Richard Pope: Beyond user needs – A new design philosophy for the digital public sector

Richard Pope: Thanks everyone, good afternoon. Yeah, I’m going to re-say something. So my background is I was part of the founding team of the Government Digital Service in the UK way back now. It seems like a long time ago. I didn’t have a grey beard. In 2011, I worked on gov.uk. Obviously, Mel was the first product manager on that. I also worked on Universal Credit, which I’ll talk about quite a bit. I’ll talk about both of these quite a bit today in the talk. Universal Credit is the UK’s digital welfare, or digital first working-age welfare system. I’ve recently published Platformland, which is an anatomy of next generation public services, and also a short book on government as a platform. I’m an independent consultant working mostly with governments, national and city governments around the world. So I’m going to try, in the time we’ve got available, to talk about three things today. And hopefully it’s going to be a bit challenging. It’s the last talk, so I think I’ll try and give some food for debate. And I’ve been asked to leave some time for questions, so I’m going to rattle through some of these so we’ve got some time. I’m going to try and cover, and this is all from the experience of projects I’ve worked on, so what I think the limits of, some of the limits of user needs are, and how we need to expand beyond those in some context, expand beyond those. Why seamless design in public services is not the target, and what we can do about all this, and why it matters right now. When we started with the alpha of what would go on to becomegov.uk in January 2011, we started with user needs. The whole design was based around the idea of having a definitive web page, a definitive URL for each need, each task a user needed to do. We printed out and cut up on tiny bits of card the existing web pages of existing central government websites, and lots of exercises to reform those into user needs. We built something, a screenshot here called Benidatron, which was, we bolted onto the front of a publishing system, and nothing was allowed to get in unless it was in the Nedatron. It all seemed very radical back then, but actually it was already, the idea of user needs was actually quite an old idea. If you ask most people to think about user-centered design, they’d probably picture a Silicon Valley designer or something, but they’d have the wrong picture in their heads. They should actually be thinking about a coal mine in northwest England, and rather than a West Coast designer, they should be thinking about a Liverpudlian social scientist. So after World War II, the UK nationalised its coal industry, along with lots of other industries, and the new National Coal Board was experimenting with new technologies, and they hired a young researcher called Enid Mumford, she was a social scientist, to try and understand the impact of that technology. This is her in later life, and she went on to become the first business professor, a first professor of a business school in the UK, and various other things. But early in her career, she was down the mines, talking to miners, trying to understand how technology affected them. According to her obituary, she apparently would smother herself in perfume, to give them lots of warning, to make themselves decent before she arrived. In another project, she also took up a job working in the canteens in Liverpool docks, so she could understand the working practices of people working in the docks. And she’d go on to apply those, what we now call user-centred design principles, to computer systems. So she worked with users in newly computerised organisations, so banks, but also government departments, which is mostly back-office users that she was dealing with there. And she developed what she called the needs-fit theory of job satisfaction. And she said that computer systems should be built around the working practices and preferences of the people using them, and that successful system design required a fit between the balance, the needs of users of the system, the social system, and the capabilities of the technology, the technical system. And failure to involve users meant that projects had a high risk of failure. So this is in the 60s and 70s. This is not new, but it’s taken a long time to work its way through to government. A couple of years later, in 1977, the idea of user-centred design was formulated by Rob Kling. He wasn’t a designer either. He was a computer scientist. He was from Indiana, a bit closer to the West Coast. Anyway, he published this paper here on the organisational context of user-centred software designs. And just to pick out a couple of quotes on that, which I think will sound familiar to many of us who worked in the public sector, systems that were imposed on users, and designers, this is designers of software systems, see themselves as change agents, reforming an inefficient organisation, and systems which are poorly designed and do not meet the needs of users are actually not effectively utilised, and they don’t satisfy people to use them. So he based a lot of his work on Enid Mumford’s work. And then the idea got popularised by this book here, which many of you may be familiar for, Don Norman, who went on to go and work at Apple. And then kind of fast forward, by the 2000s, user-centred design, user needs have become a discipline in and of itself. And most internet age companies had adopted practices like user-centred design, user research, prototyping, agile delivery. But I think on the way it