Splitr Onboarding Case Study: From 9% to 36% Completion
Splitr was attracting signups, but most new users never created their first club. Step-level and version-level analysis revealed one weak handoff, giving the team a focused problem to fix.
Splitr helps sports clubs track shared expenses and settle balances. This case study uses the onboarding cohorts reported for app versions 1.4.14 and 1.4.16.
The apparent acquisition problem
Splitr had 110 users complete signup in version 1.4.14, but only 10 completed club creation. Overall onboarding completion was 9%.
At first glance, weak growth can look like an acquisition problem: the obvious response is to find more traffic or redesign the entire onboarding experience. The funnel showed that neither was the right starting point. Users were already signing up; they were disappearing immediately afterward.
110
entered onboarding
10
completed club creation
9%
overall completion
Finding the exact leak
OnRamp separated the journey into meaningful milestones. The largest loss was concentrated in one transition:
signup_completed → club_name_step
Users were completing signup but failing to reach the first club-creation screen. Most never reached the later setup steps, so changing those later screens would not address the main loss.
The team focused on four things:
- Making sure the next action opened reliably after signup
- Removing friction between signup completion and club setup
- Tracking the handoff as a distinct conversion event
- Comparing the next release separately instead of blending app versions
→ Insight
The useful finding was not simply that completion was low. It was that one transition accounted for most of the loss.
What changed in version 1.4.16
The focused onboarding changes shipped in Splitr 1.4.16. The next measured cohort included 140 users, of whom 50 completed onboarding.
| App version | Entered | Completed | Completion |
|---|---|---|---|
| 1.4.14 | 110 | 10 | 9% |
| 1.4.16 | 140 | 50 | 36% |
That was a 27 percentage-point increase in overall completion. At the targeted post-signup handoff, the percentage reaching club_name_step rose from 9% to 50%—more than five times the earlier rate.
Because the results compare release cohorts rather than a randomized experiment, they should not be read as proof that every point of improvement came from one change. They do show that the biggest measured improvement occurred at the exact transition the team targeted.
Looking beyond completion
Completion is only useful when it helps more users reach lasting value. Splitr's early weekly cohorts showed returning users after onboarding, but the newer cohorts were still too young for a firm retention conclusion. Onboarding and retention should therefore be monitored together rather than treating a higher completion rate as the finish line.
The practical sequence is:
- Find the first meaningful step users fail to reach.
- Check whether the loss is isolated to a device, platform, or app version.
- Fix that transition and compare the next release separately.
- Measure whether the users who get through onboarding return afterward.
For Splitr, this turned a vague growth problem into a specific product problem. Instead of asking how to attract more users, the team could see where existing users stopped moving forward—and whether the next release fixed it.
Read the original Splitr analysis on X, or explore OnRamp's public demo to see the funnel and version breakdown workflow.