Choose a decision with a real tradeoff
Pick one thing you built, removed, delayed or refused. Write down the original constraint and an alternative you considered. ‘We delayed the dashboard because the underlying records were inconsistent’ gives an interview something to explore. ‘We are revolutionizing analytics’ gives it little source material.
Prepare facts, observations and open questions separately
Before speaking, distinguish what shipped, what you personally observed and what remains untested. If a number belongs in the story, bring its source, date range and meaning. A few informal conversations are not a representative survey. Remove customer names, internal credentials and material you do not have permission to share before entering the interview.
Invite questions about why the alternative lost
Explain who faced the problem, what the team tried first and what made that approach unsuitable. A productive follow-up clarifies the decision: ‘What could a user not decide yet?’ adds more than repeatedly describing an incidental word. If the conversation drifts, restate the purpose and the decision you want the reader to understand.
Preserve uncertainty during evidence review
Check who did what and in which order. Correct a draft claim that turns ‘we think this may help’ into ‘we proved this works.’ Exclude sensitive or unsupported details. Keep the difference between a feature being available and a customer achieving a result. Approve the evidence you want used rather than accepting every generated statement.
Give each account a reader-specific angle
Choose up to 6 generation accounts. For founders on LinkedIn, focus on the tradeoff and lesson. For a product newsletter, explain the change, how to use it and its current limits. Set those audiences and purposes in the account settings. The facts stay consistent; the level of detail and reader’s next step can change.
Close with a testable next step
A useful ending can invite the reader to inspect a prepared example, try a documented workflow or answer a precise question. Avoid promising revenue, adoption or time savings you have not measured. Recheck the product’s current behavior and availability before publishing an announcement; a convincing draft is not a release checklist.
Example and takeaway
Illustrative case: a founder decides to show a plain status label before building a detailed dashboard. A LinkedIn angle explains the rejected alternative; a newsletter angle shows where readers can find the label and what it does not yet show. This fictional example reports no customer outcome or growth result.
The useful story is the decision and its evidence. The product can be the subject without every paragraph becoming a sales claim.
Explore next
Try correcting evidence in the example · All practical guides