became quite minimalist, and functional and utilitarian, the priority was getting a task done. Also quite individualistic compared to the early, the origins of user-centred design, which was more like understanding a social system, it became about a focus on designing the interaction between the user and the service. So Google, for example, say in their design principles that services should work so well that people don’t have to consider how they might have been designed any differently. But it’s about a user interacting with a service. So it wasn’t new. Despite the public sector origins just about of user needs, the user-centred design, it took 30 years before the approaches that Mumford and Kling described to become accepted practice in the creation of digital services, at least in the UK. The situation when we started on Gov. UK was very much as Kling described in the 70s, that the practice was one of insulating ministers and civil servants from software engineers and designers. It was just like it was a thing that was done over there. With Gov. UK, we started to really bring the focus on user needs and bring the development and design surface closer to the organisations themselves. And we really baked in this idea of start with user needs right from the beginning. So this is a screenshot from the Design of a Year Awards and it shows Gov. UK as it was when it was launched, but also the first of our design principles, which was start with user needs. In 2013, something I worked on, we published the Digital by Default Service Standard, as it was called then, that had requirements to understand user needs and to build a working prototype around user-centred design methods. And this, over time, this became, or started to get accepted as good practice within the UK Central Government Civil Service. And it made things better, definitely. In the first five years of Gov. UK, sorry, of GDS, it saved, by some estimation, up to £4 billion. And it made services simpler and faster and clearer for users as well. News stories about failed IT projects, which were common beforehand. There was less of them. Ministers wouldn’t go on the news and then announce something and it crashed two minutes later, which was happening quite a lot. So, it made a big difference. But there were already signals a couple of years in, certainly on the projects I was involved with, that user-centred design and user needs were necessary, but not necessarily sufficient for creating the types of digital services that we needed to see for the digital age. So, this is where I’m going to talk about universal credit a little bit. So, universal credit is a single working age benefit for people who are either unemployed or in work, but in need of financial assistance. The bulk of people are actually in work in some shape or form. And I was part of a small team made up of the Government Digital Service and Department of Work and Pensions, which is our welfare agency, social security agency, tasked with creating a workable implementation of it. An initial attempt, which had followed a very old-school rights and requirements, give it to a couple of big companies, hope the best, had spent £700 million and not actually produced anything that worked. So, the task of the team was to try and make something that worked. But we found out that, actually, some of the approaches from Gov.uk and some of the approaches that worked on Gov.uk didn’t really map onto universal credit. So, the first one was actually the fact that Gov.uk had followed this approach of, you know, just by discovery, but like an alpha, beta, it was a linear progression. Universal credit didn’t work like that because it, there’s a class of problem, a class of design problem where you only really understand what people need from a system and how it should work and how to design it through use. So, we had this approach of more like test and learn. So, we started with real users, really, really small scale, in a single postcode district in South London on one day. We opened the tap, let some real users in, shut the tap off again. So, we had real users to work on. And it was just like test and learn from there. there was no design phase, beta phase. It was much more iterative. We also found that the user-centered design, certainly the one that we, all the reference points being from commercial services, tend to talk, as I mentioned a moment ago, around like the user as a singular entity who has needs. Universal credit is actually a family benefit. It’s paid to a couple. And then suddenly, we’re designing a world where we need to explain the inner workings of a couple to each other in some instances. There is no single user. That comes with it, comes with a lot of complexities, as I’m sure you can imagine. But actually, the language wasn’t there to be able to explain it necessarily. User needs also tend to be quite task-based and transactional. A user wants to achieve an outcome. And that’s great when it’s like, I want to pay a parking fine or get a passport. But not all of government policy operates in those areas. Sometimes government policy wants to shift mental models or help people reflect and change. So this example here is some tools to help people think about the type of work they’ve done in the past and might do in the future. It’s not really something that a user directly needs to do. It’s not an outcome-based thing. So the language didn’t quite map over. And the way that the universal credit policy was designed, it was very, very data-intensive. So this long list you can’t read here is all of the things. and