business / product-market-fit · article
business/ · 2,156 w · 10 min · ✎ dialogue

Product-Market Fit

Product-market fit (PMF) is the moment when a product stops being something founders push onto users and becomes something the market pulls toward itself — when customers ask to pay, invite friends, and call in panic when the site goes down. Drawing on Lenny Rachitsky's research across 25+ iconic companies, real pivot stories from founders like Bilal Mahmood (ClearBrain) and Alexander Embiricos (Remotion), and Jason Shen's own founder experience, this article covers what PMF looks and feels like, how long it actually takes, what it requires to be "true," and how to use usage patterns as a more reliable signal than satisfaction surveys.

What PMF Actually Feels Like

The term "product-market fit" is famously slippery. Marc Andreessen defined it as "being in a good market with a product that can satisfy that market," but founders often don't recognize it until after it's happened. Rachitsky's research across 25 iconic companies produced three recurring categories of recognition — not a single template.

1. Sudden and Significant Pull

About half of companies in Rachitsky's study found PMF through a sharp, unmistakable inflection. The common thread is that something starts working faster than the team can respond to it.

  • Dropbox: Drew Houston described his demo video hitting the top of Digg and Reddit as "like getting pressed into the back of your seat by a fast car or a plane taking off." Signups from strangers, not friends — that was the signal.
  • Netflix: Marc Randolph tried hundreds of variations for 18 months before testing the combination of no-due-dates, no-late-fees, and subscription. "Within days of testing it we knew we had a winner. Where before we were struggling to get traffic, all of a sudden we couldn't keep up. Our previously prodigious amounts of inventory were suddenly not enough."
  • Uber: Ryan Graves put it plainly: "Zero marketing budget and we were growing like a weed. Word of mouth was uncontrollable."
  • Stripe: Early users in the Ruby community started asking "Can we pay for this??" during a free beta. Tom Preston-Werner: "They liked it so much they wanted to pay for it. That was the first sign this was going to work."
  • Robinhood: Vlad Tenev woke up on a Saturday to 600 concurrent visitors on a site nobody knew about yet. "I'm just screenshotting the page; I'm calling my parents saying, 'Oh, this is crazy. It might actually be working.'" 10,000 sign-ups that first day.
  • CartaHR (Carta): Henry Ward's team misconfigured their phone system for three weeks, routing all sales calls to dead air. A prospect finally reached them via chat: "I've been trying to get to your sales people for weeks and you keep hanging up on me!!! Why won't you take my money???"

The last example is worth sitting with. PMF here was visible not despite the broken process, but because of it — demand was so strong it found a crack.

2. Steady and Compounding Pull

For products with a narrower initial market — high-end tools, niche B2B, or community-dependent products — PMF accumulates rather than erupts. You don't get a single "wow" moment; you get a series of growing confirmations.

  • Segment: Calvin French-Owen described the shift as moving from push to pull. "Before we had any sort of fit, it always felt like we had to push our ideas on other people. When we hit PMF, we started feeling 'pull' for the first time. We solved one problem for people... so they started asking us to solve a second, third, fourth, and fifth."
  • Superhuman: Three years of building. Rahul Vohra's moment came at a conference when he checked the analytics dashboard his growth head had prepped: "Every single metric was amazing — growth, DAU/MAU, PMF score, NPS, virality, CAC payback, CAC:LTV, activation rates, long-term retention rates." Every metric hitting simultaneously was the signal.
  • Instacart: Max Mullen's framing captures the gradient well: "Initial signs of product-market fit feel a bit like a calm breeze, while true product-market fit feels like a powerful wind at your back, accelerating you forward and compounding over time." Early PMF existed with a small segment (people who wanted instant delivery from any store); real PMF came after forming retailer partnerships and hitting a larger, more natural customer segment.
  • Nextdoor: Sarah Leary knew they had it when an engineer took the service down at noon for an hour without advance notice. "About ten minutes after taking the servers offline, I started getting emails and phone calls from concerned users." Less than a few hundred users, and they panicked. That's the signal — not size, but urgency.
  • PagerDuty: Two years, consistent >10% month-over-month growth, solving a "hair on fire" problem (critical incidents falling through the cracks). Alex Solomon: "By the middle of 2011, we felt pretty confident we had PMF."
  • Substack: Chris Best described the transition as PMF anxiety slowly being replaced by execution anxiety: "Gradually the how-do-we-keep-up anxiety got bigger and bigger until there wasn't time left in the day to worry about whether we had product-market fit."

3. Hitting a Meaningful Milestone

For some companies, PMF crystallized around a single symbolic proof point — not a metric, but a moment that stood in for the whole.

  • Airbnb: Joe Gebbia: "When my mom booked her first Airbnb, I said to myself, I think we got something here." That "my mom" test — a non-tech-enthusiast voluntarily using the product — meant it had crossed a chasm.
  • Product Hunt: Ryan Hoover was walking into a coffee shop and overheard strangers talking about their upcoming launch on Product Hunt. "Seeing a stranger use what we've built unprompted might seem trivial but it's an impactful moment."
  • Amplitude: 1,000+ product managers showed up to their first conference. While standing in the badge line, a team from a large bank was overheard: "I've been in Product for 10 years and this is the first time a company has ever focused on us."

