Product-Market Fit
Does demand pull product out of your hands?
Make something people want — Paul Graham
Perceive
Product-market fit is the point where the market pulls product out of your hands faster than you can make it. Before PMF, you're pushing. After PMF, you're steering.
Marc Andreessen: "You can always feel product-market fit when it is happening. The customers are buying the product just as fast as you can make it. Money is piling up in your company checking account."
You can also feel when it's not happening. Everything is hard. Sales take forever. Churn is high. Features don't move metrics. You're explaining instead of delivering.
Before and After
| Before PMF | After PMF | |
|---|---|---|
| Sales | Push — convince, explain, discount | Pull — inbound, referral, waitlist |
| Churn | High — customers leave after trying | Low — customers stay and expand |
| Growth | Paid, manual, exhausting | Organic, viral, compounding |
| Feedback | "It's interesting" | "When can I get more?" |
| Team | Demoralized by lack of traction | Overwhelmed by demand |
What PMF Is Not
| Common Mistake | Reality |
|---|---|
| Lots of signups | Signups without retention is a leaky bucket |
| Press coverage | Awareness is not demand |
| Investor interest | VCs buy narrative, not PMF |
| Feature requests | Requests without willingness to pay is wish-listing |
| Revenue from one big client | Dependency is not fit |
Question
Why do most startups die?
Not from building the wrong thing. From building the right thing for the wrong market. Or the right thing at the wrong time. Or a good thing that nobody needs badly enough to pay for.
The Struggle Before Fit
PMF is not a binary switch. It's a gradient. Most businesses live in the zone between "nobody wants this" and "everyone wants this" — the uncomfortable middle where some signals are positive but nothing compounds.
The danger: optimizing prematurely. Building features for users who don't represent the real market. Scaling sales before the product retains. Raising money to buy growth that doesn't stick.
Sean Ellis Test
Ask existing users: "How would you feel if you could no longer use this product?"
| Response | What it means |
|---|---|
| Very disappointed (>40%) | PMF achieved |
| Somewhat disappointed (25-40%) | Close but not there |
| Not disappointed (<25%) | No PMF |
Act
Find PMF
| Step | Action | Signal |
|---|---|---|
| 1. Narrow the market | Pick the smallest viable audience | You can name 10 people who need this |
| 2. Solve one pain | Do one thing that eliminates real friction | They use it without being asked |
| 3. Measure retention | Do they come back without prompting? | Week 1 → Week 4 retention > 40% |
| 4. Ask for money | Will they pay? How much? | Willingness to pay without negotiation |
| 5. Check pull | Are they referring others? | NPS > 50, organic growth > 0 |
Measuring Fit
| Metric | No PMF | Weak PMF | Strong PMF |
|---|---|---|---|
| Retention (30-day) | <20% | 20-40% | >40% |
| NPS | <0 | 0-30 | >50 |
| Organic growth | 0% | 1-5% | >10% |
| Sean Ellis score | <25% | 25-40% | >40% |
| Sales cycle | Months | Weeks | Days |
| Churn (monthly) | >10% | 5-10% | <5% |
Checklist
- Can you name 10 people who desperately need this?
- Are users coming back without being reminded?
- Would they be very disappointed if it disappeared?
- Are they paying without negotiation?
- Is any growth organic?
Context
- Unit Economics — Unit economics without PMF is fiction
- Distribution — Distribution amplifies PMF, never replaces it
- Cash Flow Is King — PMF is the fastest path to positive cash flow
- Zero to One — PMF validates that your "one" is real
Links
- Andreessen: The Only Thing That Matters
- Rahul Vohra: How Superhuman Built an Engine to Find PMF
- Paul Graham: Do Things That Don't Scale
Questions
If you stopped all marketing tomorrow, would anyone notice?
- What percentage of your users would be "very disappointed" if your product disappeared?
- Are you iterating toward fit or scaling before finding it?
- Is your growth organic or manufactured — and how do you know the difference?