What Stage Do You Find Product Market Fit?

The Question Everyone Gets Wrong

Most founders ask "do I have product market fit?" when they should be asking "which stage of finding it am I in right now?" That shift in framing changes everything about what you do next.

I've been through this with multiple companies and exits. PMF is not a switch that flips. It's a progression through distinct stages, each with its own signals, its own traps, and its own definition of "done." Miss the signals in one stage and you'll waste months - sometimes years - optimizing the wrong things.

Here's the other thing nobody says clearly: most of the time when a startup fails, it's not because the founder was lazy or the product was bad. It's because PMF was never actually found. The data backs this up - a lack of product-market fit is one of the most common reasons startups don't make it. The stat I keep coming back to is that only about 10% of startups survive beyond five years. Product-market fit - or the absence of it - is the single biggest variable separating that 10% from everyone else.

So let's break this down properly. Not the VC blog version. The version you can actually use to diagnose where you are right now and figure out what to do next.

What Product Market Fit Actually Means (Before We Talk Stages)

Before we get into stages, let's be precise about the definition - because the vague version is what causes so much confusion.

PMF is when a defined segment of customers consistently gets meaningful value from your product, pays for it, stays, and tells others. Every word in that sentence matters.

• Defined segment - not "everyone," not "businesses," not "people who need this." A specific slice of a market with shared characteristics.
• Consistently - not one good month, not a cohort of early adopters who eventually churned. Repeatedly, over time.
• Meaningful value - not "this is interesting" value. "I would genuinely miss this if it disappeared" value.
• Pays for it - money is the most honest signal in business. If they won't pay, you don't have fit, you have a hobby.
• Stays - retention is the heartbeat of PMF. If people leave, the value wasn't real enough.
• Tells others - organic referral is what separates a business from a business with PMF. When customers sell for you, that's the clearest signal in existence.

Marc Andreessen, who popularized the term, has a simple version: product-market fit means being in a good market with a product that can satisfy that market. That's clean, but it's the output, not the process. What we're talking about here is how you get there, stage by stage.

The other crucial frame: PMF is not permanent. Markets evolve, competitors ship your core feature, customer needs shift. Companies that found PMF and then coasted on it have died because of it. Treat PMF as a living thing you have to keep measuring and defending, not a trophy on the shelf.

The Dan Olsen PMF Pyramid (The Framework Under the Framework)

Before mapping the stages, it's worth understanding the structural framework that the best PMF work is built on. Dan Olsen's Product-Market Fit Pyramid, from his book The Lean Product Playbook, is the most practical model I've seen for thinking about what PMF actually requires.

The pyramid has five layers, each built on top of the one below it:

1. Target Customer - who exactly are you building for? Not a demographic. A specific person with specific problems in a specific context.
2. Underserved Needs - for that target customer, what are their acute pains that existing solutions don't address adequately?
3. Value Proposition - how does your product address those needs better than alternatives? What's your specific claim?
4. Feature Set - which capabilities are actually necessary to deliver on that value proposition? Only those.
5. User Experience - the execution layer. How well does the product actually deliver the features in a way that creates real value?

The bottom two layers are the market side - you don't control them, you respond to them. The top three are the product side - you control these. The reason most products fail to find PMF is that teams jump straight to features and UX while ignoring the foundation of who they're building for and what that person actually needs. They work on the visible parts of the pyramid without getting the invisible base right first.

This is why I always push founders to spend more time at the bottom of the pyramid before touching the top. You can build a technically beautiful product with terrible product-market fit. Happens constantly.

Stage 1: Problem-Customer Fit (Before You Build Anything)

This is the stage almost everyone skips in a rush to ship. It happens before a product exists. Your only job here is to validate that a real, painful problem exists for a specific, reachable group of people.

You're looking for three things:

• Is the problem real? Not assumed. Not theorized. Real - meaning people actively try to solve it today with workarounds, spreadsheets, or competitors.
• Is the problem painful enough to pay to fix? A lot of problems are annoying but not worth money. You need a problem that causes enough friction, cost, or embarrassment that someone will open their wallet.
• Is there a reachable segment that has this problem consistently? If your ideal customer varies wildly - different industries, different roles, different triggers - you're not targeting, you're guessing.

