The Question Superhuman Asked 100 Users Before It Found Product-Market Fit

In April 2017, Rahul Vohra sent a survey to Superhuman's users. The company had about 100 paying customers at 30 dollars a month, a product two years into development, and a founder who couldn't tell his team whether any of it was working.

The first question decided the mood of the quarter: how would you feel if you could no longer use Superhuman? Twenty two percent said very disappointed.

Vohra had borrowed the question from Sean Ellis, who coined the term growth hacking. Ellis benchmarked it across more than 150 companies and found the same threshold each time: below 40 percent very disappointed, growth is a grind, and above it things start pulling. Slack scored 51 percent.

Twenty two percent is a bad number without being a dead one. Most customer discovery stops right there. A founder reads a mediocre score, feels bad about it, and goes back to building whatever the roadmap already said.

Vohra went at it differently. He stopped reading the answers like a researcher and started reading them like a recruiter. He pulled out only the people who said very disappointed and read what they said next: what type of person would most benefit from this product?

People who love a product describe the benefit in the words that matter most to them, and that sentence becomes the shape of a customer. The segment he found was a person with a name and a job. Nicole, a founder or executive living in her inbox, reading a hundred to two hundred emails a day, proud of how fast she replies.

Vohra took the frame from Julie Supan's branding work and called her the high expectation customer. Then came the discipline that makes the engine worth copying. The people who said they wouldn't be disappointed at all, which was most of the sample, got no vote on the roadmap.

The people who loved the product shaped it instead: speed, keyboard shortcuts, the feeling of flow. What remained was the useful middle. Users who said somewhat disappointed but named the same main benefit as the people who loved it, which meant they were close and one specific thing was holding them back.

For Superhuman that thing was a mobile app, plus a long tail of smaller objections that no internal product meeting would have surfaced. The roadmap split 50/50. Half the resources went to more of what the very disappointed already loved, half went to systematically removing what blocked the somewhat disappointed.

After the segmenting exercise the score jumped to 33 percent. A quarter later it hit 47, then 56, then 58, where it flattened into a healthy curve.

None of that ran on intuition. It ran on a corpus. Superhuman logged roughly 12,000 pieces of verbatim user feedback and kept every one of them in a CRM.

When it came time to build calendaring, Vohra asked the product manager to read the raw sentences for hours until they were all in his head, then write the spec from there. His reason for that ritual is the part most discovery advice skips. Talking to users in person, people are polite, while thousands of their sentences read with the distance of a screen carry what they actually need.

Four questions, sent on a schedule, answered in the customer's own words, and kept in one place until a roadmap decision needs them. The benchmark is the easy part, and any founder can run it this week. The record is the half that turns a survey into an operating habit, and it is the half founders skip.

Vohra's own rule for a score under 15 or 20 percent: do the segmenting exercise first, and if the number holds, the product or the market is wrong. Learning that inside a quarter beats learning it a year later.