Growth · Featured Debate

4 guests 8 episodes 3,291 words

The PMF Measurement Problem: Eight Voices, One Question, No Easy Answer

How do you know when you have product-market fit?

Every founder has heard "you'll know product-market fit when you feel it." Eric Ries, the creator of the Lean Startup methodology, put it most bluntly on Lenny's Podcast:

"People call me sometimes they're like 'I want to know if I have product market fit.' Well, the fact that you have time to call me and ask me this question already tells me the answer. If you have had product-market fit you would not have time for this. When it's working there's no time for naval gazing." -- Eric Ries ▶ 1:13:49

It is a satisfying line. It is also, for most founders, completely useless. The problem is that PMF is not a single moment -- it is a signal that arrives through different channels at different times, and optimizing for the wrong signal can lead you catastrophically astray. Sean Ellis says survey your users. Naomi Gleit says watch your retention curves. Sarah Tavel says count core actions. Todd Jackson says PMF progresses through four levels. Hamilton Helmer says none of it matters without a moat. Elena Verna says it now expires every three months. Nikita Bier says you will just know.

They are all right. They are measuring different things at different stages. The real skill is knowing which lens to use when.

Your startup has some traction. Users are signing up, some are sticking around, a few are enthusiastic. How do you actually know if you have achieved product-market fit, and which measurement approach should you trust?

The 8 Positions

1 / 8

The "Very Disappointed" Survey -- Your Fastest Leading Indicator

Sean Ellis

Sean Ellis, creator of the PMF survey and former Head of Growth at Dropbox and Eventbrite, developed the most widely adopted PMF diagnostic in the startup world. The test is deceptively simple: ask users "How would you feel if you could no longer use this product?" If 40% or more say "very disappointed," you have signal.

But Ellis himself is careful about what this signal means:

"I would say it's a leading indicator of product market fit. The lightning indicator is, do they actually keep using it? So probably retention cohorts are more accurate, but the problem is, how long do you have to look at a retention cohort before you know that you've actually long-term retained someone?"
Sean Ellis
▶ 00:06:40

The survey's value is speed. You can run it on day one with an MVP, without analytics infrastructure. The 40% threshold was not designed from theory -- it emerged empirically from benchmarking dozens of YC-backed startups. Ellis ran the survey at a company after leaving Dropbox and got back just 7% "very disappointed," demonstrating how the test can deliver harsh truths early.

But the real insight is not the 40% number. It is what to do with the data. Ellis's most counterintuitive advice:

"Just ignore the people who say they'd be somewhat disappointed. They're telling you it's a nice to have. If you start paying attention to what your somewhat disappointed users are telling you and then you start tweaking onboarding and product based on their feedback, maybe you're going to dilute it for your must have users."
Sean Ellis
▶ 00:00:25

This is the trap most teams fall into. They see 30% "somewhat disappointed" and think: if we fix a few things, we can convert them. But optimizing for the lukewarm middle can dilute the experience for the people who already love you. The job is to understand what your "very disappointed" users love and double down on that, even if it means alienating the middle.

Ellis also observed that retention problems are usually onboarding problems in disguise:

"It's usually much more function of onboarding to the right user experience than it is about the kind of the tactical things that people try to do to improve retention."
Sean Ellis
▶ 00:00:45

And perhaps most reassuringly for founders who worry about losing PMF:

"I haven't seen it really fade back down, but I've seen companies fail despite having it. And I think a lot of times then, it becomes an execution challenge."
Sean Ellis
▶ 00:32:07

The survey measures something real and durable. But it is necessary, not sufficient.

When This Applies

Essential in AI-era products where both technology and market shift quarterly. Ship continuously. Re-validate continuously. The team that finds PMF and the team that scales it must be the same team.

Operationalize the Survey as a Cultural Gate

Sean Ellis · Jag Duggal

Jag Duggal, Chief Product Officer at Nubank (100M+ customers, formerly at Facebook and Google), takes the Ellis survey one step further. At Nubank, the score is not a one-time diagnostic. It is a mandatory product review metric that gates every launch.

