Product-market fit — knowing for real

One signal isn't enough. PMF is the AND of these:

1. Retention curves that flatten — cohorts plateau at a non-zero number, not a long decay to zero. SaaS: ≥80% week-4 retention for the target segment. Consumer: ≥40% week-4 D-stickiness for the relevant DAU/MAU bucket.

2. Organic pull — referrals, word-of-mouth, organic search, inbound signups happen without you pushing. If 100% of new customers are from founder outreach, you don't have it yet.

3. Pricing pull — you can raise prices 20-30% and customers stay. "Can I pay you more for X?" is the loudest PMF signal there is.

4. Customer expansion — existing customers buy more (seats, modules, usage). Net revenue retention >100% in B2B SaaS, or repeat purchase in consumer.

When all four are true for one segment, you have PMF for that segment. Three out of four is "early signal, not PMF." Don't fundraise off three.

Sean Ellis's test: ask current users "How would you feel if you could no longer use this product?" with options Very disappointed / Somewhat disappointed / Not disappointed.

≥40% "very disappointed" = decent PMF signal. Below 40% = not yet.

Pair it with: "What type of person would benefit the most from this product?" and "What's the main benefit you receive from this product?" Cluster the answers. The cluster that gives you your 40% IS your ICP — sharper than anything you've written down.

What the test misses:

• It surveys current users; the people who churned aren't there. Run on a recent cohort.
• It says nothing about TAM, growth rate, or unit economics.
• "Very disappointed" from 40% of 200 customers ≠ from 40% of 50 customers. Sample size matters.

Use as one signal among the four above. A 40% score plus flat retention curves plus organic pull = real.

What changes:

• Distribution becomes the bottleneck. Building stops being the constraint.
• Pricing becomes a strategic lever (you can move it).
• Hiring profile shifts: more sellers, more support, more ops; fewer "find PMF for us" generalists.
• Fundraising terms improve dramatically (and become a distraction).

What doesn't change:

• The bugs still hurt. The team is still small.
• The founder is still mostly selling and supporting in year 1-2 post-PMF.
• Most product instincts are still wrong; customer conversations still teach more than the data.
• You can lose PMF if the segment evolves and you don't.

The "PMF for one segment, not all" case — most common. Acknowledge it; don't paper over it. Double down on the segment that's working. Don't dilute the message trying to fit a second segment until segment #1 is winning hands-down.