User centred design and the rule of law
Starting with GOV.UK in 2012, the changes to how the UK government designs and builds digital services over the past decade have been significant. User centered design and agile development practice are now the norm, at least in some form. User centred design means digital services are tested with real users throughout their development to check they meet their needs. Agile means services are launched early and iterated based on learnings. Design standards, assessments and in-house digital skills help set a high baseline.
Long gone are the days when ministers would launch a service on the Today Programme only for it to crash 20 minutes later, and the basic usability of digital services rarely makes headlines, as it did for example when DWP launched the fated big-bang launch of Universal Jobmatch in 2012 which did not work for users and was full of security and privacy issues and damaged public trust in digital. The group may also be aware of the initial attempts to deliver UC before the adoption of these practices.
The introduction of user centred design to government was a necessary and, it now seems, inevitable adoption of industry best practice. But it took a while. The idea that digital systems should be designed to meet the needs of users, not the designers of that system was an old one by the 2010s. User-centred design has it’s origins in the 1970’s when computer-human interactions was an emerging area of study. The British social scientist Enid Mumford developed the needs-fit theory of job satisfaction in 1975 from studying the design of back-office banking systems. She emphasised the importance of involving users in the design process of technical systems.
In 1977 US academic Rob Kling built on Mumford’s work and coined the term User centred design, which was then popularised across the design industry by the book The Design of Everyday Things. These approaches are not new. If you visited any credible digital-age organisation today you’d see them applying user centred design, and that includes UK central government departments. It took over 35 years, but eventually those ideas made their way into government and our digital services are better for it.
It is fair to say though that user centered design tends towards the functional and the utilitarian. The priority is getting the proximate task done and within the bounds of a particular service. It is also individualistic. It deals mainly in the interaction between “the user” and “the service” with as little effort as possible. Externalities remain external. It’s an approach summed by Apple’s Steve Jobs in the phrase: “It just works seamlessly”. Or as we put it at the Government Digital Service when we were creating our design principles “Do the hard work to make it simple”.
One school of thought is that it is enough to apply Jobsian principles to the public sphere – services should ‘just work’, and people should not have to understand anything about the workings. This is probably the prevailing view within digital government practice. The collective instinct is to abstract and simplify, but at the expense of legibility.
Recent work by the Administrative Fairness Lab has highlighted this utilitarian nature. Public officials working on UC prioritise usability, accuracy and efficiency. Whereas welfare rights advisors prioritise that claimants get everything they are entitled to under the law. And claimants prioritise how the relationship between them and government feels - does it feel respectful and fair, for example, when they are appealing a decision.
Now, the UC team should absolutely be designing a system that is usable, accurate and efficient. It’s an exceptional digital team, probably the best in government. And the UC design includes some groundbreaking features like a journal people can use to prove their interactions with government that I wish other services would copy. But is it possible to accommodate the quality of relationship between claimant and government, ensure people are able to understand and and exercise their rights and help civil society hold a complex socio-technical system like UC to account against the law within government design practice? For me, the question is not whether user centred design is necessary for delivering digital public services - it is clear that it is and that it’s here to stay - but whether it is on its own sufficient.
Personally, and I say this as someone who led the initial design work of both GOV.UK and Universal Credit, I don’t think user centred design is, on its own, sufficient. That is something that has been evident for some time, but gets lost in some of the fog created by the debate around digitisation.
Firstly, there is the question of what government decides to design and automate, and what it does not. There is currently little pressure to apply user centred design and automation to things that save users time and cost government money (such as appeals). We need to see an incoming government make the systematic elimination of administrative burden a core part of government digital strategy. Government could use digital to remove learning, compliance and psychological burdens from users and help them exercise their rights.
Secondly, we need to be clear about how digital public services are different when it comes to intelligibility and predictability. Digital services like UC can be highly personalised and the software running them is highly mutable and can change daily or hourly. Campaigners, legal professionals’ and parliamentarians need to be able to operate within the iterative development cycles they can articulate their concerns in a way that digital teams can act on.
This will require the systematic and realtime publication of information about how digital services work and how they are changing so there is a ‘ground truth’ to understand policy gaps, including software code, screenshots of user interfaces and release notes all with links back to legislation and policy.
Finally, it is a characteristic of a democratic society that public services don’t just work. If democracy is government by explanation, digital services need to be designed to explain something of how government works to the users of those services. So it should always be possible for a user to understand which part of government their are interacting with, find the law that relates to the service they are using and identify how to access their rights. They should be able to do so at the point of use, even where it makes designs a little less simple. Rather than seamless design, we should aspire to well designed seams that surface the legal underpinnings of digital services.
At the moment we are designing our public services like an iPad. Functional but good luck if you want to understand the rules it runs on.
User centered design and digital services are here to stay. But the design practice that emerged from 2012 onwards was fundamentally incomplete at inception. As digitisation moves from periphery to the core of how all public services are delivered, a broader coalition needs to reforge and grow what design means in the context of public services. Digital public services need to do more than just work.