the way. But none of the interesting projects got past the scrutiny panels and the editorial policies.
At one point I was invited to spend a week at the Stanford Research Institute in California and that changed my view completely. It taught me the power of fast iteration with small teams, the notion that you’ll be wrong many times before you’re right, the idea of optimising for feedback loops, how to pitch ideas up the chain of command, and so on.
When I returned to the BBC I was determined to fly under the radar. If I just had enough top cover I could invent the future at the BBC with a small team of just three or four people. We acted like a startup. We just did it our way and managed to get quite a few interesting projects out the door, including a developer network and a precursor to what is now iPlayer, BBC’s online streaming app.
The important lesson I learned was that if you do something good, something you believe in, it will generate its own positive momentum regardless of whether you have permission or not. By showing a better way and backing it up with results, you can achieve a lot with a little. And it’s almost impossible to get fired for doing good work.
[Offscreen] When did you decide to become more politically involved?
That actually started a bit earlier. In 1998, a group of friends and I ran an online campaign to try to persuade the UK government not to adopt a range of very stupid encryption policies. We created a service called ‘Fax Your MP’ to turn up the heat. At the time, MPs (Members of Parliament) didn’t use email; not even their fax numbers were available online. Luckily, through a personal connection, I got hold of a parliamentary booklet that contained a list of MPs and their fax numbers. We built a web-to-fax gateway that allowed people to send faxes to their MPs very easily. I remember running around London with a fax modem under my arm, moving our base closer to Parliament in order to qualify for free local calls, which was crucial given that we were sending over 100,000 faxes in a short period of time.
In the process we realised that most people didn’t even know who their MPs were, much less what their opinion on specific issues was or how they voted on them. So in our spare time we built a system that would scrape the Parliamentary record to compile a plain-English profile page for each MP, making everything that person said and voted on easily accessible to the public. TheyWorkForYou.com instantly shot to number one on Google for MP-related searches and went on to become quite influential, showing Parliament how to operate more transparently through digital media. The project was a forerunner of a charity called mySociety which now provides freedom of information and other digital services to parliaments and governments around the world.
[Offscreen] In 2010 you got the chance to make a much bigger splash and help start GDS (Government Digital Service), a new publicly-funded office to transform the way the UK government operates online. How did this opportunity come about? In other words, what conditions do you think were necessary for government to come around and commit to a new way of doing things?
I think setting GDS up was one thing. Being able to go as far and as fast as we did was another. As such, the conditions were not unique, but they were pretty rare. Quite a few people, of which I was one, had spent a considerable amount of time before 2010 talking to politicians in a non-partisan way, making them aware of the fact that reforming public services through digital means was not just the way to go, it was absolutely necessary. We essentially said, “If you get into power, we will help you do this, but don’t underestimate quite how much change is required and how much top cover we need to make it happen!”
Once there was a new administration coming in, we had a set of ministers of different parties who understood and agreed that the traditional way of ‘procuring IT services’ in government was an absolute joke. There had been so many expensive failures, so much political damage, with some IT projects costing tens of billions of pounds coming disastrously off the rails. So we were at this unique juncture where there was a real appetite for a new ‘internet-era way of working’.
In Francis Maude we were also lucky to have had a minister who had been around the block, knew a lot of other ministers, and was a superb operator with a great forceful personality. He didn’t have any ambitions to be Prime Minister, which meant he was willing to spend his political capital, not try to earn it. He pushed for bringing on very talented external advisors, like Martha Lane Fox, a successful internet entrepreneur and respected digital spokesperson here in the UK. I came on with Martha to help establish GDS.
[Offscreen] Anyone who’s ever worked on publicly-funded projects has experienced the endless bureaucracy and the political power games. How did you manage to create an agile, dynamic process that wasn’t stifled by the environment it operated in?
Having a deep understanding of how bureaucracies work is essential. Mike Bracken, who joined us from The Guardian to lead GDS, and I both knew that the logic inside the civil service was insane. It was a strong logic, but it was insane.
We basically set up and scaled a startup inside the government and then tried to protect its culture by any means possible. We intentionally chose an office location not inside existing government buildings. We hand-picked a talented team with some people from inside and some from outside government who all had that kind of anger with the current way of working, a strong desire to see it improve, with skills centred around user needs, and no interest in playing games. The idea was to build a bubble in which the main strategy was delivery: let’s deliver great stuff so quickly that bureaucracy can’t catch us. And we would be completely transparent about it. We’d blog, tweet, talk to journalists. Even the code was open.
