Feedback arrives everywhere: support tickets, sales calls, reviews, messages to the founder. Without a routine, the roadmap follows whoever asked last. These four steps put feedback in one place, group it by the underlying problem, and decide with clear criteria — so customers and the team can see why something is or is not being built.
Step 1 — Collect in one place
Goal: Stop feedback getting lost across channels.
- List every channel where customers give feedback today.
- Choose one shared place to record it, with a short standard form.
- Record who asked, what they asked for, and what they were trying to do.
- Make one person responsible for keeping it tidy each week.
You finish with: A single feedback log fed from every channel.
Step 2 — Find the problem behind the request
Goal: Group requests by the job customers are trying to get done.
- Rewrite each request as the problem it solves for the customer.
- Group items that share the same underlying problem.
- Count how many customers and which segments each problem affects.
- Follow up with a few customers per group to confirm the problem.
You finish with: A short list of confirmed customer problems with evidence.
Step 3 — Decide with written criteria
Goal: Choose what to build using reasons you can repeat.
- Agree criteria in advance: customer impact, fit with strategy, effort and risk.
- Score each problem against the criteria with the team.
- Pick a small number for the next cycle and name what is deliberately not chosen.
- Record the reason for each decision.
You finish with: A ranked set of problems and a short decision record.
Step 4 — Close the loop
Goal: Show customers and the team what happened to their input.
- Tell customers when a problem they raised is being worked on, or why it is not.
- Share the decision record with sales and support so they answer consistently.
- Check after release whether the problem is actually solved for those customers.
- Review the log and criteria every cycle.
You finish with: Customers informed, a consistent answer for the team, and a checked result.
Common pitfalls
- Building the exact feature requested instead of solving the problem.
- Counting requests without asking who they come from.
- Changing criteria to justify a decision already made.
- Never telling customers what happened to their feedback.