Democratic technical debt
Working on delivering public services can be a patrician process. Even with researchers, policy people and ministers from a democratically elected government, you can end up missing a layer of democracy, and you’re deciding whats best for citizens in an artificial environment. Sure, there’s research to check whether you’ve built the right thing or, if you’re lucky, to suss out the design constraints of the project before you start, and all research is better than no research. But often, there are institutional barriers to research that could challenge a policy direction, or create a project that costs too much. This isn’t to say that those things aren’t valid reasons, but to say that the way that we understand qualitative research as a source of popular authority then rubs up against the democratic design constraints that are also meant to serve the interests of The People.
As an industry, we’ve created a pattern for working and building that is meant to be practical, replicable, scalable and produce results. Most of the time these results are good (yay for better online services, more open data from government, permanent secretaries on twitter), but as we scale up and these processes become more about how normal government runs rather than the bunker of nerds, we need to think about how to make ourselves responsive to change.
In the efforts put in to deliver services (good), there is often a neglect of the linking infrastructure that holds the idea of “government” (small g) together.
User research in the public interest
If you don’t consider what the long term impact of how you build something is, you’re accruing democratic technical debt. We’re already seeing something of that now in the way that we deal with user research. In the 30s through to the 60s, Britain had a semi-state way of understanding its populace. It used research techniques coming out of anthropology, psychology and sociology to build a national picture of Britain. Today, we’re insisting the social research plays a part in delivering government services in an adaptive, relevant and agile way that’s been out of fashion in the era of the focus group.
However, there are two problems here: firstly, the knowledge generated is ephemeral and disappears; and secondly the design constraints mean that the questions we’re seeking to answer are limited.
On the first point: there is no publication of user research in government beyond, potentially, a blog post with insufficient detail. Let me repeat that: the government spends money researching how a service could help, how to build it, how to improve it and some deeper life stuff about the needs of the service users and it will inevitably languish on a google drive somewhere rather than being shared (anonymised) either with other teams, academics or the public. Qualitative data is still data, and more data about the practice of government should be open.
I think this an omission rather than a design, but the benefits of sharing at the very least for design purposes should be obvious.
I think that the issues for withholding data boil down to essentially how easy it could be to de-anonymise data (fair question), who would you hire to be the archivist for this and where do they sit (easy question), and how do you deal with qualitative research that contradicts a policy or, say, a democratic vote (no idea, but we should be talking about this).
Point two follows on: if we’re only finding out data about the service and then never sharing it, we lose the ability to be strategic. We are always tactically building services for immediate need, rather than going all #LongTermServiceDesignPlan. Would we prefer to always build a service for people falling between the cracks, or would we like to identify the cracks an fix them?
As short-term, small scale research becomes embedded and more parts of government follow a service design informed approach, we need to add a zeroth law of designing public services, above “what is the user need?”. Something along the lines of:
Does this work improve the structure of the democratic state for users?
We need to acknowledge that:
- All the choices we make are political
- None of the interpretations we make are objective
And that’s ok, let’s be ok with making subjective judgements that we can defend.
I started by talking about how public servants act in a way that is “in the best interests of citizens”, and this needs to be something that we make explicit as a part of the process. Not just thinking about “does my service meet user needs”, but adding the service into the context of government.
The internet nerds building the rules for the way that government is going to work from now on. And that leaves us with a responsibility to think about the governance of that, and how we serve the public, and how that balances with democracy, and to acknowledge that no solution is going to be perfect, but through our inaction we allow the public to come to harm.