"We are amongst the most fanatical followers of his methodology anywhere in the world. We rarely scale a project, a product we've launched, until we know the Sean Ellis score and we know that it's hit a threshold that we find really compelling."
Jag Duggal
▶ 00:15:05

Critically, Nubank adapted the threshold. Brazilians are culturally more optimistic and polite, so the standard 40% was too easy to hit:

"Brazilians are a culturally happier group than the global average. So for us, our threshold isn't 40%, we've generally moved it up to 50%."
Jag Duggal
▶ 00:16:09

The deeper lesson: universal thresholds are a myth. The 40% is a starting point, not gospel. Nubank's lead designer coined the mantra that now drives their product culture: "Good enough isn't good enough. Is it great enough?" Combined with NPS scores that range from the 70s to an almost unheard-of 94-95 in Mexico, the Ellis survey becomes one instrument in a quality system, not a standalone verdict.

When This Applies

Use when you have 30-100 users and need a fast, cheap signal. Weight it toward users who have experienced the full product. Calibrate the threshold for your market and culture. At Nubank, the threshold is 50%, not 40%. Operationalize it as a gate, not just a check.

Retention Curves -- Watch Behavior, Not Sentiment

Naomi Gleit

Naomi Gleit, Head of Product at Meta and the company's 29th employee, helped pioneer the growth team model that reshaped how Silicon Valley thinks about product-led growth. Facebook's growth team discovered that retention, not acquisition, was the biggest lever:

"What we found was the churn in resurrection lines were actually much larger than the new user line, which implied to us that retention and driving those two lines was actually our biggest lever to drive net growth."
Naomi Gleit
▶ 00:26:30

The famous "7 friends in 10 days" metric? It was never a magical insight:

"Seven friends in 10 days was a thing. 10 friends in 14 days was also a thing. They're the same thing, they're just different points on a retention curve."
Naomi Gleit
▶ 00:26:30

Gleit is refreshingly honest about the limits of her own team's contribution:

"As much credit as we give to the growth team, I'm actually not sure how much credit we deserve and how much incremental growth we drove above and beyond the fact that this was a product that had product-market fit."
Naomi Gleit
▶ 00:38:49

This humility reveals a critical truth: retention curves confirm PMF, but they cannot create it. And they require scale -- thousands of users across multiple months -- to produce statistically meaningful results, making them a lagging indicator rather than a fast diagnostic.

When This Applies

Use when you have thousands of users across multiple months. This is your proof metric. But remember Gleit's own humility: retention curves confirm PMF, they do not create it.

Core Action Completion -- The Behavioral Foundation

Sarah Tavel

Sarah Tavel, General Partner at Benchmark and the first product manager at Pinterest, developed the hierarchy of engagement framework from her experience leading Pinterest's discovery team. Her argument: MAU is a vanity metric. What matters is whether users complete the product's core action.

"When a user completes this action, it's clear that they both understand the utility of the product, they understand what that product is all about, and it's an action that, if they perform the action, they're very likely to come back."
Sarah Tavel
▶ 00:05:45

The hierarchy has three levels. Level 1: users complete the core action (pinning for Pinterest, friending for Facebook). Level 2: the product gets better the more you use it, and you have more to lose by leaving:

"The test for me, of whether you're building a product that has the ingredients to create a retentive product on a micro level, just at the user level, is that the product should get better the more you use it, and you'll have more to lose by leaving it."
Sarah Tavel
▶ 00:11:03

Level 3: network effects make the product self-perpetuating. Picking the right core action requires real discipline. Pinterest experimented with follows, clicks, likes, pins, and time-on-site before settling on pinning. Getting this wrong means optimizing for the wrong outcome entirely.

PMF Progresses Through Four Levels

Todd Jackson

Todd Jackson, Partner at First Round Capital with deep operating experience across Gmail, Facebook News Feed, Twitter, and Dropbox, argues that PMF is not binary. It progresses through four levels -- nascent, developing, strong, extreme -- measured across three dimensions: demand, satisfaction, and efficiency.

"We have found a very consistent set of patterns, demand satisfaction, and efficiency. But the interesting thing is that you don't go for all three of them from the very beginning."
Todd Jackson
▶ 00:00:11

The sobering data from studying hundreds of portfolio companies:

"Roughly, 60% are never going to get past L2."
Todd Jackson
▶ 00:00:27

When founders are stuck, Jackson points to the "four Ps" -- persona, problem, promise, and product -- as the levers to change:

"You've got the persona, the problem, the promise, and the product. Lattice kept the first one but changed the others. Vanta changed all four."
Todd Jackson
▶ 00:00:34

This framework resolves a core tension in the PMF conversation. The "you'll know it when you see it" crowd and the "iterate for years" crowd are both right -- they are describing different levels. Nascent PMF with 3-5 customers does not feel like a whirlwind. Extreme PMF at scale absolutely does.

