Why we re-thought discoveries

In August 2021 I joined a team at the Department for Education tasked with uncovering improvements in teacher training. These opportunities would then be addressed by either policy changes or digital solutions.

This kind of project would typically be a discovery as described by the Government Digital Service (GDS). Discoveries are projects that aim to explore a broad problem space to better understand it and to suggest potential solutions. They involve a multi-disciplinary team and typically last 8–12 weeks.

This is a story of why a discovery was not the right approach for the task and why a continuous discovery was.

Why not to do a discovery

I’ve seen discoveries save millions, because this is the phase where you make sure you “build the right thing”. As a user researcher, this is the most exciting phase, because you get to observe and ask open ended questions.

However, our policy colleagues’ experience was not great. For them discoveries:

• delivered large slide decks that proposed what they already knew
• made recommendations that were impossible to deliver and ignored political context
• lacked a shared language and ways of working

Discovery findings went into a policy black hole and did not shape policy.

It almost hurt me to realise how inadequate discoveries have been for our colleagues’ needs. It was humbling.

These were very good reasons for us not to follow the script and try to build a different way of working. No huge slide decks, no moonshot recommendations, no confusing language.

The problem you are solving is too big and too complex

There were other reasons not to do a discovery.

If you work in government the kind of problems you might look at solving exist within complex, interconnected systems — crime, education, health, welfare, environment, etc. These are ‘wicked’ problems that took years to form and will take years to untangle. They take a long time to properly understand.

If you’re working on such a problem, chances are, you won’t understand it well enough in the time allocated to a discovery. You might feel the pressure to have some ideas ready at the end of a discovery, so you can test them in the next phase — alpha. This approach does not work for wicked problems because of how they interconnect and influence each other. If you rush, you will probably just pass the real problem to another team, another programme or another department.

As an example, we know that one of the burdens on teachers is paperwork. Instead of joining up internal departmental processes to reduce the amount of communications sent to teachers, which is really hard, a temptation might be to come up with a workload reduction toolkit for teachers as part of a discovery. This would be easier to implement, but would invariably lead to more paperwork and not solve the real problem.

Not enough time to build trust

As part of understanding how the problem came to be and how it continues to be, you also need to understand the formal and informal power structures. This is the stuff that you won’t find on an org chart. Who are your allies? Where are the roadblocks? Who really has the power to change things?

In one of my projects we set out to change the teacher training providers’ contract (easy, once you get through the legalese), but discovered that the contract was fine. It was the behaviours of the people who enforced the contract that we had to change instead (much more difficult!).

To understand what really happens in an organisation, and who pulls what levers, you’ll need to listen, understand, follow through, be open and work through tricky situations. Trust and psychological safety are prerequisites for effective working.

It’s already been tried

You might not know this when you start, but your colleagues may have already lived through a similar discovery.

In one of my previous discoveries, a data division colleague was openly obstructive in one of our early show and tells. We were really taken aback. Later we learned that he’d already been through 2 other discoveries on the same topic. None of them sufficiently built on, in his opinion, the wealth of data that already existed. So, it was his way of expressing frustration to us, the third team working on this.

This discovery ended early so our team could be re-assigned to COVID-19 response programme, but when the pandemic ended, a similar discovery was comissioned again. It’s costly to write things down, but it’s even more costly to forget them (Giles Turnbull gives some excellent examples, see how teams remember).

If not a discovery, then what?

In re-thinking discoveries the biggest opportunity was to bring policy and ‘digital delivery’ closer together. We were imagining a world where future policies were designed with the understanding of real people’s lives, tested and iterated with people’s input.

So we looked at the learnings from the following:

• experience of doing discoveries in government (including Will Myddleton, Kate Tarling and John Waterworth)
• policy colleagues’ reflections on what failed when working with ‘digital’ teams
• hopes and fears of the programme leadership, policy colleagues and us, the team doing the work

We became known as the Explore team who created and iterated a new continuous discovery way of working within DfE’s CPD programme. We:

• spent the time to connect the dots and understand complexities together
• designed our comms to bring our colleagues on the learning journey with us
• involved our colleagues in our activities from user research to co-creation

Better decisions

The impact of this way of working was that:

• policy colleagues made more user-centred policy decisions because they were grounded in the lives of real people
• we made more realistic design decisions because we understood and worked within the constraints (including ministerial pressures)

We realised that sharing insight and modern ‘digital’ ways of working was more impactful than building digital solutions (although we also built a digital solution along the way).

In retrospect…

I really believe that the continuous discovery approach extends the original GDS discovery and makes it more appropriate for complex problems. It could help teams to:

• appreciate that complex problems are interconnected
• consider not only ‘user needs’ but all actors’ needs
• reimagine future states and a pathway to transition towards them

For the past 2 years we’ve been busy creating and applying this new way of working. We were hugely influenced by Terry Irwin’s work on Transition design and we also often referred to the Systems thinking toolkit for civil servants.

If you have your own discovery story, especially working in complex problem spaces, please share how you worked and what you learned.

With thanks to…

• Isabelle Andrews, the product manager on this team who helped me see complexity, then possibly regretted it as I went down the systems thinking rabbit hole
• Mark Avery, the content designer on this team who helped me write better and for “two eyeing” this piece
• many DfE colleagues who took part in this project and made this way of working possible