it’s so big you can’t even fit it on screen. All of the things that a user needs to know that they have to report to get paid the correct amount of money. And if you follow the approach of designing each one of those as an individual service, it’s not going to work for people. So we needed a different approach. This example is common in welfare. It’s common in some other policy areas as well where actually government chooses to place burdens on people and those change over time. So actually, in universal credit, if you’re not working, you’re expected to do a lot more than if you are what the organisation, what DWP would say is earning enough. So you get these conflicts between what and actually… Users don’t feel like they want to do something. They don’t feel like they have a need, but the system’s expecting it of them. And going back to Rob Kling and Edith Mumford, so Kling noted in his original article on user-centred design that it only works when users of the system don’t have incompatible interests. Actually, in the welfare system, sometimes government officials and the public have incompatible interests. And for Edith Mumford too, she describes her idea of needs fit. Job satisfaction was based on the idea of not just understanding what people want for the system, but actually making it more democratic, giving people a say in the design of systems. So beyond user research, actually, being able to help people have a committee in a bank that helps the staff design the system, for example. Now, in Universal Credit, there’s not a mechanism for the public to negotiate what is an appropriate burden or not outside of elections for users of the system there isn’t one. It also turns out, and there’s quite a big body of research, that people don’t just care about outcomes. So this kind of task-based idea that people need to get an outcome for something is not the totality of it. People care about how services make them feel and how services make people feel actually potentially has a carryover into how they interact with government elsewhere. So people really care about how services make them feel and that kind of very transactional approach didn’t carry over very well either. So rather than user needs, like user needs were very important, we used that language, but actually when it came to the design of the account, it was more about designing something that allowed users in government and in the public to manage fluctuating what’s called administrative burdens. So this idea of like, there’s a bunch of stuff you’ve got to do over time and you need to stay on top of it and sometimes things go wrong and you need to be able to go and check things. So we had this idea of a to-do list and a journal and they were really trying to make sure that the burdens that the state was placing on the public were really, really clear and vice versa. There’s an obligation on government to be really transparent about what it’s expecting people. It can’t hide behind what you should just know, which is what it’s done in the past. You should know you should have filled out that paper form and put it in the posters. So it’s trying to make it really clear. And the journal was partly a way of talking to government officials so saying, I’ve moved house for example and providing a change of information but it was also a record. So it provided a record of all of the things that someone had done, all of the information had been reported, all the meetings and I know from having heard MPs speak, members of Parliament speak, they get people coming into their constituency surgeries with these things with that screen loaded up on their phone going, look, they’ve messed up, they should have done this, they should have paid me this amount. So it’s a record for the public. It’s helping people and that’s not a really kind of tangible outcome. It’s not a I need to apply for this or I need to report this. It’s really about helping people manage this flow of information and manage and give people a way of spotting when things go wrong, if they go wrong. I think this is the final one of this description of what’s different about universal credit or how we approach it differently. It’s also an inherently relational service. It’s not just a digital service, it’s a digital service and face-to-face meetings and phone calls. being able to decide giving the organisation the choice of when and how it interacts with people is really important because some things, if you want to explain to somebody that maybe they should try a different job or commute somewhere else or think about childcare anyway, that’s not something you can get over to somebody in a text message or an email and expect that they’re going to change their behaviour. It’s a relationship so you need to be able to allow creating a digital service where people can switch between online, offline and it’s not about a single channel it’s about fluid movement between them is really important. So I would say one of my takeaways from working on that project and work is that user needs are necessary but not sufficient in delivering digital public services that we need. User-centred design is definitely here to say and definitely critical to the delivery of public services but we also need to think about how we create the language and the processes for behaviour change for things that are less individualistic for managing fluctuating burdens for thinking about things that are not just outcomes but kind of how people feel about services and design of more relational services and later I’ll come back to some examples of how I think