The tool at this stage is conversations, not surveys. Go talk to 20-30 people in your target segment. Don't pitch. Just ask about their current process, what breaks, what they've tried, what they've given up on. The words they use to describe the pain become your marketing copy later.

One signal that you've completed Stage 1: multiple people, unprompted, say some version of "I would pay for something that solved that." If you have to drag that answer out of them, keep digging.

The trap most technical founders fall into here is building before they've talked to enough people. The instinct is to get moving, to have something to show. But the cost of going fast in the wrong direction is far higher than going slower in the right one. I've watched founders spend six months building a product that nobody wanted - not because they were building the wrong thing, but because they never stopped to validate the problem before writing a single line of code.

A minimum viable audience is a useful concept at this stage. Before you define your minimum viable product, define the smallest group of people with a distinct, acute problem that you can solve better than anyone else and who are willing to pay for it. That audience definition is the real foundation everything else sits on.

If you're building something for a B2B audience and need to identify who to interview first, a B2B lead database can help you pull a targeted list of prospects by title, industry, and company size - so you're talking to the right people from day one, not just whoever you can find on LinkedIn. Pull 50-100 names that match your target profile, reach out with a simple "I'm doing research and not selling anything" message, and start conversations. The data on who responds and what they say becomes your early customer intelligence.

Stage 2: Problem-Solution Fit (MVP Validation)

This is where most founders think they're finding PMF. They ship an MVP, a few people use it, and they declare victory. That's almost always premature.

Problem-solution fit means you've confirmed that your proposed solution actually resolves the problem you identified in Stage 1. Not that people find it interesting. Not that they say nice things about it. That it changes their behavior.

The trap here is polite feedback. People will tell you the product is "really cool" or "definitely something I could see using." That is not signal. Signal is someone changing how they work because of what you built - even a rough prototype. Signal is someone pulling out a credit card before you've even finished the demo.

There's a pattern called "hallucinating PMF" - mistaking early sales growth or positive reactions for true product-market fit. Early adopters are often enthusiastic not because your product is right, but because they're the type of people who try everything new. Their enthusiasm doesn't mean the broader segment will respond the same way. Treat early adopter feedback as a starting point, not a conclusion.

At this stage you're running small, controlled experiments. Give your MVP to 10-15 people in your target segment. Watch how they use it. Don't explain it. Watch what confuses them, what they skip, what they keep coming back to. That usage pattern tells you more than any interview will.

Key questions to answer before leaving Stage 2:

• Are people using the core feature you built, or a workaround inside your product?
• Would they be genuinely upset if this went away?
• Can you get anyone to pay - even a token amount - for the early version?
• Are they using it the way you intended, or finding unexpected use cases?
• What is the one thing they do first every time they log in?

That last question is important. The feature your early users gravitate to first and return to most often is almost always the core of your real product - not necessarily what you thought you were building. Follow that behavior signal wherever it leads.

One more tool at this stage: cohort tracking. Even with a tiny user base, start tracking who signed up when, what they did, and whether they came back. Averages hide everything. Break your users into groups by acquisition date and watch what happens over time. That early cohort data - even if it's rough - will tell you whether behavior is improving or not as you iterate.

If you want to pressure-test your business concept before committing to a full build, run it through the Business Idea Roaster - it's a free tool I put together to poke holes in ideas before they cost you real money.

Stage 3: Early Product-Market Fit (The Real Stage)

This is the stage people actually mean when they say "finding PMF." And it's where the magic - and the confusion - lives.

Early PMF happens when a defined segment of customers consistently gets value from your product, pays for it, and sticks around. The key word is consistently. One customer who loves you is a case study. Ten customers who behave the same way is a pattern. That pattern is early PMF.

