Home / Discovery

Finding the feature worth making.

Discovery is the cheapest part of product development and the part most often skipped. This is what it involves, and what it produces.

Four risks we look for

Every product idea carries four kinds of risk. Value: will people choose to use it? Usability: can they work out how? Feasibility: can it be built with the time, skills and systems available? Viability: does it work for the business, including cost, support load and any legal constraints? Discovery tries to identify which of these is largest for your idea and to test that one first.

What we actually do

  • Problem interviews. Conversations about recent, real events. "Tell me about the last time you had to rebook a delivery" teaches more than "Would you use an app for rebooking?"
  • Workflow mapping. For internal tools, we sit with staff and map the steps, hand-offs and workarounds as they really are.
  • Sketch testing. Rough clickable screens, put in front of intended users with a task to complete. We watch, and we do not explain.
  • Technical spikes. Short, disposable tests of the riskiest integration or data question.
  • Assumption log. A living list of what must be true, how sure we are, and what would change our mind. You can try a version on the tools page.

What you receive at the end

  • A product brief of one to two pages
  • Tested sketches of the primary journeys
  • A first-release scope with a cost range and its assumptions
  • A list of known risks and how we propose to handle them
  • A clear recommendation: build, change course, or stop

When discovery says stop

This happens, and it is not a failure. Finding out in four weeks that customers will not change their habit is far better than finding out after a six-month build. We will present the evidence plainly and, where we can, suggest what else might work.

What discovery cannot do

It cannot guarantee success. It reduces uncertainty; it does not remove it. Some answers only arrive when real people use real software in their own context, which is why our releases are small and our learning continues after launch.

Who should take part from your side?

One decision maker, one person who knows the day-to-day work, and access to five or more intended users. More people are welcome but not required.

Can we skip it?

If the problem, users and scope are already well evidenced, discovery can be short. We will look at what you have and tell you honestly.

Do you need our customer data?

Rarely, and never at the start. Discovery works with conversations and sketches. If data is needed later, we agree a lawful, minimal approach first.

Is the output useful if another team builds it?

Yes. The brief, sketches and scope are written to be handed to any competent developer.