Within eighteen months we’d replaced hundreds of government websites with GOV.UK and fundamentally shifted the way information was published through a massively collaborative effort with like-minded people across all parts of government. Then we realised that the real heart of public services isn’t just providing information, it’s transactional service: paying taxes, receiving benefits, buying a fishing licence, receiving power of attorney, et cetera. It’s great to have a website that tells you what to do, but doing the thing, that’s the actual service. And that’s a much harder problem to solve. They were traditionally seen as ‘IT problems’, but we knew they were service design and operational problems.
So after two years we cashed in all our political capital. We stopped being a bubble and tried to spread our way of working to twenty or thirty umbrella locations around the UK, which turned out to be a much bigger undertaking than setting up GDS in the first place. After four years, many parts of the bureaucracy were starting to adopt digital ways of working.
[Offscreen] As you said, GDS openly shared a lot of its internal language, principles, and guidelines, one of which was ‘start with user needs’. This seems pretty obvious to most people working on the web, but why is this such a revolutionary thought for people in government?
It’s hugely challenging to the internal logic of *any* organisation made up of multiple silos, not just people in government. At one point it becomes about what a specific department needs, rather than what the customer needs. That’s what drives most decisions. The internal games of climbing up the greasy pole become the purpose of those running each silo.
The intention behind that principle was to shock people out of their internal logic. Actually, our version of the principle ‘start with user needs’ had a little asterisk. If you followed the asterisk to the bottom it said ‘not government needs’ and that was actually as important as the first part. I’m afraid the footnote disappeared recently – deemed no longer necessary, though I’m afraid it still is very much so.
[Offscreen] How did you enforce that principle more practically? With a lot of user research, I assume?
Yes, we made use of a lot of videos of user research, people expressing their frustration and fears navigating existing sites and services. That was very powerful, particularly when we showed it to ministers.
But often the first conversation would be about the conceptual framework that people were working in. A civil servant would say, “We need to rebuild the site that allows people to apply for welfare!” – to which we would push back by asking, “What’s the user need here? Is it to build the world’s most efficient welfare system or is it to support people to get healthy quickly and get back to work?” Or another example: fishing licences. The established attitude would be, ‘We need a licence issuing system so that people can print their licence instead of our sending them in the mail.’ So what’s the user need? To be able to fish without breaking the law. You don’t need a piece of paper to get the green light – a simple text message makes it even more convenient.
It sounds like a little thing, but framing the problem in a different way can sometimes be all that’s needed to save millions if not billions of pounds. By putting user needs first you suddenly open up possibilities for service design, the operating model, *and* policy that are far more empathetic and more efficient. I think that concept is now understood more deeply in the UK. People accept that starting with the user’s need, then applying a solution assisted by technology, offers far bigger breakthroughs than simply digitising an analogue, paper-based process.
[Offscreen] Another principle you mention in a lot of your talks is ‘the unit of delivery is the team’. What do you mean by that?
It really means two things. Firstly, it means that the culture within a team is the most important factor for it to be effective. We never hired rock stars, no matter how talented or driven they were. We hired people who had the empathy and openness required to work well in a collaborative environment, people who recognised that there is value in leadership *and* followship coming from *all* members of a team. It means celebrating the output of the team and not making heroes of individuals. And that is the unit you should be assessing against: the output of a brilliant team.
Secondly, we recognised that there is an optimal size for a team. Once there are fifteen people you may need to split it into two or even three teams. In that sense, scaling doesn’t mean putting more people in a team, it means creating more teams. Again, this sounds easy, but with traditional programmes in government or large organisations you very quickly end up with enormous teams of hundreds if not thousands of people, which is a recipe for an expensive disaster. Start small and fast; stay small and fast.
[Offscreen] Your work was also guided by the idea of ‘policy visible in code’. Can you give us an example that illustrates what you mean by that?
That idea came from my former colleague Richard Pope, one of the smartest people I know. He always felt that the fate of people is increasingly decided by algorithms – not just in everyday life but also as they interact with government services. Many policies are just instantiated by an algorithm that, for instance, determines whether you are eligible for a benefit or not. Richard felt obligated to make those algorithms more transparent.
In a very basic way, we did this by open-sourcing all of our code. And it didn’t take long to realise the benefits of doing so. There was a case where a member of the public noticed an error in our code and notified us within thirty-six hours of going live. We were able to push out a fix within the hour. As this was happening, a story made the newspaper that a solicitor had found a mathematical error in an interactive PDF calculator provided by the government to determine divorce payouts. For eighteen months this solicitor had tried to persuade the government that there was an issue with the calculator but he simply wasn’t heard. Going to the media was his only option to make the government aware of the issue. And mind you, that issue affected tens of thousands of divorce payouts.