Here's what the signal looks like in practice:

• Retention curves flatten. When you plot cohort retention over time, you stop seeing everyone churn to zero. A percentage of users stabilize and keep using the product month after month. That flat line is one of the clearest PMF signals available.
• Organic growth starts happening. You didn't run an ad. You didn't have a launch. Someone found you because a customer told them about you. Word-of-mouth, even at small scale, is strong evidence of real value.
• Trial-to-paid conversion improves over iterations. As you refine the product based on user feedback, more people decide to pay. Each iteration should move that number up.
• Customer feedback converges. Instead of getting wildly different requests from every user, you start hearing the same things. That convergence tells you your product has a real job-to-be-done, not just a set of loosely related features.
• Support requests shift. Early on, support is mostly "how do I do X?" - confusion about the product. Post-PMF, support shifts toward "can you add Y?" - engagement from people who already rely on it.

The 40% rule from Sean Ellis is the most widely used benchmark here: survey your active users and ask "How would you feel if you could no longer use this product?" If 40% or more say "very disappointed," you likely have PMF in that segment. Below 40%, keep iterating. Ellis developed this benchmark after studying nearly 100 startups and found that products clearing that threshold almost always achieved sustainable growth, while those below it almost always struggled.

A few important nuances on running this survey properly:

• Survey people who have actually used the core product - not people who signed up and never logged in.
• Target users who have engaged at least twice and been active recently. Someone who loved your product six months ago and hasn't touched it since is telling you something different than a current active user.
• Aim for at least 40-50 responses before treating the number as meaningful. Below that threshold, the result is directional at best.
• Segment the results. Your overall PMF score might be 28% - but when you segment by industry, role, or company size, you might find one segment scoring 55%. That's your real ICP, and that's where to focus all your energy.

That last point is how Superhuman went from a PMF score of 22% to 58%. They didn't rebuild the product. They segmented their users, identified the high-expectation customers who loved what they already had, focused the entire company on serving that precise group, and stopped trying to please everyone. The score followed.

Notice the phrase "in that segment." Early PMF is almost always narrow. You've cracked one specific persona, in one specific context, with one specific use case. That's enough to build on. Don't try to expand the segment before you've fully understood and served the one you've got.

One more signal worth watching: churn rate. If churn is running at 10% or higher per month, you're effectively pre-PMF regardless of your other metrics. High monthly churn means customers are not finding sustained value - and no amount of acquisition growth can fix a fundamental retention problem. You'll just burn money filling a leaky bucket.

Stage 4: Repeatable PMF (Scaling Into It)

Most founders don't think of this as a separate PMF stage, but it is. You can have strong early PMF with a small group of customers and still have no idea how to find the next hundred like them. That gap - between "we have happy customers" and "we know how to get more happy customers predictably" - is Stage 4.

Repeatable PMF means:

• You have a clear, written definition of your ideal customer - specific by role, industry, company size, and the trigger that makes them ready to buy.
• Your sales or acquisition process consistently converts prospects who match that profile.
• A new salesperson or growth channel can replicate your results without heroic effort.
• The economics work at scale - LTV is significantly higher than CAC and the gap is widening, not shrinking.

This is the stage where outbound becomes a major asset. If you know exactly who your customer is, you can build a list of them, reach out directly, and test conversion at scale. Tools like Clay for enrichment and Smartlead for sequencing can help you run systematic outbound to your ideal customer profile without burning out your team.

For list building, if your ICP is a specific type of B2B company, you can use ScraperCity's B2B database to filter by job title, seniority, industry, and company size and pull exactly the segment you've identified as your strongest PMF match. The point isn't to blast the list - it's to test whether your conversion rates hold when you reach outside your warm network. If cold outbound converts at rates similar to warm referrals, your positioning is working. If it collapses, your sales process may still be dependent on personal relationships rather than repeatable messaging.

I also cover the mechanics of building repeatable outbound into early-stage companies inside Galadon Gold - the frameworks I've used across multiple SaaS exits.

One important signal that you've reached Stage 4: a new hire can close deals without you in the room. If you personally have to be part of every sale to close it, you haven't documented the repeatable process yet - you're still the process. That's not PMF at scale, that's founder-dependent selling. Fix the process before you fix the headcount.

