How Superhuman Built an Engine to Find Product Market Fit
by Rahul Vohra (founder and CEO of Superhuman)
TL;DR: This article presents the most actionable framework for measuring and systematically improving product-market fit. Vohra was struggling pre-launch at Superhuman - two years in with no launch and no way to articulate the situation to his team. He discovered Sean Ellis's benchmark: if 40% of users would be 'very disappointed' without your product, you have PMF. Superhuman started at just 22%. The four-step engine: (1) Segment to find supporters - use the 'very disappointed' group to narrow the market and create a High-Expectation Customer profile; (2) Analyze feedback - ignore 'not disappointed' users (lost causes), focus on 'somewhat disappointed' who share your main benefit; (3) Build roadmap 50/50 - half doubling down on love, half addressing blockers; (4) Repeat quarterly and make PMF score the key OKR. The survey uses four questions about disappointment, ideal customers, main benefit, and improvements. Superhuman jumped from 22% to 33% just by segmenting, then to 58% after three quarters of execution.
Key Insights
- 40% 'very disappointed' is the PMF benchmark - this is a LEADING indicator
- You can get directionally correct results with just 40 survey respondents
- Segmenting by 'very disappointed' users immediately boosts PMF score
- High-Expectation Customer (HXC) profile focuses the entire company on serving a narrow segment
- IGNORE feedback from 'not disappointed' users - they're lost causes that will distract you
- Focus on 'somewhat disappointed' users who share your main benefit - something small holds them back
- Split roadmap 50/50: half on love, half on blockers - either alone fails
- Make PMF score the most important metric - track weekly, monthly, quarterly
Actionable Takeaways
- Survey users with 4 questions: disappointment level, ideal customer, main benefit, improvements
- Target 40% 'very disappointed' as your PMF benchmark
- Segment respondents by persona and focus on the 'very disappointed' group
- Create detailed HXC profile using Julie Supan's framework
- Politely disregard feedback from 'not disappointed' users
- Focus improvement efforts on 'somewhat disappointed' who share your main benefit
- Split roadmap 50/50 between doubling down on love and addressing blockers
- Rebuild roadmap quarterly based on new survey data