Here's the thing almost nobody tells you about the Sean Ellis PMF survey: its power is entirely in the exact wording of one question. Swap "very disappointed" for "how satisfied are you" and you've built a completely different, far less useful instrument that happens to look the same. So before you send anything, get the questions right — down to the word.
There are five. One produces the number. The other four explain the number and tell you what to do about it. Let's go through each, why it's worded the way it is, and then the whole set in one copy-paste block.
The one question that defines product-market fit
This is the question. Everything else is supporting cast.
"How would you feel if you could no longer use [product]?"
Respondents pick one of these — and the options matter as much as the question:
| Answer option | What it captures |
|---|---|
| Very disappointed | Real fit — they'd feel genuine loss. This is the only answer that counts toward your score. |
| Somewhat disappointed | The fence-sitters — a pool of potential converts worth studying. |
| Not disappointed (it isn't that useful) | No fit for this user. Useful signal, but not your target. |
| N/A — I no longer use it | Churned. Excluded from the score so it doesn't distort the result. |
The share of valid respondents who answer "very disappointed" is your PMF score. 40% or above is the benchmark for product-market fit. Why "disappointed" and not "satisfied"? Because satisfaction is polite and cheap — people say they're satisfied about things they'd drop tomorrow. Disappointment measures loss, and loss is where real attachment shows up.
That N/A option isn't a formality, either. Without it, churned users pick "not disappointed" and quietly drag your score down for the wrong reason. Drop the N/A answers from your denominator so you're only scoring people who actually use the product.
The 3 follow-up questions that turn a score into a roadmap
A lone score tells you whether you have fit. These three open-ended questions — the set Rahul Vohra used to build the Superhuman PMF engine — tell you why, and what to do next. Ask them right after the main question.
2. "What type of person do you think would most benefit from [product]?"
Answered by your "very disappointed" users, this hands you your ideal customer profile in their own words. Patterns in these answers tell you exactly who to target — and who's dragging your score down because they were never the right fit.
3. "What is the main benefit you receive from [product]?"
This is your value proposition, written by the people who actually value it. The words your happiest users use here are the words that belong on your landing page and in your onboarding — usually sharper than anything you'd write yourself.
4. "How can we improve [product] for you?"
Read this one carefully by segment. Improvement requests from your very disappointed users are your roadmap to keep them; requests from somewhat disappointed users show you what's standing between a fence-sitter and a fan. Ignore requests from the "not disappointed" crowd — building for them pulls you toward the product death cycle.
5. (Optional) "Have you recommended [product] to anyone? If so, how did you describe it?"
A light fifth question that surfaces organic pull and gives you ready-made word-of-mouth language. Skip it if you're worried about length.
The full question set — copy & paste
Here's everything in order, ready to drop into your survey tool. Replace [product] with your product name.
1. How would you feel if you could no longer use [product]?
◦ Very disappointed
◦ Somewhat disappointed
◦ Not disappointed (it isn't that useful)
◦ N/A — I no longer use [product]
2. What type of person do you think would most benefit from [product]? (open text)
3. What is the main benefit you receive from [product]? (open text)
4. How can we improve [product] for you? (open text)
5. (Optional) Have you recommended [product] to anyone? If so, how did you describe it? (open text)
That's the whole survey. Four questions, five with the optional one. Resist the urge to add more — every extra question costs you completed responses, and you already have everything you need to measure fit and act on it.
Skip the setup — run this exact survey free
PMFtracker ships with these questions built in, targets your engaged users automatically, drops the N/A responses, and calculates your score the moment answers come in. No survey tool to wire up, no spreadsheet math.
Run your PMF survey free → 14-day free trial · No credit cardWording rules that make or break the survey
- Use the exact phrasing. "Very disappointed" is calibrated against decades of results. Reword it to "very upset" or "how satisfied" and your 40% benchmark no longer means anything.
- Keep the N/A option. It's what separates your real users from your churned ones in the math.
- Don't lead. Ask "how can we improve [product]," not "would you like feature X?" The open text is where honest signal lives; a leading question buries it.
- Only survey engaged users. Send it to people who've used the core product recently — ideally at least twice. Who you ask matters as much as what you ask.
- Keep it short. Four to five questions. Every extra one trades insight for a lower completion rate.
Questions to avoid (and why)
Three "PMF survey" questions look reasonable and quietly wreck your read:
- "How satisfied are you?" (1–10). Satisfaction is not fit. People rate things 8/10 and cancel next month. It measures politeness, not attachment.
- The NPS question ("How likely are you to recommend…?"). NPS is related but not the same — willingness to refer isn't the same as personally not wanting to lose the product. Use the disappointment question for fit; keep NPS as a separate metric if you like it.
- "What features do you want?" Asked cold, this invites your least-fit users to design your roadmap. Only read feature requests after segmenting by the main question.
Get the five questions right and the rest of the product-market fit journey gets a lot clearer — because for the first time you're measuring the real thing, in the users' own words, on a number you can watch move.
Turn these questions into a tracked score
Ask the right questions once, then let PMFtracker keep asking — it re-runs the survey on your engaged users, scores it, and trends it over time so your PMF number is always current.
Start free → 5-minute setup · No credit card
