The Most Famous Advice on PMF Is Also the Most Useless

Marc Andreessen once wrote that product-market fit means "being in a good market with a product that can satisfy that market." His practical guidance? "You can always feel when product-market fit isn't happening. And you can always feel it when it is happening."

That sounds nice. It's also completely useless if you're a founder sitting at your laptop at 11pm, staring at a dashboard that shows some growth but not explosive growth, some retention but not stellar retention, and a handful of users who really love what you've built alongside a larger group who signed up and never came back.

Most founders don't experience PMF as a thunderclap. They experience it as a persistent question: "Do we have it? Are we close? Or are we fooling ourselves?" The answer matters because it determines whether you should be stepping on the gas (hiring, spending on ads, raising money) or still searching (talking to users, cutting features, repositioning).

This article is about making the question answerable. Not with vibes, but with concrete signals you can actually measure.

The Sean Ellis 40% Test

Sean Ellis, who coined the term "growth hacking" and led early growth at Dropbox, LogMeIn, and Eventbrite, developed the most widely used quantitative test for product-market fit. It's simple enough to run in a single survey question.

You ask your existing users: "How would you feel if you could no longer use this product?"

If 40% or more of respondents say "very disappointed," you have product-market fit. Below that threshold, you need to keep iterating.

Ellis benchmarked this number against hundreds of startups. Companies that went on to scale successfully almost always cleared 40%. Companies that stalled out usually sat in the 20-35% range. The precision of the threshold is debatable, but the direction is reliable. If only 15% of your users would be very disappointed to lose your product, you do not have PMF. Full stop.

A few practical notes on running this test well:

The 40% test isn't a replacement for judgment. But it gives you a number, and numbers are better than vibes when you're deciding whether to spend $50,000 on your next marketing push.

The Five Real Signals of Product-Market Fit

Beyond the Ellis test, PMF shows up in observable patterns. These are the signals that actually indicate fit, as opposed to the ones that feel good but lie to you.

1. Organic word-of-mouth is your biggest growth channel

When people use a product that genuinely solves a problem for them, they tell other people. Not because you asked them to. Not because you gave them a referral bonus. Because solving the problem was remarkable enough to mention in conversation.

If your "How did you hear about us?" responses are dominated by "a friend told me" or "my coworker shared it," that's PMF talking. If you're getting inbound inquiries from people you've never marketed to, that's PMF. This is different from viral mechanics or referral programs. Those are growth levers. Word-of-mouth is a symptom of genuine value.

2. Users come back without prompting

Look at your retention curves. Not the Day 1 numbers (everyone looks good on Day 1). Look at Day 7, Day 14, Day 30. Are people returning to your product without you sending them an email reminder, a push notification, or a discount code?

A product with PMF has retention curves that flatten out rather than declining to zero. Some percentage of users (it doesn't have to be huge) forms a habit around the product. They come back because the product is woven into their workflow, not because your lifecycle marketing is effective.

3. People pay without heavy convincing

Sales cycles tell you a lot. If prospects need four demo calls, a custom proposal, and two weeks of follow-up emails before they'll commit, you might have a product people are interested in but don't urgently need. If prospects see a demo, ask about pricing, and say "let's do it," you probably have fit.

This shows up in conversion rates too. Products with PMF typically convert free-to-paid at 3-5% or higher. Products without it convert at fractions of a percent. The gap is enormous. When the product solves a real problem, the pricing conversation becomes simple.

4. Usage grows without proportional marketing spend

Plot your monthly active users against your marketing budget over the last six months. If usage is growing faster than your spending, organic forces are at work. If usage tracks exactly with your ad spend (up when you spend, down when you don't), you're buying users, not earning them.

This is a signal that gets overlooked because founders are taught to "invest in growth." But there's a critical difference between pouring fuel on a fire and trying to start a fire with dollar bills. PMF is the fire. Marketing spend is the fuel. If there's no fire, more fuel won't help.

5. Customers describe the product the same way you do

This is subtle but powerful. Ask five of your most active users: "How would you describe this product to a friend?" If their answers roughly match your positioning, your messaging and your value delivery are aligned. If they describe something completely different from what your landing page says, you either have a messaging problem or a product identity problem.

When Basecamp had PMF, users described it as "the simple way to manage projects without all the complexity of Microsoft Project." That mapped perfectly to what 37signals (now Basecamp) was saying about their own product. The message matched the experience.

Three False Signals That Trick Founders

For every real PMF signal, there's a convincing fake. These are the patterns that make you feel like you've found fit when you haven't.

Friends and family signing up

Your first 50 users will include your college roommate, your mom, three people from your co-working space, and that investor you had coffee with who said "I'd love to try it." These people are signing up because they care about you, not because they have the problem your product solves.

There's nothing wrong with this. Early traction has to start somewhere. But don't confuse social support with market demand. The test is whether strangers with no personal connection to you are adopting the product. If your user base is entirely people who would come to your birthday party, you don't have data on PMF yet.

