27 founders visited my product. 90% left. I'm realizing I don't know what problem I'm actually solving.
A few days ago I changed Ashive's onboarding after a stranger gave me incredibly useful feedback.
Before that, users had to sign up pretty early. Now they can actually experience part of the product before creating an account.
After the change I got around 27 visitors.
Almost 90% bounced.
My first instinct was:
The onboarding still isn't good enough.
But after reading a lot of discussions here recently, I'm no longer convinced that's the right conclusion.
Maybe it's onboarding.
Maybe my positioning attracts curiosity but not the right founders.
Maybe the product promise doesn't match the experience.
Or maybe I simply don't have enough users yet to distinguish signal from noise.
That's the part I'm finding hardest.
With 27 visitors and one detailed piece of feedback, almost every explanation still feels like a hypothesis.
Right now I'm trying to spend less time guessing and more time talking to founders, because every meaningful conversation has taught me more than another week of building.
For those who were at this stage:
When did you know you had enough data to confidently say this is an onboarding problem instead of this is a positioning or ICP problem?
I think one of the hardest parts is that a bounce is an observation, not an explanation.
With 27 visitors, several hypotheses can fit the same outcome: onboarding, positioning, ICP, expectation mismatch, or simply too little data.
What has helped me think about these situations is asking, "What evidence would rule out each hypothesis?" So you don't have to believe any pre existing hypothesis.
Sometimes a single piece of contextual feedback can eliminate an entire line of thinking, while another week of analytics just produces more uncertainty.
To me, the goal is to collect the kind of signal that reduces ambiguity, not more data.
I really like that framing.
A bounce is an observation, not an explanation is probably something I'll remember for a long time.
I also like the question about ruling hypotheses out instead of trying to confirm one. I think my instinct has always been to explain the data as quickly as possible instead of asking what evidence would actually distinguish one explanation from another.
Appreciate this.