These two examples perfectly highlight the benefit of running government services more transparently. To be fair, the member of public who alerted us to our bug was a developer. The long-term goal is to make those algorithms understandable to non-coders as well. We need a change of language so that everyone knows how and where their data is being used. We need new design patterns and consent models that put an end to the black boxes that are government policies.
[Offscreen] Where do you see the biggest opportunities for digital government services? Is the goal to ‘get out of the way’ and reduce government to a search bar?
In government the risk of ‘designing for people like us’ is much greater than for most businesses. Many of us are in a position from where we see anything government-related as a hassle. We want it to disappear or reduce the friction of government interaction to zero. I think that’s a dangerous over-simplification, because the people who really need government the most don’t want the government to disappear. Quite the opposite. For those who are really struggling – and it could be any one of us at some point in our lives – the government is a safety net, and that safety net cannot be invisible nor completely frictionless. We need to facilitate a proper relationship between someone with a difficult problem and a public servant empowered to build a relationship with them.
That’s where digital poses a real opportunity. For those who would like government to ‘get out of the way’, let’s make it frictionless. This will save the government money and the individual time. Let’s then invest that money to make government more present and more useful for those who really need it – be it digitally, via phone, or face-to-face. No matter the medium, people in need require the state to be empathetic, efficient, effective, and human.
In the pre-digital era it kind of made sense to have one usually paper-based process to serve sixty million Brits. This inevitably caused a lot of bad friction for a large amount of people. But now it doesn’t have to be a one-size-fits-all solution anymore. Digital technology allows us to offer a differentiated service design that suits the needs of the user much better. With the concept of ‘government as a platform’ we can build unified systems that underpin all government services across the board – things like identity assurance, payments, notifications, and so on. This makes developing a range of different, more targeted service designs much easier and cheaper.
Unfortunately, no current government has really grabbed that opportunity yet. The thinking is still very much, ‘We have one policy, ergo we have one service design to deliver that policy.’ Whereas my vision is to deliver a particular policy through a range of service designs that adapt to the user’s needs.
[Offscreen] Do you think an improvement in service delivery is enough to regain the public’s confidence in government and politics in general?
That’s our only bloody hope! There is currently a gigantic gap between service quality in government and service quality in the best of the commercial world. The wider that gap gets, the bigger the distrust in government becomes. The result of this is dangerous: just look at recent elections around the world. Our only hope is to make government serve its people in a way they expect it to – efficiently and empathetically.
That’s just the baseline though. I think we can then go beyond that and do some interesting things around accountability, transparency, and potentially allowing people to shape policy. But right now, we’re far from that. We’re in a delivery crisis in which we have nineteenth-century processes that over the years have been badly corrupted by ‘adding IT’, costing billions, and getting ever further behind expectations.
What we’re talking about here is civil service reform. It’s a reform of the machinery of government more than a reform of politics itself. I think the vast majority of politicians are now reasonably aware of the power and ability of digital technology – just to get elected you need a good understanding of it. What they are not aware of, though, is the rotten, antiquated state of the machine they will inherit once they get into power. That machine is not able to deliver the simplicity and quality of service citizens now expect. It’s going to take something reasonably dramatic to overhaul the current shape of public services, which is made up of lots of vertical silos. What we need instead are horizontal layers connecting all parts of government.
So there is a crunch point coming. The good thing is that we can become involved. In fact, I believe those of us who work on the internet have a duty to pull up our sleeves and do our part to reform the machine.
[Offscreen] What would a digital-first government of the future look like?
It would most certainly have fewer people, but those who remain would be far more empowered. They would be working in small, multidisciplinary teams consisting of policy makers, lawyers, operations people, developers, designers – you name it. Those teams would have the authority to make decisions quickly to iteratively deliver a service design that meets users’ needs and achieves the minister’s intended outcome. The system would be underpinned by a twenty-first-century data infrastructure with shared platforms around identity, location, consent, payment, notifications, and so on. These platforms would be operated by the government but also made accessible to the commercial and not-for-profit worlds in order to increase the economic and cultural performance of the whole country.