How Long PMF Takes

Lenny Rachitsky collected PMF timelines from marketplace and non-marketplace founders alike and found:

  • Average time from launch to PMF: ~1.5 years. Building v1 often takes another year or more before launch. Expect to put in at least two full years before you feel "this is working."
  • Marketplace PMF takes roughly the same time as non-marketplace PMF. The advantages (network effects, cost efficiency) don't compress the timeline meaningfully.
  • Local marketplaces (Lyft, DoorDash) often found early PMF in city 1 but only felt confident after replicating it in city 2. Lyft's LA launch "didn't have a single ride request for hours" after a successful SF launch. They had to redesign their marketplace mechanics before it worked outside their home city.
  • Every successful marketplace started hyper-concentrated — single city or single category, no exceptions.
  • Thumbtack took five years: two years building supply, two years building demand, one year developing a revenue model. All three had to come together before the flywheel started spinning.

A counter-intuitive finding: most founders are not simply impatient — the underlying work often genuinely takes this long. The companies that seem to find PMF fast (Caviar: 3 months; Lyft: immediate in SF) typically had the benefit of immediately visible network effects or a pre-existing, deeply frustrated user base.

True PMF Requires Three Things Simultaneously

Feeling product-market fit is not the same as having it durably. Rachitsky and others are explicit: there are three conditions, and all three must hold.

  1. A product people want — the most obvious and the most commonly discussed.
  2. Profitable delivery at scale — companies like Cherry (on-demand car washing) and Luxe (valet parking) had genuine demand but could not deliver the service profitably. They failed not for lack of product love but for lack of viable unit economics.
  3. Sustainable acquisition and retention — finding new customers costs something; keeping them costs something else. If either is broken, the business eventually runs out.

Investors and founders sometimes confuse strong engagement with PMF. ClearBrain's story is instructive: Bilal Mahmood and his team found a real audience for their ML-as-a-service product, landed an Airbnb contract worth $500k, and felt they had PMF. But growth stalled just under $1M ARR because the product wasn't used frequently enough to be essential. "It doesn't matter if you give someone a million dollar insight they get once a month, versus a $2 insight they get every 10 minutes. Build something that is repeated in use." Their third pivot — from ML-generated insights to a real-time visual analytics platform (where the predictive/causal capabilities were an embedded bonus) — produced both regular usage and eventually an acquisition by Amplitude.

The principle: a high-value, low-frequency tool is not PMF. If customers must consciously remember to use it, it won't survive competitive pressure.

PMF Dimensions: Fit Quality × Market Breadth

The intensity of the pull you experience depends on two variables:

  • Quality of fit: How much better is this than existing alternatives? A 10x improvement produces a stronger signal than a 2x improvement.
  • Initial market breadth: How many people immediately experience the fit?

A 10x product in a large, undifferentiated market (Dropbox, Netflix, Tinder) produces sudden, broad, unmissable pull. A 10x product for a narrow segment (Superhuman, Substack for newsletter writers, Remotion for engineering teams) produces steady, compounding pull that can look like nothing for months before it becomes obvious.

This distinction matters because many founders misread slow compounding pull as a signal that the product isn't working, when it may simply be working for a niche they haven't yet fully identified. The Remotion team's PMF survey process is worth studying: by running the Sean Ellis survey ("how would you feel if you could no longer use Remotion?"), segmenting results by team type, and crossing that with usage data, they found that engineering teams mixing direct calls and coworking rooms had a 60% "very disappointed" score — well above the 40% threshold — while the broader user base was at 39%. Their PMF was real, but concentrated.

The Regular Usage Test

Before celebrating PMF, ask: how often are users returning, and is it driven by habit or emergency? The ClearBrain insight generalizes widely: if customers use your product monthly because it's valuable but not daily because it's essential, you are vulnerable.

From the Remotion experience: "We're always an afterthought. It has to be the primary tool the customer used every day." Remotion's most engaged users were those who had built it into their workflow for quick, lightweight engineering conversations — not those who used the social/coworking features on a more casual basis.

A related diagnostic: track not just whether users return, but what triggers their return. Is it a scheduled reminder (weak), a workflow integration (stronger), or an unprompted habit (strongest)? The Nextdoor test — users calling in a panic within 10 minutes of an outage — is the extreme version of this. The signal isn't engagement metrics; it's urgency.

Measuring PMF: The Sean Ellis Survey

The most operationalized approach to measuring PMF before it's obvious is the Sean Ellis / Superhuman framework: survey users who have used the product at least twice in the past two weeks with the question: "How would you feel if you could no longer use [product]?" If 40% or more answer "Very disappointed," you have PMF for that segment.

Key caveats:

  • Survey the right users (active users, not sign-ups or churned users).
  • Don't over-index on the overall score — segment by role, team type, use case, and look for where the score spikes above 40%.
  • Pair the quantitative result with qualitative follow-ups to understand why those users are disappointed and what specifically they'd lose.

What PMF Doesn't Guarantee

PMF does not guarantee a successful business. The full stack of requirements:

  • PMF (people want it)
  • Unit economics (you can deliver it profitably at scale)
  • Distribution (you can find and keep these people sustainably)

Ridejoy, Jason Shen's YC-backed rideshare marketplace, had real user behavior and genuine engagement. But when Craigslist — their primary traffic source — sent a cease-and-desist in August 2012, the entire supply and demand flywheel collapsed overnight. Distribution was a single point of failure, and when it failed, PMF became irrelevant. The lesson: distribution is part of PMF, not separate from it.

Thread · 0 replies+ add reply
no replies yet — be the first to write back to this article.