One big customer loves you

Landing an enterprise contract feels like validation. A Fortune 500 company is paying you $80,000 a year. That must mean you have PMF, right?

Not necessarily. One customer tells you that one organization finds your product useful. PMF means a market finds your product useful. If that enterprise client needed extensive customization, if their use case is unique to their industry, if you couldn't describe their requirements in a way that maps to hundreds of similar companies, then you have a consulting client, not product-market fit.

The question to ask: "Could I sell this same product, configured the same way, to ten more companies like this one within 90 days?" If the answer is no, the single customer is a false positive.

Press coverage without retention

You got featured in TechCrunch. Traffic spiked 10x. Sign-ups went through the roof. This is exciting, and it feels like the beginning of something. Then you look at the numbers a month later. 95% of those new users never came back after their first session.

Press is a distribution event, not a validation event. It tells you that your story was interesting enough to write about. It says nothing about whether your product solves a real problem for real people. Products with PMF can survive a press spike because they retain the users who show up. Products without PMF get a dopamine hit and then return to baseline.

How Slack Knew: The 2,000-Message Threshold

Slack's growth story is well documented, but the specific moment they identified PMF is instructive. Stewart Butterfield and his team discovered a metric that predicted whether a team would become a long-term Slack user: 2,000 messages.

Teams that exchanged 2,000 messages on Slack almost never churned. The conversion from "trying it out" to "this is how we communicate now" happened somewhere around that threshold. Below 2,000 messages, teams might or might not stick around. Above it, they were locked in.

This insight did two things for Slack. First, it gave them a quantitative definition of their "aha moment." They could measure exactly how many teams were crossing the line from trial to committed usage. Second, it told them where to focus their onboarding efforts. Everything about the early Slack experience was designed to reduce friction between sign-up and the 2,000th message. Slackbot, the default channels, the GIF integrations, the smooth mobile experience. All of it was oriented toward getting teams to that activation point faster.

The broader lesson: PMF isn't a binary state for your whole product. It's a threshold that individual users or teams cross. Your job is to find that threshold, measure it, and clear the path to it.

Superhuman's PMF Engine

Rahul Vohra, the CEO of Superhuman, built what might be the most systematic approach to finding product-market fit ever documented. He wrote about it in a 2019 essay called "How Superhuman Built an Engine to Find Product-Market Fit," and the framework has become a reference for founders ever since.

Vohra started with the Sean Ellis survey. Superhuman's initial score was 22%. Not great. But instead of panicking or pivoting, he did something methodical. He segmented the responses.

He looked at which types of users were saying "very disappointed" and which were saying "not disappointed." The very disappointed users tended to be founders, managers, and executives who received a high volume of email and valued speed. The unimpressed users tended to be people who didn't process much email or who used email primarily for marketing campaigns.

This segmentation revealed the path forward. Superhuman didn't need to make everyone happy. They needed to make their core audience (high-volume email users who valued speed) even happier, while deprioritizing features that only mattered to non-core users.

Vohra then looked at the "somewhat disappointed" group and asked what would make the product better. These people were almost fans. Their feedback was a roadmap for turning lukewarm sentiment into strong advocacy. He cross-referenced their suggestions with the qualities that the "very disappointed" group already loved, and built features that deepened the core value without diluting it.

Over several quarters of running this loop (survey, segment, build, re-survey), Superhuman's Ellis score climbed from 22% to 58%. They didn't stumble into PMF. They engineered it.

The takeaway is concrete. If your Ellis score is below 40%, don't guess at what to build next. Survey your users, segment the responses, identify your most enthusiastic cohort, and double down on what they love. PMF can be a process, not just an outcome.

Notion's Slow Burn to PMF

Not every PMF story looks like Slack's explosive adoption curve. Notion is the counterexample that founders who are grinding through a slow build need to hear.

Notion launched in 2016 and spent nearly two years with modest traction. Ivan Zhao, the co-founder, was so committed to the vision of an all-in-one workspace that he relocated the team to Kyoto, Japan, to cut costs and focus on product development. They had some users. They had some revenue. They did not have PMF.

What they did have was a small group of power users who were building elaborate systems in Notion: personal wikis, project trackers, content databases. These users weren't just using Notion. They were creating templates and sharing them with others. They were writing blog posts about their Notion setups. They were doing the marketing for free because they were genuinely excited about what they'd built inside the tool.

Notion's PMF emerged from this template-sharing behavior. The company leaned into it, making templates easier to create and share, building a template gallery, and letting the community become the primary distribution channel. By 2019, Notion had crossed the PMF threshold and growth accelerated dramatically. By 2020, they were valued at $2 billion.

The lesson from Notion is patience, paired with observation. Zhao didn't try to force PMF by adding features for users who weren't engaged. He watched what the engaged users were doing, identified the behavior that signaled genuine love for the product, and built around that behavior. PMF sometimes announces itself quietly. You have to be paying attention.

What to Do If You Don't Have PMF Yet

Here's where most founders make the same mistake. They don't have PMF, they know they don't have it, and their instinct is to add more features. More capabilities means more value, right? If the product did more things, more people would want it.