When This Applies

Use as a diagnostic at any stage. Know your level (nascent, developing, strong, extreme). Know which dimension to optimize (satisfaction first, then demand, then efficiency). When stuck, examine the four Ps: persona, problem, promise, product.

PMF Without Power Is a Castle Without a Moat

Hamilton Helmer

Hamilton Helmer, author of 7 Powers (credited by Patrick Collison, Peter Thiel, Reed Hastings, and Daniel Ek as foundational to their strategic thinking), originally believed founders should find PMF first, then worry about strategy. He changed his mind:

"Before I thought it was you do product market fit and then you do strategy, and if you try and put strategy before product market fit, it is not much you can do with it. That's wrong."
Hamilton Helmer
▶ 00:04:38

Now he argues that even pre-PMF, founders should ask: of the business propositions I am considering, which have characteristics that tilt toward the availability of power?

"Even before you have product-market fit, it's worth thinking about strategy. The questions are of the business propositions that you're thinking of in trying to get to product-market fit, what are the underlying characteristics that might tilt them towards the availability of power or not?"
Hamilton Helmer
▶ 00:05:41

Speed alone is not power:

"You're on a treadmill and if you stop running on that treadmill, you get creamed, but it's not power. The things that drive operational excellence can be mimicked."
Hamilton Helmer
▶ 00:00:20

When This Applies

Use once you have identified your product's core action and have enough users to see patterns. This is your ongoing health metric. For Pinterest, it was pinning. For Facebook, friending. For Airbnb, booking. If you cannot name your core action, that is your first problem.

In Consumer, PMF Is Binary and Unmistakable

Nikita Bier

Nikita Bier, who built two #1 App Store apps (TBH, acquired by Facebook, and Gas) with a team of four people, represents the visceral end of the spectrum:

"A lot of people ask me like, 'What's the benchmark for product-market fit?' And this founder that I'm friends with, his name's Roger Dickey, he told me this one time, 'If your product's working, you'll know. And if there's any uncertainty, it's not working.'"
Nikita Bier
▶ 00:36:06

When TBH hit PMF, standard metrics were too coarse to capture what was happening:

"Our metric was hourly actives per day. Not daily active users, hourly active users."
Nikita Bier
▶ 00:36:06

This is the view from the extreme tail: viral consumer social apps where PMF announces itself through chaos -- servers breaking, geofencing to slow growth, customer support systems failing every three days. But as Lenny's own research shows in his newsletter "What it feels like when you've found product-market fit," about half of the 25 iconic companies he studied found PMF immediately, while the other half spent months or years iterating (Netflix: 18 months, Airbnb: 2 years, Superhuman: 3 years, Amplitude: 4 years). The binary view is real for some companies and misleading for others.

When This Applies

Trust your gut for consumer products. If you have to ask, the answer is probably no. But do not over-index on this for B2B, marketplace, or infrastructure products where PMF curves are slower.

PMF Is Now a Treadmill, Not a Destination

Elena Verna

Elena Verna, Head of Growth at Lovable ($200M ARR in under one year with fewer than 100 people -- one of the fastest revenue ramps in software history), argues that the entire concept of PMF has fundamentally changed in the AI era:

"The rate of that evolution was measured in years. What is it that you need the next product-market fit step function change, which often was called second horizon or third horizon. Sometimes five, 10 years. Now, it's three months, and all of a sudden you have to face that question again."
Elena Verna
▶ 1:00:49

Both the technology (new LLM model releases) and consumer expectations shift quarterly. Even a $200M company is not safe:

"Every three months, I feel like we have to recapture our product-market fit and not just recapture on the same technology and with same customers. It's both of those pieces of the equation change every three months, and it's terrifying in a way."
Elena Verna
▶ 1:03:43

The organizational implication is profound:

"The team that finds your product-market fit is very different than the team that usually scales your company, yet we have to find the team that is capable of doing both on ongoing basis. Now, I think every AI company is on this product-market fit treadmill."
Elena Verna
▶ 1:04:26

Even OpenAI lost roughly 6% market share in a single week when Gemini 3 launched. Nobody's future is bulletproof.

When This Applies

Start thinking about this from day one. Not full strategic planning, but asking: does this path tilt toward defensibility? At Lovable's scale, even with $200M ARR, the question of power remains open.

