Choose one sheet, not all six

If you are stuck on Use Finish with
Whether the problem is realProblem evidence logA testable problem statement and the next missing fact.
What belongs in version oneMVP scope sheetOne complete flow with explicit exclusions.
Which technology to commit toStack risk registerA choice tied to maintenance and failure recovery.
How to chargePricing decision sheetA value metric, package boundary and test.
How to learn from early outreachBuyer conversation ledgerA record of fit, timing, objections and next steps.
Why users do not returnRetention traceThe earliest break between signup and repeated value.

1. Problem evidence log

Use this before treating an idea as a product opportunity. Add one row for each real incident, not one row for each person who says the idea sounds useful.

FieldWhat to record
Date and triggerWhen the workflow started and what created the need.
Person and responsibilityWho did the work and who was accountable for the result.
Current stepsThe tools, handoffs, files and checks used today.
Failure or frictionWhat was slow, confusing, risky or repeatedly corrected.
ConsequenceWhat changed because of the problem: delay, extra work, missed follow-up, rework or uncertainty.
Buying authorityWho could approve a change, and what else would influence the decision.
UnknownThe most important fact still based on assumption.

Decision prompt: “For [buyer role], [trigger] creates [specific workflow problem]. They currently use [workaround], which leads to [observable consequence]. We still need to learn [unknown].”

Next move: If incidents do not share a buyer, trigger or consequence, collect more evidence. If they do, choose a test from the validation hub.

2. MVP scope sheet

Use this when the backlog contains several reasonable ideas but the first useful release is still hard to describe.

Write the core flow in one line

“When [trigger], the user provides [minimum input], the product helps them [critical action], and they leave with [useful result].”

Part of the flowDefine itBoundary question
EntryWhat the user must already know, have or upload.Can setup be handled manually for the first pilot?
Core actionThe smallest transformation from input to result.Which steps need judgment rather than automation?
ResultWhat the user can review, send, decide or complete.Can the user tell whether the result is correct?
Risky stateBad data, missing permission, failed payment or unavailable integration.What must be prevented, explained or recovered?
Support fallbackWhat the operator can inspect or repair.Can you help without editing production data blindly?
Explicitly outAdjacent jobs, custom requests and edge cases postponed for now.Would removing this break the core result?

Decision prompt: Keep an item only if removing it prevents the first useful result, creates unacceptable risk or makes support impossible. Move the rest outside the pilot boundary. The MVP hub has deeper guides for each part.

3. Stack risk register

Use this after the workflow is clear. Comparing feature lists too early rewards novelty and hides the cost of operating the product.

ComponentQuestions to answer before choosing
Customer dataWhere does it live? How is access separated? Can customers export or delete it?
AuthenticationWho manages identity, password recovery, organisations and permissions?
Critical integrationWhat does the user see when it is slow or unavailable? Can the job be retried safely?
AutomationWhere are runs logged? Who receives failure alerts? Which steps must be idempotent?
BillingWhich states affect access? How are failed payments, refunds and plan changes handled?
AI or variable-cost serviceWhat is the cost unit? How is poor output reviewed? What fallback exists?
Deployment and backupWho can restore the service, and what has actually been backed up?
Exit pathWhat would be difficult to migrate if cost, policy or product needs changed?

Decision prompt: For each realistic option, record the likely failure, who can recover it and what switching away would require. Prefer the choice whose risks you can explain and operate. Continue in the stack hub.

4. Pricing decision sheet

Use this when the price came from a competitor page, a round number or a fear of asking too much.

Decision prompt: “The price changes with [value unit] because [reason]. The base package includes [complete outcome]. The next package is for [different operational need], not merely more buttons.”

Test: Ask a buyer to explain the offer back to you before defending it. If the unit or boundary is confusing, revise the package with the pricing hub.

5. Buyer conversation ledger

Use this for outreach, discovery calls and demos. It keeps a quiet “not now” separate from a true mismatch and stops useful objections from disappearing into an inbox.

FieldRecord
Reason for contactThe sourced situation that made the buyer relevant. Do not write a generic persona label.
Problem languageThe buyer’s own words for the job, friction and consequence.
Current approachTools, people and workarounds already in place.
TimingWhy the issue matters now, later or not at all.
ObjectionFit, trust, price, switching effort, authority, priority or another specific barrier.
Product evidenceWhat the conversation confirms, weakens or leaves unknown.
Next stepA dated action agreed by both sides, or a clear closed outcome.

Weekly review: Group conversations by trigger and objection. Change the message when relevant buyers misunderstand it; change the target when the problem is not important to the people reached. The acquisition hub helps choose the next channel or conversation.

6. Retention trace

Use this when signups, usage or cancellations do not explain why value fails to repeat. Trace individual accounts before averaging them together.

MomentEvidence to inspectQuestion
PromiseLanding page, sales notes, plan selected.What did the customer expect to achieve?
SetupInvites, permissions, imports, integrations.What had to be true before useful work could begin?
First resultCompleted workflow, saved output, sent report or resolved task.Did the customer reach a result they could judge?
Repeat useSecond workflow, return interval, shared output.Was there a natural reason to come back?
FrictionErrors, abandoned steps, support requests, manual intervention.Where did confidence or momentum break?
OutcomeContinued use, downgrade, cancellation or silent inactivity.What happened after the break, and what replaced the product?

Decision prompt: Compare a successful trace with a stalled one. Fix the earliest difference that the product can reasonably influence. Do not use a late reminder to cover an early setup failure. Continue with the growth hub.