we can do that. So the next thing I wanted to talk about was seamless design. Now I’m sure many of you worked on government digital strategies. and you’ve probably seen this word pop up a few times. It’s normally held up as the ideal. We should just make it seamless to do X, Y, or Z. My argument is that it’s not an appropriate target in the public sector. I’ll explain why. So on YouTube there’s a Steve Jobs video where he says it just works seamlessly like 44 times. I don’t know who made it or why they decided to make it. They obviously really love Steve but someone did it and that really sums up his and Apple’s approach to design. Design should get out of the way. It’s not there. It shouldn’t distract. So from the beginning Apple products were designed as tightly sealed boxes and when they even when they discovered that people were repairing them in repair shops they changed the screws. You can see there they changed so people couldn’t meddle with the design. There’s one school of thought that it’s enough to take those principles that it should just work to the state. And you do hear it sometimes. You hear people say if only all the council need to do all the municipality need to do is collect the bins on time and collect the taxes. That’s all they need to do. But those sorts of arguments tend to come from people who have quite a privileged position of how they interact with the state. They tend to be people who mostly just deal in transactional services renewing a passport every once in a while. For people who have more messy complicated interactions with the state it doesn’t really wash. It also I think fundamentally misunderstands the idea of what makes public services public which I’ll talk to a little bit now. So public services are partly public services are different because people have to use them. I know that’s the thing people say is obvious but it’s important it’s true. They must work and they must work for everybody. And I don’t think we want our public services to be working like an iPad. They’re functional but good luck if you want to understand how it works. The reason being that public services only get better through co-production with the public. I think some of the things Oleg was saying a minute ago around providing opportunities for the public to engage and help shape how their government works is really important. That’s a core feature of public services in my opinion. And for reasons I’ll explain in a second, seamless design makes that harder. I just wanted to say a little bit about the concept of co-production before though. This is the second history lesson hidden within this presentation, if you’ll forgive me. So the concept of co-production comes again from the early 70s from this woman Eleanor Ostrom. So she’s an economist, a Nobel Prize winning economist. And her lifetime’s work was informed by field work with communities who relied on shared resources, common resources of things like fisheries and watersheds and forestry, but also public services. She was very anti the idea that the tragedy of the commons was a real thing. Her argument was that no, we have much more complicated governance structures which are partly bottom up and top down. And actually communities have ways of developing sustainable governance mechanisms. Anyway, I digress. Co-production. I think the first time she defined it was when she looked at violence. She was asked to explain why the crime rate in Chicago had gone up, even though they’d made the police more efficient. So they put the police, they’d made, they’d obviously done some calculations and worked out where they needed to put all the police cars and they’d put the police officers, they’d taken them off the beat and put them in the police cars and it was super efficient so they could get to the crime quicker. And she, through her research, pointed out that actually the police need the community as much as the community need the police because crime reduction is something that is co-produced between the public and the police through conversations, through the reassurance of seeing police officers, through police officers understanding really what a community wants. So it’s a co-production. That is an example of what she would call the service paradox. So this is where the better services are, and this is a challenge to all of us who design things for a living, the better services are, as defined by professional criteria, the less satisfied citizens are with the outcome of those services. So to give you a specific example, if you think about a textbook in a classroom, if it’s so well designed that kids don’t need to put their hand up and ask a question or talk to the person sitting next to them, then actually you get worse outcomes because education in a classroom is a co-production between students and a teacher and a textbook. There’s also a study, which is referenced in the book, I can’t remember the name of the researcher now, that students at universities have shown a preference for library books that have got scribblings in the margins and highlights, because it’s a useful thing. One of my go-to quotes is this, which no one’s quite sure who it’s from, it’s either from A.J. Balfour, who was a Prime Minister of the UK in the turn of the last century, or Geoffrey Howe, who was a Cabinet Minister and Deputy Prime Minister a bit more recently in the 80s. And it’s democracy is government by explanation. And this gets to the root of why I think the design of seamless services is an anti-pattern. it’s that, so just in the same way if we design textbooks for students, it’s