Milestone map
Milestone map
3 milestones
Define your target PMF signal
1–2 weeks (definition) + ongoing measurement
Define what product-market fit means for your business before measuring it. 'You know it when you have it' isn't actionable. Common signals are Sean Ellis's 40% 'very disappointed' test and NPS above 50. Others are flattening retention curves or a category retention benchmark. The right signal depends on your business type. Defining it first prevents rationalising afterwards.
Proof required
Write a PMF hypothesis document. Name your target signal and why it fits your business model. Give its current measurement, even if below target. List the leading indicators you watch that might predict it.
What gets checked
- A named quantitative signal with a target threshold. 'When it feels right' doesn't count.
- The current measurement is shared honestly. Being below target is valuable data.
- Leading indicators you watch day to day are named.
Common mistakes
- Picking a signal you've already hit. That's rationalisation, not a hypothesis.
- No current measurement. The hypothesis then has no evidence.
- A qualitative-only signal, like 'customers say they love it'.
Resources
Foundationstart here
Depthgo deeper
Arena
What a verifier looks for
- Ask when the signal was defined, and whether that came before the measurement.
- Claiming PMF here without evidence suggests rationalisation.
- Ask which metric they check each morning to track progress toward PMF.
You'll sign in first, then come straight back here.
Cross your PMF threshold
4–12 weeks (depending on measurement cycle)
Show your Milestone 1 signal crossing its threshold. PMF isn't permanent; it's a signal for a specific cohort at a specific time. State which cohort you measured and when. Say whether the signal held across at least 2 measurement periods.
Proof required
Share the data showing the threshold crossed, like survey results or a retention curve. State the cohort measured and when. Say whether you've measured twice and both results cleared the threshold.
What gets checked
- The actual data is shown, not a description of it.
- The cohort is specific. A recent cohort is stronger evidence than all users ever.
- Two measurement periods are covered if available. One result is preliminary.
Common mistakes
- Changing the signal after missing the original threshold. That isn't PMF.
- Measuring all users ever. It mixes engaged early users with disengaged recent ones.
- Only one measurement. Two results above threshold is the minimum credible claim.
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- The data must match the Milestone 1 signal. If it changed, ask why.
- Ask whether they measured all users or a recent cohort.
- Ask for both measurement dates. With one, ask when the next is planned.
You'll sign in first, then come straight back here.
Name segments with and without PMF
2–4 weeks (analysis and write-up)
Identify the segment that has PMF and the segment that doesn't. PMF usually concentrates in one customer type, use case or geography. Naming both tells you where to focus acquisition and where not to. It shows you understand what you found.
Proof required
Write a segmentation analysis. Name the segment with PMF by customer type, use case or observable trait. Name the segment without it and say why. Say which original product assumption the second segment disproved.
What gets checked
- The PMF segment is specific enough to target, not 'engaged users'.
- The non-PMF segment is a real customer type where the product genuinely underperforms.
- At least one disproved product assumption is named.
Common mistakes
- Claiming PMF with every customer. That's almost never true.
- Describing the segment by behaviour, like 'regular users'. Name a customer type.
- Saying they 'weren't the right fit' without naming the wrong assumption.
Powstik Guide
A segment analysis naming where you have PMF and where you don't.
Steps
- Split users by type, use case or trait.
- Compare retention or survey scores per segment.
- Name the segment with PMF and the one without.
- Say what you'll stop doing.
Template
| Segment | Users | Retention or score | PMF? | |---------|-------|--------------------|------| Strongest segment: Weakest segment: What I'll stop doing:
What gets sent back
- Segments with no data.
- Only one segment named.
- No decision.
Resources
Foundationstart here
Depthgo deeper
What a verifier looks for
- Ask how they'd find 10 more customers exactly like the PMF segment.
- Ask whether they have current customers in the non-PMF segment.
- Ask whether the disproved assumption was in their original pitch or founding hypothesis.
You'll sign in first, then come straight back here.
Part of