This instinct is almost always wrong. The path to PMF isn't wider. It's narrower.

Narrow the audience, don't broaden the product

If 15% of your users say "very disappointed" on the Ellis survey, those users are telling you something important. They're telling you that your product does work, for them. Your job is to figure out what makes them different from the 85% who are lukewarm or disengaged.

Maybe they're all in the same industry. Maybe they share a company size. Maybe they all use your product for the same specific workflow. Whatever the pattern is, that's your real ICP (ideal customer profile). Narrow your focus to serving those people exceptionally well. Cut features that only matter to the other 85%. Rewrite your messaging to speak directly to the 15%.

Counterintuitive as it sounds, shrinking your addressable market often increases your traction. A product that's perfect for 500 people will grow faster than a product that's okay for 50,000.

Talk to churned users

The users who signed up and left are a goldmine of information, and almost nobody mines it. Send them a short email: "Hey, I noticed you tried [product] but stopped using it. I'm the founder, and I'd genuinely love to understand what didn't click. Would you have 10 minutes for a call?"

You'll get a 10-15% response rate, which is more than enough. What you'll hear will be specific and actionable. "I couldn't figure out how to import my data." "It didn't integrate with the tool I already use." "The pricing didn't make sense for my team size." These aren't abstract product strategy problems. They're fixable issues that sit between your current product and PMF.

Kill your darlings

If a feature took you three months to build but nobody uses it, remove it. Every feature you maintain is cognitive load for your users and engineering load for your team. The bloat that accumulates during the search for PMF actively works against you. A lean product that does one thing well converts better than a sprawling product that does ten things adequately.

The "Hair on Fire" Framework

There's a mental model that's been circulating in the startup world for years, often attributed to various YC partners. The idea is straightforward: find people whose hair is on fire.

If someone's hair is on fire, they'll use any available solution to put it out. They won't comparison shop. They won't ask for a free trial. They won't need a beautiful UI or a seamless onboarding flow. The problem is so urgent that they'll use a half-finished, buggy, ugly product if it solves the problem right now.

This framework reframes the PMF search entirely. Instead of asking "is my product good enough?", you ask "is the problem I'm solving urgent enough?" If you have to convince people they have the problem, you're selling vitamins, not painkillers. And vitamins are a much harder business.

Here's how to apply this practically:

  1. List the top three problems your product solves. Be specific. Not "helps with productivity" but "reduces the time spent formatting sales proposals from 2 hours to 15 minutes."
  2. For each problem, ask yourself: who is currently losing sleep over this? Not who would find a solution nice to have. Who is losing actual money, actual customers, or actual hours every week because this problem exists?
  3. Go find those people. If you can't identify them specifically (by company, by role, by industry), the problem might not be urgent enough. If you can identify them, reach out directly and describe the problem, not your solution. See if they lean in.
  4. If they lean in, show them the product. If they start using it immediately, despite the rough edges and missing features, you've found a hair-on-fire problem. That's the foundation of PMF.

The companies that find PMF fastest are usually the ones that got the problem selection right before they wrote a line of code. They talked to enough people to find a problem so painful that the solution almost didn't matter. Then they built the solution and, predictably, people used it.

"The best startups tend to start with a problem that is so urgent for a specific group of people that those people will use almost anything to solve it. Then the product just has to be better than 'almost anything.'"

Putting It Together: A PMF Diagnostic You Can Run This Week

If you've read this far and you're still not sure where you stand, here's a practical checklist. Set aside three hours and work through it.

  1. Run the Ellis survey. Send it to users who've experienced your core value at least twice. Aim for 40+ responses. Calculate your "very disappointed" percentage.
  2. Check your retention curves. Pull your Day 1, Day 7, Day 14, and Day 30 retention. Does the curve flatten or does it decline toward zero? If it flattens, at what percentage? (Above 20% on Day 30 is encouraging for most B2B products.)
  3. Audit your growth sources. What percentage of new users come from organic channels (word-of-mouth, direct traffic, organic search) versus paid channels? If organic is above 50%, that's a strong signal.
  4. Look at your sales cycle. How many touches does it take to close a customer? Is the number trending down or up? A shortening sales cycle usually means the market is pulling your product, not the other way around.
  5. Ask five active users to describe your product. Do their descriptions match your positioning? If three out of five describe a different product than the one you think you're building, you have a fit problem.

If three or more of these checks come back positive, you likely have PMF or are very close. Double down on what's working. If fewer than two come back positive, stop spending money on growth and go back to the fundamentals: who is your customer, what is their most painful problem, and does your product solve it well enough that they'd be upset to lose it?

PMF isn't magic. It's the measurable result of building something that a specific group of people genuinely needs. The founders who find it are usually the ones who stopped guessing and started measuring. And when the measurements told them they didn't have it yet, they narrowed their focus instead of broadening their feature set.

That's the whole playbook. Run the diagnostics, be honest about what the data says, and keep narrowing until something clicks.