The P/MF Matrix: 4 steps, 3 tools, 12 questions to find product/market fit
Short answer: The P/MF Matrix is a 4 × 3 grid. The rows are the four steps to Product/Market Fit, in order: ICP, offer (value proposition), product, business model. The columns are the three tools you apply to each step: assumptions, experiments, and narrative. Each of the 12 cells is one question. Work through them row by row, and you always know what to do next.
Founders don't fail for lack of effort. They fail because they work on the wrong thing at the wrong time. Usually that means building the product before they know who it's for.
The Matrix fixes the order.
The four steps (rows)
- The ICP. Your Ideal Customer Profile is your beachhead: where the opportunity is yours to take. Get the ICP right and the rest follows.
- The offer. Your value proposition is what your ICP is after. When it lands, they'll opt in to whatever you ask in return.
- The product. Once the ICP and the offer work together, you know exactly what to build. Think lean, and iterate on feedback, not dreams.
- The business model. Products deliver value to customers. Business models drive value back to you. Without a sustainable one, you have nothing to satisfy investors, or yourself.
The three tools (columns)
- Assumptions. Your main tool. Write down what you believe about each step, then find the riskiest beliefs.
- Experiments. Direct and indirect interactions with the market that test those assumptions in the real world. Surveys, landing pages and MVPs are all experiments.
- Narrative. A well-founded story tells people plenty about you as a founder. What's your story, and do you have enough evidence to base it on?
The Matrix
| Step | Assumptions | Experiments | Narrative |
|---|---|---|---|
| 1. ICP | Who in my addressable market is most likely to become a super user with minimal effort from me? | Run surveys and interviews to find the traits your potential power users share. | Write an anecdote that shows your ICP's pain. |
| 2. Offer | What is the ICP after that competing products can't deliver? | A/B test messaging and value propositions to see which resonates with potential ICPs. | Describe the desired future in your ICP's own words. |
| 3. Product | Which features get the ICP to the Aha! Moment most efficiently? | Test prototypes and betas with different ICPs. Look for emotional responses: delight and expectation. | Describe how your product creates a transformation no other product does. |
| 4. Business model | How can I retain the most value from the value my product creates? | A/B test business models to find which one keeps the most value. | Explain how the value you create for users comes back to your company. |
How to use it
Go row by row. Don't start on row 3 because building is more fun. Each row depends on the one above it: you can't write an offer without knowing the ICP, and you can't scope a product without knowing the offer.
Go left to right within a row. Write the assumption, test it, then tell the story. If the experiment proves you wrong, go back to the assumption, not forward to the next row.
Expect to loop. Change the ICP and the rows below it change too. That's not failure. That's the Matrix doing its job.
Use the narrative column for investors. The right-hand column, filled in with evidence, is your pitch. It tells the market story and the founder story at once: here's who it's for, what they want, how we deliver it, and how value comes back. How to prove P/MF to investors goes deeper.
A quick self-check
Go through the 12 cells and mark each one:
- Evidence: you tested it, and you have real market data.
- Belief: you have an answer, but it's an assumption.
- Blank: you haven't thought about it.
Your next step is the first non-evidence cell, reading top to bottom, left to right.
Common mistakes
- Filling the Matrix with opinions. An answer isn't evidence. A cell is done when the market has answered it, not when you have.
- Skipping the narrative column. If you can't tell the story, you don't understand the step yet.
- Starting at row 3. The most common and costly mistake founders make is starting with the product.
What to do next
- Copy the 12 questions into a doc.
- Mark each cell evidence, belief, or blank.
- Design one experiment for your first non-evidence cell.
Prefer to work through it with an AI coding agent? Plan PMF, my free Claude Code plugin, walks you through the first three rows and sets your validation thresholds.
Want the full method in one place? Get the free P/MF Quick Guide, or tell me which cell you're stuck on.