so easy to understand that people don’t have to ask questions. What if we design our government services in the same way? What if we don’t give the public an explanation of how things work? Because democracy is about more than voting every four or five years, it’s about, again, as earlier we say, it’s about the opportunities that people have to shape the world around them and the services that they rely on. So, early in the days of GDS, our version of it just worked seamlessly was this and do the hard work to make it simple. And it’s a really important principle. I think it went to the heart of what we were trying to do with Gov.uk and a bunch of other services around just really sweat the design to make sure that people can just use it and it works for them. It’s a really important principle. because people shouldn’t have to understand how government works to get something done. Absolutely, they shouldn’t have to. But I think as design practice spread across central government in the UK, it almost took on a bit of a life of its own and somewhere along the line it crossed the line from people shouldn’t have to understand how government works to people shouldn’t understand how government works. It’s an inconvenience. It’s going to get in the way of serving the core need here of get that passport. sign a petition or whatever it is. So people shouldn’t have to. But that actually had a real effect. There’s a couple of examples here. I’m not sure if you can see this on Facebook. This is two screenshots for the Appeal of Benefit Decision Service. This is the place you go if you want to take the UK government to court because you think they’ve paid you the wrong amount of social security money. And the reason for showing it here is the one on the left here is branded as government UK. Legal campaigners really kicked up a stink about this, rightly so, because they were saying that it’s important that the judiciary is independent and is seen as independent and it is a separate thing. And it was not badged as that, it was badged as the government. So for a user, they’re like, oh, I’m kind of seeing the same thing. I’m reporting you to yourself. How’s this work? Eventually it got changed, so it now, at least on the service pages, says the courts and tribunal service. And I think something we didn’t really acknowledge in the creation of government UK is that we were, sometimes we were making things simple and clear for people, but we were also making some things harder to understand as well. There’s a concept in geography called wayfinding, which I’m sure many of you heard of. It actually comes in this book here, which I really recommend, if you can find a copy of it, it’s brilliant. It’s called The Image of the City by Kevin Lynch. And he came up with this term wayfinding, so how we find our way around a city based on cues and the environment. He also came up with the idea of imageability. So he compared some different cities all in the US, so Boston, LA and New Jersey, I think. So Boston, if people have been there, it’s quite European, it’s winding streets, there’s lots of landmarks. New Jersey and Leicester is very kind of just grids. So imageability was the extent to which a place invokes a clear mental image in people’s minds. And it’s how people navigate cities. We look for maps like this if they’re there, but if not, we look for that church tower, we look for the river, we look for the road or whether shops change in some indeterminal way. And in design of services like gov.uk, there’s a risk. So things that abstract away the complexity of government for very valid reasons, there’s a risk that we end up designing low imageability services. so actually it’s really hard to understand the bit of government you’re dealing with and why. And most of the time you won’t need to, most of the time you can just ignore it. But if things go wrong or people need reassurance, am I dealing with the immigration system or the health system here? I’m not sure. If people really need to know, there needs to be a way for them to find out. It’s not just the public who need to understand how services work though. Civil society organisations can find their job hard to do as well. Their job in part being helping makes public services better. So long after I worked on Universal Credit, I was there, the original team that was there for six months and then the department ran with it and built a brilliant digital team. I was involved with a little bit of work with this organisation here on an advisory panel and so I became familiar with it. This is the Child Poverty Action Group and they’re brilliant. They’re a civil society organisation who campaign to help people get the entitlements that they deserve. But they also really think quite hard about digital and they tried to understand whether the law had been applied properly through the design of Universal Credit. And they found for the most part it had. But they found some gaps. They found some cases where if you, a certain set of criteria, people were going to be missing out on some of their benefits. But to get to that point, they had to, you know in courtrooms where you have illustrators who draw a picture because you don’t have to have photos in the courtroom, they had to do something a bit like that. They were leaning over the shoulders of people using the system, drawing pictures of it so they could understand what bits were missing because there’s very little on the public record about how Universal Credit works. Same for most services. Some services will publish their source code but if you want to understand