The Dan Olsen Pyramid Applied to Each Stage

Here's how the PMF pyramid layers map to the four stages, because understanding which layer you're currently working on helps you avoid doing Stage 4 work when you're still in Stage 1:

The reason this matters: if you're getting poor retention (a Stage 3 problem), the root cause might be in your target customer definition (a Stage 1 problem). You can't fix retention by improving the UX if the fundamental issue is that you're targeting the wrong person. Trace the problem back down the pyramid before trying to fix it at the top.

The Signals That Tell You You're There (And the Ones That Lie)

Let me give you the honest list, because a lot of founders mistake noise for signal at every stage.

Signals that actually mean something:

• Low churn relative to your category. If customers are staying, they're getting value. This is hard to fake.
• LTV outpacing CAC by a significant margin. When the lifetime value of a customer is significantly higher than what it costs to acquire them, the economics are working. That's PMF in financial form.
• Customers describing your product in their own words, consistently. When you start hearing the same unprompted phrases - not phrases you've fed them - your positioning has landed.
• Inbound leads from unexpected channels. When people find you who you never tried to reach, something is pulling them in. That pull is PMF.
• Sean Ellis score at or above 40%. Not just a high score overall - a high score within a clearly defined segment.
• NPS scores above 50. Net Promoter Scores at this level indicate the kind of customer enthusiasm that precedes referral growth.
• Expansion revenue growing. When existing customers buy more, upgrade, or add seats over time, they're voting with their wallet that the product is worth more to them the more they use it.

Signals that are mostly noise:

• High sign-up numbers without retention data.
• Press coverage or social media buzz.
• Compliments from people who aren't customers.
• A big enterprise pilot that hasn't converted to paid.
• Enthusiastic early adopters who don't match your ICP.
• Revenue growth driven entirely by new customer acquisition with flat or declining retention.
• Investors saying they're "interested" - interest is not investment and investment is not PMF.

The difference comes down to one thing: are people paying, staying, and telling others? If yes to all three, you've found it. If you're missing any leg of that stool, you're still working toward it.

How to Run the Sean Ellis PMF Survey (Step-by-Step)

Because so many founders do this wrong, here's exactly how to run it:

Step 1: Choose your audience carefully. Only survey users who have experienced the core of your product - not people who signed up and browsed, not people who used it once six months ago. Active users who have used your primary feature at least twice in the last two weeks. If you're surveying the wrong group, any score you get is meaningless.

Step 2: Send one question first. "How would you feel if you could no longer use [product name]?" with three answer options: Very disappointed / Somewhat disappointed / Not disappointed. That's it for the first question. Keep it simple and don't bury the lead in a long survey.

Step 3: Follow up for context. After the primary question, add follow-ups: What's the main benefit you get from [product]? What type of person do you think would benefit most from it? What would you use instead if it went away? What can we improve? These qualitative answers are where the real roadmap lives - not in the score itself.

Step 4: Get at least 40-50 responses. Below that threshold, the percentage fluctuates too much to be meaningful. Aim for 100+ if you can get them.

Step 5: Segment before you conclude anything. Break responses by industry, role, company size, plan type, or however your customer base varies. Your aggregate score might be mediocre while one segment is crushing it. That segment is your real PMF beachhead.

Step 6: Repeat it on a schedule. Run this survey every quarter. Watch the trend, not just the snapshot. A score moving from 28% to 35% to 41% tells a better story than a single 45% reading from an optimally selected audience.

What PMF Looks Like Across Different Business Types

PMF shows up differently depending on your model. Worth knowing before you benchmark yourself against the wrong standard.

SaaS/subscription: Look for flattening retention curves, rising MRR from expansion (not just new customers), and trial-to-paid conversion improving across iterations. Monthly churn above 10% is a strong signal you're still pre-PMF regardless of how fast you're growing. Pay attention to expansion revenue - it's one of the clearest signs your product is genuinely valuable to the people using it.

Agency/service business: PMF here looks like a repeatable client type who closes easily, stays long, and refers others. If every client feels like a new fight - new objections, new scope creep, new reasons to churn - you haven't found the segment that fits yet. When you have PMF in a service business, your close rate on qualified leads should be consistently high and your average engagement length should be growing, not shrinking.

