Free: The 7-Day Idea Test — validate any business idea in one week Get the Playbook →

Product Market Fit Examples: 5 Real Pivots Explained

11 min read
In this article

Slack was not supposed to be a chat app. It was the leftover of a failed video game.

Glitch, the game, shut down in December 2012 because too few people played it. The internal tool the team built to talk to each other while making it did not shut down. That tool became Slack, and by October 2014 Slack had raised $120 million at a $1.12 billion valuation, according to CNBC’s report on the funding round.

A 3D illustration of a checklist with green and red checkmarks against a light blue background

Product market fit is not a feeling. It is a moment you can usually point to: a support inbox that will not stop filling up, a feature nobody asked for that users adopt anyway, a survey answer that crosses a specific threshold. This article walks through five real examples, what signal told each founder they had found it, and how to read for the same signal in your own idea before you build further.

Two notes before you read on. First, every company below eventually built something very different from its first version. None of them found fit by polishing the original idea; they found it by noticing what the market did with what they had already shipped. Second, these are the companies whose pivots worked. The how to find product market fit question matters more than any single story, because most teams that pivot do not land on a second idea that scales.

What Signal Told Slack It Had Found Product Market Fit?

Internal usage that would not stop, inside a product nobody had built for outside users.

Product market fit, in Slack’s case, showed up as a tool built for internal use that the team could not imagine working without, long before a single outside customer saw it. Stewart Butterfield’s team built the messaging tool to coordinate while making Glitch. When the game failed, the tool remained in daily use. That retention, not the game, was the signal worth following.

Per CNBC’s account, the team had built the tool to communicate, share files, and search chat history while making Glitch. When the game failed, Butterfield and the team kept using the tool anyway, because they genuinely depended on it, and shut down the game to focus on it instead. That observation, surviving tool versus failing product, is the entire pivot.

What to notice. Nobody surveyed Slack’s internal team and asked if they liked the chat tool. The signal was behavioral: continued use after the reason for building it had disappeared. That is a stronger signal than a stated preference, because it costs the user nothing to say they like something and everything to keep using it when the original project around it is dead.

Your version. If you have built an internal tool, a workaround, or a side feature that your own team or early users keep reaching for even when it was not the point of the product, that is worth testing on its own. The Mom Test explains why behavior beats stated opinion when you are reading this kind of signal.

Can a Single Survey Question Measure Product Market Fit?

Yes. Superhuman built its entire product roadmap around one question and a specific answer threshold.

Superhuman measured product market fit by asking active users how they would feel if they could no longer use the product, and tracking the percentage who answered “very disappointed.” Rahul Vohra, the company’s founder, wrote on First Round Review that the team used Sean Ellis’s 40% threshold: if 40% or more of surveyed users say they would be very disappointed to lose the product, the company has product market fit.

Vohra’s account describes an early score below that threshold, in the 20s. Instead of treating that as a verdict, the team segmented respondents by role and followed up with the “very disappointed” group specifically, asking what they valued most and what would make the product even better for people like them. They used those answers to prioritize the roadmap toward the users already showing the strongest pull, rather than chasing everyone equally. Vohra reports the score eventually crossed the 40% line after several roadmap cycles built around that feedback.

A blank tablet screen with five yellow star icons above it, representing a rating or survey prompt

What to notice. The number by itself was not the useful part. The useful part was segmenting responses and acting only on the group that already loved the product, instead of trying to please the middle. A 20-something percent score with no segmentation is just a disappointing number. The same score, broken out by user type, is a roadmap.

Your version. A one-question survey costs you an afternoon. Ask your most active users the same “how would you feel” question, apply the 40% threshold loosely as a gut check rather than a pass/fail line at low sample sizes, and read the idea validation scorecard criteria on willingness to pay alongside it. A disappointment score with no spending behind it is still a soft signal.

Can Usage Data Alone Reveal Product Market Fit, Without Asking Anyone?

Yes, when one feature inside a broader product gets used heavily while the rest goes quiet. Instagram found its entire business this way.

Per Wikipedia’s account of the company’s history, Kevin Systrom and Mike Krieger built a check-in app called Burbn, closing a $500,000 seed round in March 2010 while developing it. Burbn let users check in at locations, make plans, earn points, and post photos, competing in a crowded field against Foursquare and similar apps.

Realizing Burbn was too close to Foursquare to win on check-ins, the founders looked at what people were actually doing inside the app. Photo-sharing was the feature users kept returning to. The check-in and planning features saw far less engagement. Systrom and Krieger stripped the app down to photos, comments, and likes, renamed it Instagram, and relaunched in October 2010. Facebook acquired the company for roughly $1 billion in April 2012.

What to notice. Nobody ran a focus group asking which Burbn feature people preferred. The founders looked at which feature people kept using inside a product built for an entirely different primary purpose, then rebuilt the whole company around that one feature and dropped the rest.

Your version. If you have shipped a multi-feature product, check which single feature gets used without you promoting it, reminding people, or asking for feedback. That is a stronger signal than anything collected through a feature request form. The validation signals before launch guide covers which behaviors are worth reading this way.

Before you build a second version of anything, the Idea Validation Scorecard runs your idea through the same 10 criteria these examples point back to: demand signals, willingness to pay, and reachability among them. Free. Takes 20 minutes.

What Happens When the Pivot Comes From Watching What Users Pay For, Not What They Say?

PayPal found its product by watching which of its own features actually moved money, and killing the rest.

According to the account in Eric Ries’s original MVP writeup and corroborated in multiple retrospectives of the company’s early history, PayPal launched with ambitions around beaming money between PalmPilot devices using infrared. That feature saw negligible use. What users actually adopted, heavily and without being asked, was the ability to send money by email, particularly on eBay, where buyers and sellers needed a way to pay strangers without sharing bank details.