Evidence from the Archive

GrowthHackers

YC-backed companies in Silicon Valley formed the original benchmarking cohort that produced the 40% threshold

Ellis ran the survey at a company post-Dropbox and got 7% 'very disappointed,' demonstrating how the test can surface harsh truths early

Sean Ellis · Author of Hacking Growth, creator of the PMF survey, former Head of Growth at Dropbox and Eventbrite ▶ 00:03:19

The '7 friends in 10 days' and '10 friends in 14 days' metrics were equivalent points on the same retention curve

Facebook's growth team was composed of engineers and PMs, not marketers, instrumenting every step of the user journey

Naomi Gleit · Head of Product at Meta, 29th employee at Facebook, longtime growth team leader ▶ 00:26:30

First Round Capital

The best enterprise companies reach extreme PMF in roughly 4-6 years

Lattice kept its persona but changed its problem, promise, and product to find PMF

Todd Jackson · Partner at First Round Capital, former product lead at Gmail, Facebook News Feed, Twitter, and... ▶ Watch

Nubank

Nubank recorded an NPS of 94-95 in Mexico during the first couple of years after launch

When NPS dips even one or two points, it triggers alarm and investigation

Jag Duggal · Chief Product Officer at Nubank, formerly at Facebook, Google, and Quantcast ▶ 00:15:05

Benchmark

Pinterest experimented with follows, clicks, likes, pins, and time-on-site before settling on pinning as the core action

Pinterest shipped a 'picked for you' algorithmic feed that made the experience better the more users pinned, creating Level 2 retention

Sarah Tavel · General Partner at Benchmark, former first PM at Pinterest, previously at Greylock ▶ 00:05:45

Strategy Capital

AWS, Apple's iPhone, and Intel's CPUs are examples of iconic second acts that required restarting the PMF-to-power cycle

Warren Buffett's metaphor of 'economic castles protected by unbreachable moats' captures the benefit + barrier requirement

Hamilton Helmer · Author of 7 Powers, strategy advisor to companies like Netflix and Spotify ▶ 00:04:38

The Synthesis

These eight lenses are not competing -- they are sequential and contextual. Lenny's own newsletter work synthesizes the picture. In "How to kickstart and scale a consumer business -- Step 5: RETAIN," he identifies four signs of PMF as a confidence spectrum, not a binary: (1) flattening retention curves, (2) explosive organic growth, (3) the Sean Ellis survey above 40%, and (4) visceral excitement. In "What is a good activation rate," his survey of 500+ companies shows that activation rate (the moment a user first completes the core action, connecting Tavel's framework to Gleit's retention curves) averages just 34% across all products.

Sequential Lenses

Why do eight different PMF measurement approaches all seem valid?

Stage-Appropriate Tools

What is the critical mistake in measuring PMF?

Confidence Spectrum

Is product-market fit binary or a spectrum?

The eight lenses are not competing -- they are sequential and contextual. Pre-product: use Helmer's power lens. MVP with 30-100 users: the Ellis survey. Scaling to thousands: core action framework. At scale: retention curves. Continuously: re-measure regularly as PMF can decay.

The critical mistake is using the wrong lens at the wrong stage. If you have 50 users and you are trying to analyze retention cohorts, you are wasting time. If you have 500,000 users and you are still relying solely on surveys, you are leaving rigorous data on the table.

PMF is a confidence spectrum, not a binary: flattening retention curves, explosive organic growth, the Sean Ellis survey above 40%, and visceral excitement. Know which level you are at -- nascent, developing, strong, extreme -- and optimize the right dimension at each stage.

Which Approach Fits You?

Answer 3 questions about your situation. We'll match you to the right approach.

Question 1

How many users do you have?

Question 2

How does your market evolve?

Question 3

What is your biggest measurement challenge?

Notable Absences

The Bottom Line

**Always:** Jackson's four-level model as a diagnostic. Know which level you are at (nascent, developing, strong, extreme) and optimize the right dimension (satisfaction first, then demand, then efficiency).

Sources

Sean Ellis — "Watch on YouTube" — Sep 5, 2024
Jag Duggal — "Watch on YouTube" — May 16, 2024
Naomi Gleit — "Watch on YouTube" — Oct 27, 2024
Sarah Tavel — "Watch on YouTube" — Dec 27, 2023
Todd Jackson — "Watch on YouTube" — Apr 11, 2024
Hamilton Helmer — "Watch on YouTube" — May 5, 2024
Nikita Bier — "Watch on YouTube" — Aug 25, 2024
Elena Verna — "Watch on YouTube" — Dec 18, 2025
Eric Ries — "Watch on YouTube" — Oct 29, 2023