Marketplace: Both supply and demand sides need to find value. Watch liquidity - are transactions completing without heavy manual intervention? Organic supply-side growth is especially telling. If suppliers are joining without you recruiting them directly, demand-side PMF is pulling them in. That's a strong signal.

Consumer apps: Retention and organic growth matter most. DAU/MAU ratio (daily actives vs. monthly actives) tells you how sticky the product really is. High DAU/MAU means people have made it a habit - the highest form of PMF for consumer. Word-of-mouth growth and time-in-app trends matter more than download counts.

B2B software with long sales cycles: You need to be careful here - a few $100K enterprise contracts can mask serious PMF problems. High annual contract values can give the appearance of being further along than you are. Look at expansion rates, renewal rates, and NPS within existing accounts to get the real picture.

Why You Can Lose PMF After Finding It

This is the part nobody talks about enough. PMF is not permanent. Markets change. Competitors ship. Customer needs evolve. I've seen companies that had genuine PMF at one stage lose it within 18 months by doing three very predictable things:

ICP Drift. Sales starts closing accounts outside the defined ICP because they're "close enough." The product wasn't built for those customers. Churn rises. Support costs spike. NPS falls. And suddenly the retention metrics that used to look healthy start degrading. The fix is to requalify your ICP regularly against retention data - if the customers closing most easily aren't the ones staying longest, you have a misalignment that needs correction.

Competitive Displacement. A competitor ships the core feature you built your PMF around. If your PMF was built on a single differentiator that's easy to replicate, it can be eroded fast. The only defense is continuous deepening - making the product better at the thing it's uniquely good at in ways that are harder to copy. Feature breadth won't save you. Depth will.

Market Evolution. The problem you're solving becomes less urgent as the market matures or as buyer alternatives expand. This one is the hardest to see coming because the market is changing around you while your product and positioning stay static. The founders who survive this are the ones who never stopped doing customer discovery even after finding PMF - who kept having the conversations that told them when the market was shifting before the revenue charts showed it.

The practical defense against all three is simple: keep measuring PMF as an ongoing discipline, not a one-time achievement. Run the Sean Ellis survey quarterly. Watch churn by segment. Stay close to your best customers and ask them what's changing in their world.

How to Actively Move Through the Stages Faster

PMF is not something that just happens to you. You can accelerate the process with a few deliberate moves:

• Talk to customers on a schedule, not just when something breaks. Weekly customer conversations during early stages compress the feedback loop dramatically. Treat it like a standing meeting with your most important stakeholder - because it is.
• Kill features, not just add them. Most early products are trying to be too many things. The path to PMF is often subtraction - finding the one thing that actually matters and making it undeniable. Every feature you add that isn't core to your value proposition is a feature that dilutes your positioning and adds complexity your users have to navigate.
• Use outbound to test positioning. Cold email is one of the fastest ways to test whether your value proposition lands. If your email gets replies and meetings, your messaging is working. If it gets silence, the market is telling you something. The Cold Email Manifesto covers the mechanics in detail - the process of writing, testing, and iterating on cold outreach is genuinely one of the fastest feedback loops available to a pre-PMF founder.
• Use cohort analysis, not averages. Averages hide everything. Break your users into cohorts by acquisition date and track behavior over time. You'll see which cohorts retained, which churned, and whether you're improving. A product with improving cohort retention is moving toward PMF. A product with flat or degrading cohort retention is not, regardless of what the monthly aggregate numbers say.
• Define your ICP in writing before you think you need to. Most founders think the ICP is a marketing exercise for later. It's not. Having a written, specific ICP forces you to make choices in Stage 1 and Stage 2 that would otherwise get deferred. A vague ICP produces vague customer conversations, vague MVPs, and vague feedback. Precision at the ICP level cascades into clarity everywhere else.
• Find your power users and clone them. In every early user base, there are a handful of people who use the product differently - more frequently, more deeply, in ways that generate more value. Find those people. Interview them. Understand what makes their context different. Then build your ICP around that context and your acquisition strategy around finding more people who match it.

