PMF measurement for solo founders and bootstrappers without a data team

A lot of PMF advice reads like it was written for a 12-person growth team: build a dashboard, segment by six dimensions, run it as a continuous pipeline. None of that is wrong, it's just aimed at a stage most solo founders and bootstrappers aren't at yet. If it's one person doing this part-time, alongside building the actual product, the real question isn't "what's the best-practice setup." It's "what's the smallest version of this that still tells the truth."

What you can skip

These are the things that show up in every "how to measure PMF" checklist and that a solo founder can safely ignore for now, without losing anything that matters:

• A dedicated analytics dashboard. If you're the only person who looks at the numbers, a live dashboard is a nice-to-have, not a need-to-have. A spreadsheet you update after each survey round does the same job.
• Continuous cohort-tracking software. Tracking retention cohorts automatically matters when you're running this weekly across a large team. At 40 to 100 users, you can eyeball a retention curve from raw usage exports just as reliably.
• Deep multi-dimensional segmentation. Splitting by plan tier, company size, channel, and use case simultaneously is powerful once you have hundreds of responses per segment. With 40 total responses, one segmentation cut (usually plan tier or use case) is all your sample size can support without the sub-groups becoming too small to trust.
• A formal research operations process. Panel recruitment, incentive management, structured research repositories: all real, all overkill for a founder who can just message their own users directly.

What you cannot skip

Two things stay non-negotiable no matter how small the team is, because skipping them doesn't save time, it just produces a number that looks precise and means nothing.

1. The response floor

You need roughly 40 valid responses from engaged users before a PMF score is directional, and well over 100 before it's investor-grade. This doesn't change because it's just you running the survey. A score built on 12 replies from whoever happened to answer an email isn't a smaller, scrappier version of the truth, it's noise with a percentage sign on it. The full math on sample size and margin of error is worth reading once so you know exactly how much to trust an early number.

2. Asking the right users

The other non-negotiable is who you survey. It has to be genuinely engaged users, people who've used the core feature more than once, not everyone who ever signed up. A solo founder is often tempted to blast the entire user list to hit the response floor faster. That inflates the sample size while diluting the signal, because inactive or trial-abandoning users answer "not disappointed" almost by default, and their presence drags the whole score down without telling you anything real about product-market fit. Defining "engaged" narrowly and correctly matters more than defining it quickly. If you're still building that list, this covers how founders actually get their first 100 users in the first place.

A realistic solo-founder setup

Here's what a minimal, honest version of this looks like in practice:

1. Define engaged as "used the core action at least twice in the last two weeks." Pull that list from your product's own event log or a free analytics tool you already have.
2. Send the exact Sean Ellis question and the three open-ended follow-ups using whatever you already have, in-app message, email, even a form link in a DM. The wording matters more than the delivery mechanism.
3. Chase to ~40 responses. Expect to need to reach out to more than 40 people to get 40 replies; a reminder after four or five days roughly doubles response rates for most early-stage founders.
4. Score it in a spreadsheet or a free calculator and read every open-ended answer yourself. At this volume, you'll spot the repeated themes faster reading them once than any dashboard would show you.
5. Re-run quarterly or after any real product change, not continuously. A trend from two or three data points, taken a few months apart, tells you far more than obsessing over one number weekly.

Nothing in that list requires a data team, a subscription to five tools, or a full day of setup. It requires discipline on the two things that actually determine whether the number is real: enough of the right people, asked correctly.