what an average user journey looks like, what the data collected is, what the database looks like, it’s really hard most of the time. Digital is inherently quite opaque. There’s another bunch of welfare rights campaigners created their own click-through version of the application process using Google Forms so they had a reference point. There’s no ground truth otherwise. digital can make it much harder for civil society organisations to do their job. It can make things opaque regardless of how well user research is being done, regardless of how brilliant the team is. It’s a bit like this. I don’t know how many people anyone who’s read Hitchhiker’s Guide to the Galaxy? Some people. I’ll try and explain this. In Hitchhiker’s Guide to the Galaxy, the guy in the dressing gown wakes up one morning to find out that his house is about to be demolished. He’s told by the council official, but the plans were on display. He’s like, on display, I had to go down to the cellar. He goes, that’s the display department. I had to have a torch. The lights have probably gone. The stairs have gone too. It turns out there’s a leopard waiting at the bottom of the stairs as well and it’s locked in the filing cabinet. He couldn’t find out that his house was being demolished because the friction to find out was just too high. Civil society organisations, campaigners trying to understand how digital systems work, find themselves in a very similar situation. It’s almost impossible to understand how digital systems work. It’s more important than ever because digital systems change all the time. Universal credit gets updated every two weeks. It’s personalised using an individual’s data or a couple’s data. So what one person sees is very different to another. We need these ground truths to explain how things work and how they’re changing. Just in the saying that we have public registers of planning systems, public registers of companies, we need the same sort of thing for digital services and infrastructure. If we don’t, we get things like this, what I call digital fog. So in the absence of transparency, in the absence of information about how things work, people make stuff up. Anything can become a conspiracy if you don’t know how it works. So this is a headline saying that DWP, the welfare agency, are going to use AI robots to decide if claims were true or not. It’s like probably not actually. I’m sure somewhere there’s some people doing something for fraud detection, but there’s not a single AI, almighty AI system that determines whether people get money or not. But because there’s not a reference point to how these things actually work, people make stuff up. People fill the gaps. And that has an effect on people’s trust in government and again the ability of, in this instance, the media to hold government to account. So to sum up this section, seamless services, the reason I think they’re an anti-pattern is they block, they don’t explain themselves, they block, and in doing so they block understanding and they limit co-production. It’s harder for users to understand or query something. It’s harder for civil society organisations to do their job. So it’s a democratic imperative, but it’s also a service design imperative as well. Public services get better through co-production, so you need to make sure that those routes for co-production are there at the point of use for people. So what can we do about this and why does this matter now? So my contention is that the public design practice that emerged from 2012 onwards was brilliant in many ways, but it was also incomplete, and it’s kind of our job now to fill it out a bit, and I think we’re able to, now we all know a bit more. So the question is not whether user centred design is here to stay, it’s a necessity for delivering next generation public services. The question is whether it’s sufficient on its own, and I don’t think it is. It’s not just design, it’s not just design practice, it’s the intersection of design practice and how government assigns value. So the aim of most digital programs is actually the status quo delivered more cheaply. Universal Credit is a bit of an odd exception there, GovUK is as well, because they were both new novel initiatives. most digital projects tend to be, we’re doing this at the moment, but we want to turn off that expensive mainframe or stop doing it on paper. That means there’s whole areas that user centred design, good digital design is never applied to in the first place. Efficiency is a trap, this idea that we should be going after efficiency all the time, it’s a trap. It’s very hard to digitise something and to make it more expensive if you do it right. you’ll probably get that for free anyway, but how do we decide to look in the right places to get the efficiency savings is the important thing. Digital rarely gets applied to the frictions between departments or the gaps between one service and another or gaps between the public sector and the private sector. So we need an approach to design and an approach to business cases and an approach to funding that allows us to decide where to act. So I think the two critical questions are these. Can we design for more than just transactional outcomes whilst also saving money but still keeping users at the centre? And can we reconcile this idea of it just works with government by explanation? And I think we can for both. And I’ve got a few ideas here. There’s more in the book. I’m not trying to present this as the future design philosophy for government. It’s some ideas. I’d really like a discussion and debate around