For building contact lists of people who match your refined ICP - whether you're doing discovery interviews or running outbound tests - an email finding tool can help you turn a name and company into a reachable contact quickly, so you're spending time on conversations rather than on digging for contact details.

If you're in the idea stage and want fresh angles to test, the Daily Ideas Newsletter can give you a regular stream of business concepts to stress-test against real market needs. And if you're building something in the SaaS space, the SaaS AI Ideas Pack is a good starting point for identifying under-served problems worth pursuing.

Common Mistakes at Each Stage (And How to Avoid Them)

I've watched enough companies go through this process to know the failure modes are predictable. Here's the mistake map:

Stage 1 Mistakes:

• Doing surveys instead of conversations. Surveys are useful later. At Stage 1, you need the messy, unstructured, "tell me more about that" depth that only comes from talking to humans. A survey will tell you what people think they think. A conversation will tell you what they actually do and why.
• Talking to the wrong people. This is the biggest one. Talking to people who are supportive but not representative of your actual target segment produces false confidence. Be ruthless about whether the people you're interviewing actually match your hypothesized ICP - their job title, their company stage, their trigger for having the problem.
• Pitching instead of listening. The discovery conversation is not the sales call. The moment you start talking about your solution, you've contaminated the data. Ask about their world. Let them talk. Resist the urge to show them what you're building until you've fully understood what they need.

Stage 2 Mistakes:

• Mistaking polite feedback for validation. "This is really interesting" is not signal. "I would pay for this right now" is signal. Learn to tell the difference and stop treating compliments as data.
• Building too much before testing. The MVP is supposed to be minimum. Most founders build at least twice as much as they need to validate their core hypothesis. Define the single assumption that, if wrong, would kill the whole idea. Then build only enough to test that one assumption.
• Not watching behavior. User interviews tell you what people say. Usage data tells you what they do. Both matter, but behavior is more honest. If users say they love a feature and never use it, the feature doesn't belong in the product.

Stage 3 Mistakes:

• Declaring PMF too early. A few excited early adopters or one strong sales month isn't PMF. Wait for sustained signals across multiple metrics - retention, referrals, converging feedback, improving conversion - before calling it.
• Confusing acquisition with retention. Getting customers is not PMF. Keeping them is. Many founders celebrate acquisition metrics while ignoring whether those new users stick. Growth built on weak retention is a leaky bucket that accelerates cash burn without building a real business.
• Surveying the wrong users for the Ellis test. If you include churned users, inactive users, or users who only used a secondary feature, your PMF score will be artificially low - or artificially high if you cherry-pick enthusiasts. Survey your active, engaged core users and segment from there.

Stage 4 Mistakes:

• Scaling acquisition before codifying the process. Hiring more salespeople before you have a repeatable sales playbook doesn't multiply results - it multiplies chaos. Document what works before you try to scale it.
• Expanding ICP too early. The temptation at Stage 4 is to go wider - serve more types of customers, enter adjacent markets, add use cases. This almost always backfires if done before you've completely dominated the beachhead segment. Go deeper before you go wider.
• Stopping customer discovery. The biggest mistake at Stage 4 is thinking you know everything you need to know about your customer. Markets shift. The customer who had your problem in a certain way last year might have it differently this year. Keep talking to customers even after you've found PMF - especially after you've found it.

The Role of Outbound in Finding PMF Faster

There's an underrated connection between systematic outbound and faster PMF discovery that most founders miss. Outbound isn't just a revenue tactic - it's one of the fastest market research tools available.

Here's how I think about it: when you're writing a cold email to a specific ICP, you're forced to articulate why your product solves a problem they have, in terms they'll recognize, in a way that's compelling enough to get a reply. If you can't write that email, you don't understand your value proposition yet. And if the email doesn't get replies, you're either targeting the wrong people or the value proposition isn't landing.

