The Situation
Superhuman set out to build an email client, which sounds like a thankless task. Email is a mature, crowded category dominated by free, capable incumbents. The conventional wisdom is that nobody wants to pay for email and that the problem is solved. Into this unpromising space, Rahul Vohra and his team aimed to build a fast, premium email experience that people would actually pay a meaningful subscription for. The challenge was not only competitive. It was the deeper question that haunts every product: did they have product-market fit, and if not, how would they even know?
Most teams treat product-market fit as a mystical, binary, after-the-fact judgement. You either feel it or you do not. You "know it when you see it." This vagueness is comforting and useless. It gives a team no way to tell whether they are close, no way to measure progress, and no way to decide what to work on next. The contribution of this case study is that Superhuman refused to treat product-market fit as a mystery. They turned it into a metric they could measure, segment, and deliberately improve. That reframing, more than any specific feature, is the lesson.
The Bet: Make Product-Market Fit Measurable
The central idea Superhuman adopted, building on thinking from Sean Ellis, was a single survey question asked of users: "How would you feel if you could no longer use this product?" The answer options were "very disappointed," "somewhat disappointed," and "not disappointed." The insight is that the percentage of users who say they would be "very disappointed" is a usable proxy for product-market fit. A rough benchmark held that if around forty percent of users would be very disappointed to lose the product, you likely have fit worth building on.
This is a deceptively powerful move. By converting an abstract feeling into a number, the team gained everything that measurement provides: a baseline, a way to track progress, and a way to argue about priorities with evidence rather than opinion. Product-market fit stopped being a thing you waited to feel and became a thing you could work on directly. When Superhuman first ran the survey, their score was below the benchmark, which would normally be a discouraging verdict. Instead, they treated it as a starting line. The question was no longer "do we have fit?" but "how do we raise this number?"
What They Actually Did
The genius of the approach was not the survey question alone, which had existed for years. It was the engine the team built around it. Rather than reading the survey as a single aggregate score and shrugging at it, they dissected the responses to find a repeatable path to improvement. Several specific moves made this work.
- They segmented to the high-expectation users. Instead of trying to please everyone, they focused on the people who said they would be "very disappointed." These users already loved the product. Understanding exactly who they were and what they valued became the compass for the roadmap.
- They studied the "somewhat disappointed" group too. This middle group held the key to growth. Some of them could be converted into the "very disappointed" camp if the product gave them what they were missing. The team looked for the specific gaps that separated lukewarm users from devoted ones.
- They identified the loved features and doubled down. By asking what the "very disappointed" users valued most, they learned what to protect and amplify, ensuring improvements never diluted the core that already worked.
- They identified the blockers and fixed them. By asking the convertible middle what held them back, they built a prioritised list of exactly what to improve to move the score, rather than guessing.
The result was a roadmap derived directly from the product-market-fit data. Every item on it was either protecting what the most devoted users loved or removing a specific reason a convertible user fell short of loving it. This is a fundamentally different way to prioritise than the usual mix of intuition, loudest stakeholder, and competitive feature-matching.
Ignore the Wrong Users on Purpose
A subtle and important move was deciding whom to ignore. The "not disappointed" users were, counterintuitively, not the priority. These were people for whom the product was never going to be a strong fit, and chasing them would have pulled the product away from the segment that loved it. Trying to satisfy users who fundamentally do not want what you are building dilutes the product for the users who do. The discipline to deprioritise the wrong users is as important as the effort to delight the right ones. A product that tries to be for everyone usually becomes "very disappointing" to lose for no one.
Why It Worked
The approach worked because it made product-market fit a controllable variable rather than a matter of fate. With a clear metric, a clear segmentation, and a roadmap derived from both, the team could deliberately improve their score over time. They were not hoping for fit. They were engineering it, iteration by iteration, with each release targeted at either reinforcing the loved core or removing a known blocker for convertible users. The score rose as a direct consequence of work aimed at raising it.
It also worked because it imposed focus. By orienting around the "very disappointed" segment, the team had a clear definition of who they were building for and a clear filter for every decision. Feature requests could be evaluated against a real question: does this help our high-expectation users love us more, or convert a lukewarm user, or neither? Requests that served no one in the target segments could be declined with confidence. This is the practical value of a measurable framework: it does not just tell you how you are doing, it tells you what to do and what to ignore.
What Almost Went Wrong
The risks in this approach are real, and a careful PM should understand them rather than adopt the framework blindly. The first risk is misreading a low initial score as a verdict rather than a baseline. A team that took an early below-benchmark result as proof of failure would have quit exactly when the useful work was about to begin. The score is a starting measurement, not a judgement.
The second risk is over-fitting to a narrow segment. By focusing intently on the users who already love the product, a team can build something exquisitely tuned to a small group and miss whether that group is large enough to matter. The framework tells you how to deepen love among the people who already have it; it does not, by itself, tell you whether the market is big enough. A PM has to hold both questions at once. The third risk is treating the survey as the whole truth. The metric is a powerful instrument, but it is still a proxy, and it works best alongside qualitative understanding of why users answer the way they do, not as a replacement for that understanding.
The Lessons for Product Managers
The first lesson is that product-market fit can be made measurable. Instead of treating it as a feeling you wait to experience, you can adopt a concrete metric, the share of users who would be "very disappointed" without your product, and use it as a baseline and a progress indicator. This single reframing turns an intimidating, vague goal into a tractable, trackable one. It lets a team argue with evidence and measure whether their work is actually moving fit.
The second lesson is to segment to your high-expectation users. The path to fit runs through the people who already love the product and the convertible middle who almost do, not through the people who were never going to care. Knowing exactly whom you are building for, and having the discipline to deprioritise everyone else, concentrates your effort where it compounds. Averaging all feedback together is how products become bland.
The third lesson is that this can be a repeatable engine, not a one-off study. The survey, the segmentation, the roadmap derived from loved features and known blockers, and the re-measurement form a loop you can run again and again, deliberately raising the score each cycle. Product-market fit becomes something you build on purpose rather than something you stumble into.
A Final Word
Superhuman's real contribution to product thinking is not an email client. It is a demonstration that product-market fit, long treated as an almost mystical state, can be measured, decomposed, and deliberately improved. The team took a question many people knew and built a disciplined engine around it: measure the score, find the users who love you, understand the ones who almost do, fix what blocks them, and measure again. Any PM can run this loop. The lesson is that the most important and most intimidating goal in product, fit itself, does not have to be left to luck or feel. It can be engineered.
Key Takeaways
- Make fit measurable. Use the "how would you feel if you could no longer use this?" question and track the share of "very disappointed" users as a baseline and progress metric.
- Treat a low score as a starting line. An early below-benchmark result is a beginning to work from, not a verdict to quit on.
- Segment to high-expectation users. Deepen love among those who already love you, convert the lukewarm middle, and deliberately ignore those who were never your users.
- Derive the roadmap from the data. Protect the features the devoted love and remove the specific blockers stopping convertible users from loving you.
- Build a repeatable engine. Measure, segment, fix, and re-measure as a loop, and pair the metric with judgement about whether the target segment is large enough.