Superhuman's PMF survey: how Rahul Vohra's 40% rule became the most-cited product-market-fit framework in B2B SaaS
In 2018, Superhuman founder Rahul Vohra published a First Round Review essay titled 'How Superhuman Built an Engine to Find Product/Market Fit.' The framework: regularly survey users with the question 'How would you feel if you could no longer use this product?' and treat 40% of users answering 'very disappointed' as the threshold for product-market fit. The essay went viral in startup circles and the methodology — usually called the Sean Ellis test after the growth marketer who developed the original survey — became the most-cited PMF measurement framework in B2B SaaS. The Superhuman case study is interesting not just for the methodology but for the practical product-development implications the survey drove inside Superhuman itself.
- Story: In May 2018, Superhuman founder Rahul Vohra published a First Round Review essay describing how Superhuman used the Sean Ellis PMF survey (the question: how would you feel if you could no longer use this product?) to systematically improve product-market fit. Superhuman started at 22% 'very disappointed' (below the 40% threshold) and iterated to 58% over several quarters by segmenting respondents, doubling down on what the very-disappointed segment loved, and fixing friction for the somewhat-disappointed segment.
- Why it matters: Superhuman's essay popularized the Sean Ellis PMF survey methodology in B2B SaaS startup circles. The 40% threshold became the most-cited PMF benchmark of the late 2010s and 2020s.
- Takeaway: Segment survey respondents by attachment level (very/somewhat/not) and treat each segment differently.
- Takeaway: Double down on what the very-disappointed segment loves; ignore feedback from the not-disappointed segment.
- Takeaway: Trajectory matters more than the absolute threshold; if the score is improving, you're earning the right to scale.
Superhuman PMF survey — the four-step story
Superhuman PMF survey at a glance
Quick facts
The methodology: Sean Ellis test, applied at Superhuman
The core survey asks current users one question: How would you feel if you could no longer use [product]? Options: very disappointed, somewhat disappointed, not disappointed. The percentage answering 'very disappointed' is the PMF signal. Growth marketer Sean Ellis observed across his consulting work that products above 40% on this metric tended to scale efficiently and products below 40% struggled to grow even with paid acquisition. Hence the 40% threshold.
Superhuman applied the survey systematically. When Vohra ran the survey early in Superhuman's development, only about 22% of users said they would be very disappointed. That score was below the PMF threshold. Rather than ignore the signal, the team treated it as a diagnostic and built a focused process to improve it.
The diagnostic process: segment, prioritize, build
What made Superhuman's PMF process distinctive was the second-order analysis. After running the basic survey, the team segmented respondents by their answer:
- 'Very disappointed' respondents (the PMF-positive segment) — the team studied what they loved about the product, what use cases they used it for, and what segments they came from.
- 'Somewhat disappointed' respondents (the on-the-edge segment) — the team studied what was holding them back from being very disappointed. These users often loved part of the product but had specific friction points that made them unwilling to commit.
- 'Not disappointed' respondents (the not-PMF segment) — the team de-prioritized features these users requested, on the logic that their feedback would move the product away from the segment where PMF existed.
- The framework: double down on what the 'very disappointed' segment loves, fix what's blocking the 'somewhat disappointed' segment, ignore the 'not disappointed' segment.
Iterating to 58% and what that proved
Over several quarters of focused iteration on the high-conviction roadmap, Superhuman improved the survey score from 22% to approximately 58%. The methodology became the company's North Star metric for product-development prioritization. The essay describing this process became one of the most-shared startup essays of 2018-2019 and is required reading at many YC batches.
The business outcome at Superhuman: the company has continued to grow and has been profitable enough to raise on its own terms through 2021 (Series C at $825M valuation). The product remains positioned as premium email for professionals, with a $30/month price point that's substantially above mass-market email. Whether Superhuman's PMF process produced its current scale, or whether the company would have grown anyway, is harder to disentangle — but the methodology has been replicated by many other startups.
Criticisms and limitations
The Sean Ellis test and the 40% threshold have been criticized on several grounds. First, the 40% benchmark is empirical and may not generalize across product categories — viral consumer products and slow-burn enterprise products may have different appropriate thresholds. Second, the survey response rate matters: a low response rate can mean the people answering aren't representative. Third, the survey measures current users' attachment, not market-size or growth-trajectory potential — you can have 40% PMF in a tiny addressable market.
More substantively, some growth marketers argue that the methodology can be misleading for products with strong network effects or platform dynamics, where individual-user attachment isn't the right signal. For example, a B2B SaaS used by analysts whose company will eventually consolidate vendors may have high individual-user attachment but low business-level PMF. The methodology is useful as one signal among many, not as the sole determinant of whether to scale.
How RGM thinks about PMF measurement
We've used the Sean Ellis test with multiple clients across SaaS and consumer categories. The methodology's real value is less in the 40% threshold itself and more in the discipline it enforces: continuously surveying current users, segmenting by attachment level, and prioritizing the roadmap toward the high-attachment segment.
The honest framework: use the Sean Ellis test as a quarterly diagnostic, segment respondents by their 'somewhat' vs 'very' vs 'not' answers, study the very-disappointed segment for what to amplify, study the somewhat-disappointed segment for what to fix, and ignore the not-disappointed segment to preserve product focus. The Superhuman essay is the worked example. The 40% threshold is heuristic, not a hard line — what matters is the trajectory and the discipline. If your score is improving quarter-over-quarter, you're earning the right to scale; if it's flat or declining despite product investment, something deeper is wrong with the product or the segment selection.
Frequently asked questions
Where did the Sean Ellis test originate?
Sean Ellis developed the survey question in the mid-2000s while working as a growth consultant. He used it at companies including Dropbox, LogMeIn, Eventbrite, and others. The 40% threshold was Ellis's empirical observation that companies above 40% on the metric tended to scale efficiently and companies below 40% didn't. The methodology pre-dates Superhuman's adoption by roughly a decade.
Has the methodology been validated by independent research?
There is no large-scale academic validation. The empirical evidence is anecdotal across the dozens of companies Ellis and Vohra and others have publicly described using it. The methodology is widely adopted in startup practice but has not been rigorously validated against neutral measures of long-term product-market fit. Anyone using it should treat it as a useful but imperfect signal.
Does the Sean Ellis test work for B2B SaaS with low end-user attachment?
Less well. For products where the buyer is different from the end user (HR tools where managers buy and employees use; observability tools where engineers want them and CFOs approve them) the survey of end users captures something but may not predict business outcomes. The methodology is most reliable for products where end-user enthusiasm directly drives renewal and expansion.
What's Superhuman's actual business performance?
Superhuman is private and does not publicly disclose detailed financials. The $825M Series C in 2021, ongoing operations, and apparent ability to retain a $30/month price point suggest the business is healthy. The company has not been publicly disclosed to be profitable but has not raised since 2021, which usually suggests the company has sufficient cash to operate.
Are there alternatives to the Sean Ellis test?
Yes. NPS (Net Promoter Score), retention cohort analysis (D30/D60/D90 retention curves), expansion-revenue metrics (NRR, GRR), and product-engagement metrics (DAU/MAU ratios) are all complementary or alternative PMF signals. Most mature companies use a portfolio of signals rather than relying on a single survey question.
Sources & references
- First Round Review essay (original) — Rahul Vohra primary essay describing the methodology.
- Sean Ellis original framework — Ellis original writing on the survey methodology.
- Superhuman company history — Historical reference for Superhuman context.
- Superhuman Series C coverage — TechCrunch coverage of last funding round.