That feedback loop - write email, send to ICP, measure reply rate, iterate - is dramatically faster than waiting for organic discovery to produce signal. A well-run cold email test to 200 people who match your ICP can give you more useful PMF signal in two weeks than six months of watching free trial users meander through your product.

The mechanics of running that kind of outbound - the list building, the sequencing, the iteration on messaging - is covered in depth in The Cold Email Manifesto. The short version: keep the email short, make the value proposition specific to their role and problem, ask for a conversation not a demo, and track reply rate as your primary metric. Reply rate tells you whether the market recognizes the problem you're solving. That's PMF research.

For list building to support outbound-based PMF testing, you need clean, accurate contact data. Tools like Findymail and ScraperCity's B2B database let you filter by exact ICP criteria and pull verified contact information so you're testing your messaging with the right audience, not a random slice of whoever you happen to know. The quality of your test depends entirely on the quality of the audience you're testing against.

Once you have those lists, running sequences through tools like Instantly or Smartlead lets you test multiple value proposition angles at scale - essentially A/B testing your PMF messaging in real time against a qualified audience.

What PMF Is Not (Busting the Myths)

A few things worth clearing up because they cause real confusion:

PMF is not launch day traction. Product launches create a temporary spike driven by curiosity, novelty, and the founder's network. That spike has almost nothing to do with PMF. Watch what happens to that cohort 30, 60, and 90 days after launch. The trend after the spike is the signal. The spike itself is noise.

PMF is not a single metric. There is no one number that definitively tells you whether you have product-market fit. The Sean Ellis 40% rule is a useful benchmark, not an absolute rule. An NPS of 60 is a strong signal but not a guarantee. Flat retention curves are meaningful but not the whole story. PMF is the convergence of multiple signals pointing the same direction - retention, referral, payment, converging feedback. When they all point up simultaneously, that's PMF.

PMF is not product quality. A technically excellent product can have terrible product-market fit. A rough, buggy product can have strong PMF. Quality matters for retention and word-of-mouth downstream, but it's not the primary driver of initial PMF. The question is whether the product solves the right problem for the right person, not whether it's beautifully engineered.

PMF is not revenue alone. High revenue can coexist with poor PMF if it's driven entirely by new customer acquisition and masked by growth. Look at whether the customers you acquired three months ago are still around, whether they're paying more than when they started, and whether they'd refer someone. Revenue without retention is a warning sign, not a validation.

PMF is not the same across segments. This is the one that trips up the most founders. You might have strong PMF with small B2B companies and essentially zero PMF with enterprise. Or strong PMF in one vertical and no traction in another. Always segment your analysis. Aggregate numbers are averages that often hide the truth about where your real PMF lives.

The Bottom Line on PMF Stages

To answer the original question directly: you find product market fit in Stage 3 - after you've validated the problem (Stage 1) and validated that your solution actually addresses it (Stage 2). But you only know you've found it by watching what happens in Stage 4, when you try to repeat the result with new customers who weren't part of your original network.

The mistake most founders make is trying to scale before completing Stage 3, or declaring Stage 3 complete based on the noise signals instead of the real ones. Retention, willingness to pay, organic referrals, and converging feedback - those are your scorecards. Everything else is a distraction.

The second most common mistake is thinking PMF is permanent once you have it. It's not. Markets shift, competitors copy your differentiators, customer needs evolve. The founders who sustain PMF are the ones who treat customer discovery as an ongoing practice, not a phase that ends when you find the initial fit.

The PMF pyramid framework gives you a structured way to diagnose where the problem is when things aren't working. If retention is poor, trace it back down the pyramid - you might have a UX problem, but you might have a target customer problem. If conversion is poor, the issue might be in your value proposition, not your sales process. The pyramid helps you identify the root cause rather than treating symptoms.

Finally: the fastest path through all four stages is staying close to real customers and letting their behavior - not their words, not their compliments, not their promises - tell you what to do next. When you've genuinely found it, you'll know - not because someone told you, but because the market starts pulling you forward instead of you pushing.

If you want help navigating the mechanics of outbound, ICP definition, and the GTM moves that come after early PMF, I work through this in depth inside my coaching program with founders who are in the thick of it right now.