What a good discovery phase should produce

The GOV.UK Service Manual is clear on the purpose of discovery: "Before you commit to building a service, you need to understand the problem that needs to be solved." It is equally clear on what discovery is not: "You should not start building your service in discovery." The practical question for any team is what a discovery should leave behind.

Judge it by the decision

Discovery ends with a decision on whether to proceed. The Service Manual notes that "it's not a failure to stop at the end of the discovery phase if your research shows that's the best thing to do." The test of a good discovery, then, is whether a senior responsible owner could make a sound funding decision from its outputs alone.

The outputs

• a problem statement, reframed from the original assumptions where the research shows it should be
• the users and their needs, based on research with real users rather than assumptions about them
• the constraints: legislation, policy intent, security classification, existing processes and contracts
• a map of the data and systems involved, including what each actually does in practice
• the riskiest assumptions, ready to be tested in alpha
• how success will be measured
• an indicative approach, team and cost, with the uncertainty stated plainly

Common pitfalls

• restating the original brief rather than testing it
• choosing the solution before the research is done
• ending with findings that describe the situation but support no decision
• running for longer than the questions require; the Service Manual suggests around four to eight weeks is typical, and that the purpose should dictate the length

What comes next

If the decision is to proceed, the outputs carry straight into alpha, where the Service Manual describes "identifying your riskiest assumptions and testing them" as a crucial part of the work. A discovery that has already named those assumptions gives alpha a clear starting point.

brytloop runs discovery as a short, fixed-scope phase that ends in a decision. The result is a document that supports real choices about funding, scope and approach, rather than a set of general recommendations.