Product discovery helps a team understand a problem, the people affected, and the available ways to improve their situation before committing to a solution. The 13 phases below form a practical checklist synthesized from established discovery guidance—not a universal or officially standardized sequence. Treat them as evidence-led activities that can loop as you learn, not as a rigid stage gate.
What product discovery is—and what it is not
Discovery is the work of reducing uncertainty about a user problem and whether there is a worthwhile way to address it. GOV.UK’s Service Manual advises: “You should not start building your service in discovery.” Its guidance is written for government services, but the principle also helps commercial product teams avoid mistaking a requested feature for a proven need.
Discovery should consider user goals and context, policy or business intent, technical and legal constraints, existing services, and possible improvements. It may lead to a new product or feature, a non-software change, further investigation, or a decision to stop.
The 13 product discovery phases
1. Set a discovery goal
State the decision the team needs to make and the uncertainty that could change it. For example: “We need to determine why customers abandon account setup and whether a change to the process could reduce avoidable drop-off.” A clear goal keeps the work scoped and gives the team a way to recognize when it has enough evidence to decide. GOV.UK’s discovery guidance recommends setting a clear goal.
2. Reframe the request as a problem
When someone proposes a predetermined feature or service, ask what underlying user difficulty it is intended to solve. Record the request, but do not treat the proposed solution as proof that the problem exists or that the solution is right. Research the problem before choosing what to build.
3. Make assumptions and boundaries visible
Write down what the team believes, what it does not know, and what falls outside the current problem. Turn important assumptions into questions that can be investigated. This prevents opinions or internal preferences from quietly becoming requirements.
4. Identify likely users and their context
Establish who is trying to accomplish what, how they do it today, and what surrounds the product or service. Look at the wider journey rather than only the point where a proposed feature would appear. A user’s circumstances, other tools, and prior or subsequent steps can change what a useful intervention looks like.
5. Plan research questions
Translate the most consequential unknowns into focused research questions. Early questions can be broad, then become more specific as evidence develops. For example, first learn where a process breaks down; then investigate what causes a particular failure or which users encounter it most.
Free tools Windows power users keep installed
One-click scans. No signup required.
6. Choose participants and methods
Select methods that can answer the priority questions with credible evidence and reasonable time, effort, and cost. Plan to hear from a broad range of users, including disabled people and people with low digital skills, as GOV.UK user-research guidance advises. A method is only useful if the people whose experiences matter can take part and the evidence it produces suits the decision.
When deciding among research activities, compare the question each can answer, the strength of evidence, participant access, setup time, cost, and whether it reveals the whole journey. Choose the most efficient activity that still gives a reliable answer; convenience alone is not a sound reason to use weak evidence.
Rank #3
7. Research the end-to-end journey
Learn how people currently reach their goal, including interactions with tools, transactions, support, and offline steps. A problem that appears to be a product issue may originate elsewhere in the journey, or require coordination across several services. Research current behavior rather than relying only on what people say they might do.
8. Map constraints and related work
Check the technology, legislation, policy intent, dependencies, existing services, and organizations involved. Find out whether relevant data or work can be reused and whether another service already addresses the need. This can uncover both limits and opportunities, while reducing duplicated effort. The GOV.UK Service Standard’s guidance on understanding the whole problem stresses looking beyond the immediate service.
9. Identify opportunities and alternatives
Explore possible interventions before settling on software. Depending on the evidence, a useful change might involve clearer content, better information, a partnership, data sharing, a policy or process adjustment, or a product feature. Compare alternatives against user needs, constraints, and the outcome the team wants to improve.
10. Estimate the problem’s scale and cost
Establish a baseline for the problem where possible: for example, time spent on a process or the cost of carrying it out. Record how the estimate was derived and what it includes. A baseline gives later work something concrete to compare against, instead of relying on an impression that the solution “seems better.”
11. Define success measures
Decide what evidence would show that people are better able to achieve their goals. Connect measures to the user outcome and the baseline, and identify what data would be needed to assess change. Avoid choosing a convenient metric that does not indicate whether the underlying problem improved. GOV.UK’s measuring-success guidance links performance measures to whether users can complete their goals.
12. Synthesize and share findings
Bring the evidence together into a clear account of the problem, user context, constraints, remaining uncertainty, and promising opportunities. Distinguish what the team observed from what it infers. Share learning with the people and teams who need it unless confidentiality prevents that; discovery findings can inform work beyond the immediate team.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Best Value
13. Decide what happens next
Use the findings to choose whether to proceed, investigate further, pursue an alternative, or stop. Continue only when there is a viable opportunity worth the likely cost and effort. If the team proceeds, identify ideas to test and an initial measure to carry into alpha or the next development phase. GOV.UK puts the point plainly: “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.”
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How long should product discovery take?
There is no fixed duration. GOV.UK says around four to eight weeks is typical for government-service discovery, while emphasizing that timing depends on the work. That range is contextual guidance, not a measured industry benchmark or a universal deadline for commercial products. The right duration depends on the decision, the uncertainty, the evidence already available, and how difficult it is to reach relevant users or assess constraints.
Who should be involved?
GOV.UK’s service-team guidance identifies product management, user research, and service design skills for discovery. Depending on the problem, content design, development, performance analysis, business analysis, or cyber security expertise may also be useful. The GOV.UK service-team guidance does not imply every team needs the same staffing configuration; involve the skills needed to understand the users, evidence, and constraints in your particular case.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
