kust.io: onboarding built with Clew lifted activation by 30% — and what those numbers don’t tell you
A case study from our own user-research service: what we put in, what the internal numbers showed, and the questions we would ask about those numbers ourselves.
kust.io is our own product, a user-research service, and one of the first places we put Clew. According to kust.io’s internal data, conversion to activation rose by 30% afterwards, and the share of people who finish onboarding rose by 50%. Here is what we actually built, and how much weight those numbers can carry.
Where new users got lost
In kust.io, activation means launching your first study. No launch, no sense of how the service works — and most likely no second visit.
We designed the interface to be clear. But clear to the team that built it and obvious to someone arriving for the first time are different things. Newcomers spent time working out where a study starts and where everything lives.
What we put in
A tour to the first study
It goes from sign-in to launching the first study and shows only the navigation needed for that — not a tour of every menu.
Inviting participants without leaving the app
A study doesn’t happen if nobody shows up for the interviews. Hints help people invite participants to customer interviews right from kust.io, without switching to email or a messenger.
Surveys as hints instead of chat
We used to collect feedback through a built-in chat and rarely got answers. The same question, asked as a hint in the right place in the interface, gets answered more willingly: there is no separate window to open.
Result
Both figures come from kust.io’s internal analytics, exactly as we recorded them in the case.
How to read these numbers
- It is our product and our tool. We have an interest in a good result — factor that in.
- The case doesn’t publish absolute values, the period, or how the comparison was made. So don’t treat +30% as a forecast for your product: it shows the approach worked for us, not how much it will do for you.
- Activation moves with everything at once — the interface, traffic, seasonality. The only reliable way to isolate onboarding’s share is an experiment with a control group.
Today we would test a change like this with an A/B test inside Clew: half the newcomers see one version of the tour, half see the other, and the dashboard works out significance and how long to wait. How we do that on our own site is in how we onboard Clew users with Clew.
What we took away
- Even a clear interface benefits from a guide to the first key action — to that action, not through every section.
- A question asked in context gets more answers than a chat you have to open.
- Decide how you will measure before you launch: what counts as activation, what you compare against, how long you wait.