some of these ideas. So I think we can reframe or maybe alongside the idea of user needs start talking about public services that work harder for the public and public servants. Because I think that encompasses more than just someone needs to get a task done. It also allows us to talk about automation in a way that’s not just government saving money. It’s like that plus some other things as well. So what are the opportunities to use automation, maybe even AI in the future, but certainly today’s automation that’s a relatively well-known commodity, for the public good to save people time to actively seek out what’s hard for people and try and fix it. I think the language that’s useful here is this one of the which I’ve mentioned a few times already, is administrative burden. And the definition here is by two researchers, Pamela Hurd and Dominic Moynihan, I’d recommend their book, Administrative Burden and Policymaking by Other Means. They’re also very active on Blue Sky. And they define it like this. So it’s the result of learning, compliance and psychological costs placed on citizens and public sector workers. I’m not sure that’s actually their exact definition, but it’s the language they use. They talk about learning costs, so finding out that something exists in the first place, compliant costs, fully informed, and the psychological costs, that thing I mentioned before about people not just caring about an outcome but how it makes them feel as well. That to me feels like a more rounded approach. that user needs plus administrative burden feels like there’s the beginnings of a new approach. And I actually think if we choose to take it, today’s technology, applied with that intent, presents us with the opportunity to systematically eliminate administrative burden across the public sector. Just take the faff out of things, but in a systematic way. But to do that, we need to measure it first. So we need to know where those things are. So where are the learning costs, the compliance costs, the psychological costs? We need to measure it. Brilliantly, the UK government, I did a bit of advisory work on this, the UK government published something recently called the Blueprint for Digital Government, and that actually contains this commitment to start to think about how to measure administrative burden in a systematic way. Hopefully, that will be published at some point. But only once we measure it can government decide when and where to act. I think we can also think about in the future, maybe some other rights or expectations, maybe we should say, rather than rights that the public should have. Some countries have got the expectation of wants only, this idea that you can report data once and the information is available to use elsewhere. I think we could add alongside that maybe a right to prediction in some circumstances, so that government, there’s an expectation that if government has a fairly good idea that you’re about to, I don’t know, leave school and there’s some information that might be useful to you, it should present that to you. A right to more real-time services, a right to interoperability between different services. These sorts of things start to get you to that world radically reduced administrative burden. You can articulate all these things in the language of user needs, but it just gets a little bit clunky. I think we need some new language. The second thing I think we should be thinking about now is really we need to reframe public design so we don’t just talk about the outcomes, we don’t just talk about users using services, but we also talk about making them understandable and accountable and democratic. It’s not enough for services to just work harder. So it’s this plus this. So how do we do that? As I said before, users should never have to understand the structure of government to get something done, but it doesn’t mean we should obfuscate it. Actually, we’re going to have to start thinking about how we reintroduce bits of understanding, even if there’s a little bit more friction, introduce it back into the design of services. That doesn’t mean that we should go back to designing every department and having its own website or anything like that. Absolutely not. But making sure that the institutions that are accountable to the public are surfaced somewhere. I would contest that actually there’s an obligation on public services to actively educate people about how things work, how their public services work, and where power and accountability lies. Who’s on the line if something goes wrong? Actually, I’ll talk about in a minute why that’s going to get harder in the future, but I think as a principle it should be really clear who’s accountable for a service. We can do that by designing, putting accountability at the point of use. Again, it’s a little bit of friction, but just making sure there’s opportunities for people to understand how things work. That might be who’s the organisation behind it, or it might be, and there’s some examples in the book of this, helping people when they think something is a bit wrong, helping them delve down and understand, how was this decision made? Was it an automated decision? What data went into it? Who do I call if something’s wrong? Can I get all the way down to the legislation so I understand that this thing is really doing what it should do? Most people won’t use that most of the time, but for some people who need it, they will do. I’m very fortunate not to be allergic to nuts. I ignore, for the most part, what’s on the back of ingredients list on the back of a packet of food, for