Frame the choice and its constraints
Write one question that could change what you do next. Identify the user, the job and the constraint. Should the next reporting experiment test accuracy checks or approval ownership is more useful than asking for a general roadmap.
Specify what counts as an observation in the experiment. A positive reaction to a mockup may mean interest; it does not establish adoption or willingness to pay.
Trace claims back to the source
Select the customer or market evidence for the question. Inspect attribution, date, segment and permission. In Product Decisions, review the evidence pack and complete any missing trace before treating the evaluation as decision-ready.
Separate observed behavior, a participant's requested solution and your explanation. Include an account that does not fit the preferred explanation.
Copyable working template
Claim:
Source and segment:
Observed behavior or quotation:
Interpretation:
Contrary evidence:
What is still unknown:Make alternatives solve distinct problems
Define at least two plausible actions with their costs and limitations. Include a smaller experiment or a decision to defer when it would provide better information.
Synthetic example: a review checklist may address report accuracy; an approval step may address responsibility. Comparing them as simply small versus large misses the difference in what they test.
- Checklist experiment
- Observe whether a user can find the figure they need to verify.
- Approval experiment
- Observe whether the named approver can review the report before sharing.
- Reason to defer
- No evidence yet shows that the delay affects the user's outcome.
- Information needed
- A recent example of the workflow and the consequence of its failure.
Ask for the strongest objection
Review the critique before accepting a recommendation. Ask which evidence contradicts it, which assumption carries the choice and whether the proposed experiment could distinguish the options.
A decision can be sensible under uncertainty. Record that uncertainty so the next person reviewing the result does not mistake the choice for a proven market fact.
Make and retain the founder decision
Approve, revise or defer through the required human decision. Document the chosen action, its reason and the observation that would change your view. Revisit the record when new evidence arrives.
AI can help assemble and challenge the reasoning. The founder owns the product choice. The example here is synthetic, with no live evaluation, commercial commitment or measured success implied.
Copyable working template
Decision and date:
Selected option and reason:
Evidence used:
Uncertainty accepted:
Next experiment and owner:
Revisit when:Access and review
Request access and complete approval, invitation and verified login. Applying does not start billing. Free covers limited work in Customer Discovery, Product Decisions and Marketing; current Free and early access use BYOK, billed separately by your provider. Final Free AI conditions and limits are still being confirmed. Founder — US$19/month, with access to Oqera and AI usage included within the plan’s limits — is preparing for launch. Your own API key is not required for that included usage. Founder offers all nine teams; Coding requires supported setup and separate approval. Paid checkout is not yet enabled. Trial terms for the updated offer are not yet confirmed. Separately paid third-party services are not included. The Apps catalog is available; connections and external actions require separate setup, permission and review.
