Claim and evidence
For me, the goal of segmenting was to find pockets in which Superhuman might have better product-market fitSuperhuman · review.firstround.com · 2018-11-12
Primary
The company or founder describes the tactic and its result themselves: own blog, talk, filing.
Score 22% overall, 33% after segmenting.
Working · 2026-11 Founder's own account, 2018.
What the numbers show
point at which Superhuman's founder says results start to be directionally correct
review.firstround.comWhy it works
The overall score hides pockets of strong fit. Superhuman narrowed to enthusiastic personas and speed-focused users to lift the score.
How to run it
- Cut the "very disappointed" share by persona
- Find the benefit the lovers name
- Prioritise that segment in onboarding
- Re-run next quarter
Ways to run it
Design options to test, not promised results.
- Option APersona cut
Compare the very disappointed share by persona and keep the ones with the strongest fit, as Superhuman did.
- Option BBenefit cut
Group answers by the main benefit people name and focus on users who value the same thing.
- Option CTargeted onboarding
Shape onboarding and acquisition around the best-fit segment, then re-run the survey next quarter.
Timeline
- 12018 Superhuman — Segmented survey data to the most enthusiastic personas and users who named speed.
- 22026-11 status checked: Working
Dated examples
- 2018
Superhuman
Segmented survey data to the most enthusiastic personas and users who named speed.
Score 22% overall, 33% after segmenting.
“For me, the goal of segmenting was to find pockets in which Superhuman might have better product-market fit”
✓ quote checked word-for-word on the pagereview.firstround.com · 2018-11-12
Does it still work?
Founder's own account, 2018.
Risks
Sources
- review.firstround.com
“you start to get directionally correct results around 40 respondents”
✓ quote checked word-for-word on the page
Change log
- 2026-11-20 — harvest v0.1 (activation)