DIY PMF measurement vs done-with-you diagnosis: an honest comparison

"Should I just run the survey myself, or pay for help" is a question with a real answer, and it isn't "whichever you can afford." Both paths use the same underlying method, the Sean Ellis 40% test. What differs is who does the targeting, who chases the responses, and who turns the number into a decision.

What DIY actually requires

Running it yourself with PMFtracker means you handle four things: defining who counts as an engaged user, sending the survey to that exact group (not your whole list), getting to at least 40 valid responses, and reading the segmented result correctly, very disappointed, somewhat, not, and the open-ended answers underneath each. The tool does the calculation and the tracking. It doesn't do the targeting or the interpretation for you.

This is genuinely enough for a lot of founders. If you already have an engaged user base, you're comfortable designing a survey, and you plan to track the score on a recurring cadence rather than get a one-time answer, DIY is the right call, and it's a fraction of the cost.

What a diagnosis adds

The PMF Diagnosis is a four-week, fixed-scope engagement run by Olek Uznanski, done with you rather than left for you to figure out. For products with 40 or more engaged users, that means survey design and targeting, chasing responses to a sample that actually means something, a PMF score against the 40% benchmark with a confidence band, an ICP built from the "very disappointed" cohort, and a ranked list of barriers to fix, delivered as a written diagnosis plus a 90-minute readout call.

For pre-launch products or anyone under 40 users, a percentage would be a statistical guess dressed up as data, so the engagement switches to the interview-based track instead: 8 to 12 customer interviews synthesized into a jobs-to-be-done read, an ICP, and a build-or-pivot verdict. No score, on purpose.

The part most comparisons skip: it isn't either/or

Every PMF Diagnosis engagement includes a lifetime PMFtracker licence at no extra cost. You don't choose between the tool and the service, the diagnosis hands you the tool as part of the outcome, so you can keep tracking your score after the engagement ends. Thinking of this as "buy the tool or hire the person" undersells what the diagnosis actually is: the tool plus the targeting, the chasing, and the interpretation you'd otherwise have to do yourself.

How to decide, honestly

• Go DIY if: you already have 40+ engaged users, you've read a segmented survey result before, and you want an ongoing number rather than a one-time verdict.
• Get the survey-track diagnosis if: you have the users but not the time or confidence to design the targeting and read the result correctly, or you need something you can hand an investor.
• Get the interview-track diagnosis if: you're pre-launch or under 40 users. A percentage at that sample size is a guess, not a measurement, so a score isn't the honest deliverable, a qualitative verdict is.

Whichever you pick, the mistake to avoid is skipping the measurement stage entirely and guessing. A wrong number you act on is worse than no number, and a number you never generate is worse still.