The company’s early team did not decide this in a meeting. They watched transaction data and saw where volume actually flowed, then reallocated engineering effort toward the email-payment use case and away from the original device-to-device vision. The eBay-adjacent payment product became the business; the PalmPilot feature became a footnote.

An abstract network of glowing orange lines and nodes radiating outward on a dark background

What to notice. Nobody at PayPal asked users which feature they preferred. The signal was transaction volume: real money moving through one path and almost none through the other. That is the hardest kind of signal to argue with, because it requires the user to do something costly (send money) rather than say something free (answer a survey).

Your version. If you have shipped more than one feature or use case, look at which one people actually use for the thing that costs them something, whether that is money, time, or a referral to someone else. The validation signals before launch guide lists which early behaviors are worth weighting this heavily and which are not.

Why Did Airbnb Have to Go Door to Door Before It Found Fit?

Because the earliest signal, people booking the founders’ own apartment during a sold-out conference, was real but did not scale, and the team had to manually test whether the same pull existed city by city before trusting it as product market fit.

Per Wikipedia’s account of the company’s history, Brian Chesky and Joe Gebbia rented air mattresses in their San Francisco apartment to attendees of a sold-out design conference in October 2007, after noticing local hotels were fully booked. That single event generated bookings and a lesson: strangers would pay to stay in someone else’s home when hotel capacity ran out. It did not yet prove the idea worked anywhere else, for any other reason to travel.

Paul Graham’s essay “Do Things That Don’t Scale” describes what came next: the founders personally went door to door in New York, meeting hosts, helping them take better listing photos, and improving listings one at a time. Graham credits that unscalable, manual work, roughly a month of it, as the difference between a company that found real demand and one that assumed its first lucky weekend would repeat itself.

A forked path through a park with a signpost marking two directions

What to notice. The first signal (a sold-out conference) was real but narrow. The founders did not treat it as proof. They went and manually recreated the conditions in a second city, by hand, before concluding the pattern held generally. A single strong data point is not the same as a repeatable one.

Your version. If your idea has worked once, in one specific context, resist calling that product market fit. Manually repeat the same test in a second context, a second customer segment, or a second city, even if the repeat has to be done by hand. The one-week idea test is built around exactly this kind of fast, manual re-test.

How Do You Read for These Signals in Your Own Idea?

Look for a signal that costs the user something, repeats across more than one instance, and clusters around a narrower need than you originally built for.

Across all five examples, the real signal was never a compliment. It was retained usage after the original reason for building disappeared (Slack), a segmented survey answer crossing a set threshold (Superhuman), one feature getting used heavily while the rest of the product went quiet (Instagram), transaction volume concentrated in one feature (PayPal), or a result that held up when manually repeated in a second city (Airbnb).

A magnifying glass held over a printed page of bar charts and data tables

A four-step version for a solo founder reading for the same pattern:

  1. Identify what users are already doing with what you have shipped, not what they say they want next.
  2. Separate behavior that costs something (money, time, a referral) from behavior that costs nothing (a like, a compliment, a form fill).
  3. Look for concentration: is demand clustering around one narrow use, or spreading thin across many unrelated requests?
  4. If you have a strong signal from one source, test whether it repeats in a second context before trusting it.

Keep two limits in mind. First, every story here is a survivor. Thousands of pivots do not end with a company anyone profiles. Second, a signal that looks like fit at a small scale can still fail to hold at a larger one, which is why Airbnb’s founders re-tested manually in New York instead of assuming San Francisco would repeat itself everywhere.

To build your own read on this before committing further, the product market fit checklist turns these signal types into a structured pass.

Frequently Asked Questions

What is a real example of product market fit?

Superhuman’s survey method is the most measurable example. Founder Rahul Vohra tracked the percentage of active users who said they would be “very disappointed” without the product, using Sean Ellis’s 40% threshold as the benchmark, then rebuilt the roadmap around the segment already showing the strongest pull.

How do you know if you have product market fit?

Look for a costly signal that repeats: usage that persists after the reason you built something is gone, a segmented survey score crossing a set threshold, or transaction volume concentrated in one feature. A single compliment or one unrepeated success, like Airbnb’s first sold-out weekend, is not enough on its own.

Did Slack always plan to be a messaging app?

No. Slack began as an internal communication tool built while the team developed a video game called Glitch. When the game shut down in 2012, the team noticed they still depended on the internal tool daily, and that retained usage became the pivot into Slack as a standalone product.

What percentage of users need to want your product for product market fit?

Superhuman used Sean Ellis’s 40% threshold: if 40% or more of surveyed active users say they would be very disappointed without the product, that is read as a fit signal. Founder Rahul Vohra reported scores below that line early on, then rebuilt priorities around the segment already above it.

Keep Reading

Not sure if your idea has this kind of signal yet? Run it through the evaluation framework. No sign-up required.

What to Do Next

Choose the path that fits where you are right now.

Get the 7-Day Idea Test (Free)

Download the free 7-Day Idea Test. One task per day. Four evidence signals. One clear go, wait, or kill result — before you spend months building the wrong thing.

Send Me the Playbook

Choose Your Next Step

Find the path that fits where you are right now.

Get the Playbook

Start Reading

Read the step-by-step setup guide for your platform.

Free Download

Get the 7-Day Idea Test (Free)

Seven focused days. One task per day. By Day 7 you will know — with real evidence — whether your idea is worth pursuing. Free PDF, instant delivery.

Free. No spam. Unsubscribe anytime.