some people it’s absolutely critical. So the information should be there for people who really need it. Everyone’s relationship with the state is not equal. Some people have to adopt a very different stance when dealing with particular parts of the state. It should also be a matter of public record how services work and how they’re changing. I think for people who don’t work in close proximity to software development teams, to digital teams, it can seem like digital is just like it’s a thing that you buy. It’s like it’s a fixed entity. It’s like you’re buying some paperclips or something. Rather than it being something that’s in constant flux, it’s constantly changing. It should be a matter of public record what those changes are and how they work. There are some good examples of this, particularly in healthcare actually, where you’re seeing, I think driven by clinicians need to understand changes to systems. There’s good examples in there. But it needs to become a systemic activity. Is that the right word? It needs to become the default activity of all digital teams that they publish information about how services work and how they’re changing. That’s, for example, screenshots of the service. Some teams are publishing things called design histories. Screenshots of example screens and also the thinking that went behind them. That’s part of the answer, I think. Just to summarize, why does this matter now? The next generation of public services, if we can say anything about them, they’re going to do more, probably. They’re going to have more data, they’re going to have more smarts, they’re probably going to be more joined up. That’s partly technology enables it, it’s partly people’s expectations. They’re going to be doing more, there’s going to be more automation, there’s going to be more real-time services, there’s going to be more services like Universal Credit that use a lot of data to tailor a service to an individual or a family. So they’re going to do more, and there’s a paradox in there, like hopefully they’ll do it in a way that makes things simpler for people, we can use automation to make things simpler for people, but simplification creates complexity, the system you create to do that, becomes more complex and more opaque. So the public are going to need new tools, like the examples I gave you from Universal Credit there, the to-do list in the journal, to make sense of things. So if you have the once-only principle and someone reports their address and then that has a knock-on effect to their car registration and their benefits and they suddenly find they have less money in the bank, they need to be able to go and understand what happened, why. I mean, hopefully that doesn’t happen, but things will go wrong, there need to be tools there that help people understand it, and they also need to be the information in the public domain that helps civil society organisations understand services so that the public and those organisations can both play their role as co-producers of outcomes. Finally, I think this is the last slide, yeah. So I just want people to get a bit excited again about digital and tech. I came into this world of digital government by stumbling across what was the very emergent community in the UK of civic technologists based around the organisation My Society, this is in the mid-2000s, and there was huge hope there, possibly, maybe some of it is misguided, but there was hope that there was potential for a combination of technology and design to enrich public services and democracy. That was the thing that got me excited, that was what led me to end up working at the government digital service. I think we’ve lost something on the way though. I think it’s partly become about efficiency, but like I said, efficiency is a trap because we’re looking for efficiencies in the wrong place. It’s also sometimes become a bit reductive and a bit utilitarian, just helping people get the outcome they want. I think, as I’ve said, I think public services are about more than that. I worry that rather than saving us from bureaucracy, digital’s almost displaced it, anatomised it, and paper forms have become web versions, and we’re not really thinking about the shape and type of public services that we want. So I really hope that people can get excited about the possibility of technology and design and democracy again, and working out the sometimes messy overlap between those three things. And I do think it matters more than ever, the excitement around AI, particularly from politicians. I think the things that what’s happening in the US is throwing up around accountability and trust in digital public services is going to play into the mix as well. It really matters. So now’s the time to really think about not throwing away what we’ve done in the past, absolutely. User needs, user needs, user design in the public sector is critical, and loads of brilliant work has been done, but now’s the time to expand on that and really reforge what we think public design should be, could be. And with that, I’ll say thank you.

Jeroen Schalk: Thank you for listening to this audio recording. This keynote was delivered at the User Needs First International Conference 2025 in Amsterdam. You can find all about this conference on gebruikercentraal.nl slash recap 2025. Don’t forget to subscribe to this podcast through Spotify, Apple Podcasts, or your other podcast app of choice. You want to learn more about the User Needs First community? Check out international.gebruikercentraal.nl or for our Dutch listeners, gebruikercentraal.nl. Till next time.