How Superhuman Went From 22% to 58% Product-Market Fit in 9 Months
How to put the Customer Discovery framework into action
Rahul Vohra had a problem in summer 2017.
His team at Superhuman had grown from 7 to 14 people. They’d been coding for two years straight. Investors were asking when they’d launch. The team was getting restless.
But he had no way to know if they were ready.
You know this feeling. You’re adding features. Testing pricing. Tweaking messaging. Trying five things at once because you’re not sure which one will move the needle.
Rahul built a system to know. A way to measure product-market fit before spending a dollar on growth.
The result? He went from 22% product-market fit to 58% in nine months. Not by guessing faster. By knowing what to build.
This decision—and the framework he built around it—is exactly what Customer Development is supposed to do. And it’s why I spent yesterday writing up the Customer Development framework for you.
The Survey That Changed Everything
Rahul found a metric from growth expert Sean Ellis: ask users “How would you feel if you could no longer use the product?”
Track the percentage who answer “very disappointed.”
The threshold for product-market fit? 40%.
Rahul surveyed his early users. The ones who’d used Superhuman at least twice in the last two weeks.
Result: 22% said “very disappointed.”
That’s barely half the threshold.
Rahul asked: what if we could increase this number systematically?
Discovery: Who Actually Loves This?
Rahul looked at that 22% who loved Superhuman. He assigned personas to each person. Founders. Managers. Executives. Business development.
He ignored everyone else.
Then he asked: what do these people have in common?
They dealt with 100-200 emails daily. They sent 15-40 replies. They prided themselves on being responsive. They aimed for inbox zero but rarely got there.
He named her Nicole.
Rahul made Nicole the only person that mattered.
Not “potential customers.” Not “the market.” One specific person with one specific set of problems.
The product-market fit score jumped to 33% just from this focus.
One person. One problem. That’s the constraint that forces clarity.
When you’re trying to serve everyone, you’re building features for people who don’t care. When you pick one person, every decision gets easier. Build this or that? Ask: does Nicole need it?
Validation: Why Does Nicole Love This?
Next question: what does Nicole actually value?
Rahul surveyed the “very disappointed” group again. This time asking: “What is the main benefit you receive from Superhuman?”
The answers clustered hard around three things:
Speed
Keyboard shortcuts
Focus (one email at a time)
Everything else? Noise.
Then he looked at the “somewhat disappointed” group. The ones who wanted to love Superhuman but couldn’t quite get there.
He found a pattern: the ones who valued speed but were only somewhat disappointed had one thing in common.
They needed a mobile app.
Not “would be nice to have.” Needed.
Calendaring. Integrations. Better search. Read receipts. All showed up in the data.
The Roadmap That Writes Itself
Rahul split it 50/50.
Half the roadmap: make speed even faster. Add more shortcuts. Build more automation. Polish every detail Nicole already loved.
Half the roadmap: mobile app. Calendaring. The features blocking people from becoming Nicole.
Three quarters later?
Product-market fit score: 58%.
From 22% to 58% in nine months. Not by launching. Not by “iterating with the market.” By systematically talking to customers and building what the data told them to build.
What This Actually Means for You
Customer Development isn’t about launching fast. It’s about answering two questions before you scale:
Who will be very disappointed without your product?
What makes them feel that way?
Everything else is guessing.
You don’t need 2 years to figure this out. Just the discipline to ignore everyone except the people who would be very disappointed.
Rahul built an engine instead. Survey users. Segment by disappointment level. Focus on what the very disappointed love. Fix what blocks the somewhat disappointed from getting there. Measure the score. Repeat.
The score became Superhuman’s only OKR.
That’s Customer Development working.
When Do You Use This?
You’re in one of three situations:
Situation 1: You’re pre-launch with some beta users
You think you’re close but you’re not sure. You have a gut feeling something’s missing. You keep adding features hoping one will click.
Do this instead: Run the survey. Find your 40%. Build only for them until you hit 40% “very disappointed.” Then scale.
Situation 2: You’ve launched but growth is expensive
Your CAC is too high. Conversion rates are weak. You’re getting users but they’re not sticking. You can’t figure out why.
Do this instead: Survey your happiest users. The ones who’ve stuck around. Find out what they value. Stop building for everyone else.
Situation 3: You’re stuck between ideas
You have three feature requests. Your team is split. Everyone has an opinion. Nobody has data.
The problem isn’t deciding between A, B, or C. The problem is you’re asking the wrong question. You’re trying to pick features when you should be asking: which one moves us closer to 40%?
Do this instead: Talk to the customers who would be very disappointed. Ask what benefit they get. Build what reinforces that benefit, not what sounds cool.
Customer Development gives you the framework to stop guessing. It won’t tell you what to build. But it will tell you who to build for and what they actually value.
That lets you move faster. Not because you’re cutting corners. Because you’re building the right things.
Rahul moved faster by knowing. He had a number that told him if he was getting closer or further from fit. He had a system that told him what to build next.
That’s the difference between wandering and walking toward something specific.
The hard part isn’t the framework. It’s the discipline to use it every week. To ignore the feature requests from people who aren’t Nicole. To say no to ideas that sound good but don’t move the score.
That discipline is what separates founders who find fit from founders who burn runway chasing everything.
The system works. Superhuman proved it. But you need the framework to apply it.
The 50/50 roadmap split is what makes this approch powerful. Doubling down on what existing power users love while also removing friction for the somewhat-disappointed segment creates compounding momentum instead of just chasing random feature requests. I tried a similar framwork with a B2B tool last year where we kept building for "the market" and conversion stayed flat. Once we narrowed to one specific persona and asked what benefit they actually got, we cut like 40% of planned features and saw usage patterns shift within weeks. The discipline to ignore everyone except the very disappointed group is harder than it sounds tho.