The impact of some of the more recent technological developments is difficult to predict – for example, machine learning. The impact on areas like health services or education could be dramatic, but it requires a lot of discipline around data and consent. I think that’s something governments may well get catastrophically wrong before they get it right. Another domain that’s still very scary is cyber security and the fact that a bunch of hackers can take down a whole country. Again, it’s something that governments will probably become better at over time, with many bumps along the way. Then there are lots of regulatory challenges. Global competition laws, for instance, are woefully out of date given the concentration of power that comes from data monopolies.
[Offscreen] Going back to a point you made earlier, how can we ‘do our part’ to reform government?
I’ll be honest. If you live in a stable, democratic country you should give five years of your time to uphold that stability. We owe it to our countries to help them close the technology gap. Our institutions are crucially important. Everything else is just play.
For many years I worked outside and around government trying to influence how things are done on the inside. I can tell you now that the impact you can achieve working from the inside vastly outweighs anything you can do from the outside. It’s important to understand that what you are actually changing isn’t politics, it’s the culture of an organisation. You cannot change culture from the outside.
I admit it, it’s not always sexy, but the work matters. And I have an even tougher message: if you already work for government and find that your efforts to change things are fruitless, then go and try to get fired for doing the right thing! Why are you waiting for permission? If you feel like you are a victim of the machine then just quit. We need people to go in there and ruffle some feathers, show good leadership, and inject a new culture. If the machine rejects you, fine, try another way in or do something else. As someone working in tech you’re not going to struggle to find a new job, really.
[Offscreen] So you think software designers/developers carry a bigger responsibility for improving government than everyone else?
An old friend of mine, Matt Jones, said something that really resonated with me. He started by quoting Arthur C. Clarke’s famous law which goes like this: any sufficiently advanced technology is indistinguishable from magic. In practice this means you can now tap a button on your phone and two minutes later a cab turns up in front of your house – that’s magic! Matt then added to that, ‘And magic is a power dynamic.’ If you create magic, you have power over those who consume magic. In other words, if you are writing code or designing services that hide ‘the magic’ away from the user and you are the only one with the key to changing that magic, you are in a *very* strong position of power. And with great power comes great responsibility. Let’s be honest, as developers and designers in today’s ultra-connected, digital world, we really have the power to make or break lives. It’s that simple.
Sticking with the Uber example, there has been a huge amount of negative press and commentary around their dubious business methods. What really interests me was why their engineers decided to go along with this. Clearly, Uber’s leadership had a certain goal in mind, but why didn’t their engineers think twice about the ethics of some of the code they wrote? To me that again highlights the power each individual has – that personal responsibility to do the right thing and be mindful of the impact of one’s decisions.
So, if you’re making software, you are actively reshaping power dynamics between citizens, consumers, companies, and governments. Shifting power relationships is politics. If you are in software, you are in politics. I don’t think we as citizens and consumers have yet properly understood the power technology gives us. However, we’re lucky to be part of the digital revolution, because unlike most generations we’re able to shape and build our future.
[Offscreen] That’s a great way to circle back to my first question about that *Wired* cover from the ’90s. In retrospect, has the internet so far realised its potential to ‘begin the world again’ and if so, has it been for the better or the worse?
Well, I didn’t see that headline back then and think, ‘The world is inevitably going to get better!’ What I thought was, ‘My god, we’ve got a duty to try and make it better because it’s inevitably going to change a lot!’ And that’s pretty much how I still feel today.
It’s no surprise that some bad things have happened and continue to happen, but equally, some brilliant, life-affirming things have come from it too! And so it’s a constant battle to bring our values to bear at a period of rapid change, to maximise the benefits for as many people as possible and reduce some of the inevitable side effects. Because there is only one certainty: we’re never going backwards with technology.
Many of the negative effects come down to culture. We cannot just be passive consumers or companies selling products. That’s why I worry a lot about Silicon Valley culture where anything goes – *anything* – in order to become a unicorn. The stories about normalising unethical behaviour and growing attention monopolies at any cost really appall me. History should teach that a predatory culture like that never leads to a good place. Hopefully, over time we get better at understanding the global network effects inherent in some of these companies and get better at regulating them. In the meantime, we need to work on what is socially acceptable amongst our peer group and weed out rapacious, unethical behaviour.
But at the end of the day, we’re only safe because we have trusted public institutions which respect the rule of law. That’s the only reason we’re safe on the streets. Public institutions need our help to keep up and stay relevant. So it’s time to shape the future we want and be engaged, pro-active citizens – not just consumers.
Copyright Offscreen Magazine, 2017. Republished with permission.
Subscribe to Offscreen magazine.