Where the funnel breaks, and who never entered it
Your product usage data is already there, and the top-line chart is all anyone has time to pull. DecisionBox walks every step, every customer group and every sequence, and tells you which step is costing you which users.
80% of registered hosts never create a single listing. The onboarding funnel never recovers past step one
A finding from a real discovery run. Nobody asked for it.
Five questions your product dashboard doesn't answer
DecisionBox goes after questions like these on its own, run after run, without anyone writing them down.
- 01
How many of our sign-ups never made it past step one?
Every step of onboarding, counted for everyone who started it.
- 02
Where do users from each channel actually drop off, and where is that different from the average?
The funnel broken down by where people came from, so you fix the right step for the right group.
- 03
Is our return rate growing because the product improved, or because we keep replacing the users we lost?
It separates new arrivals from people who came back.
- 04
How many sessions hit a limit or interruption we never meant to set?
It finds the friction that nobody designed but everybody feels.
- 05
What do users do in their first two days that predicts they are still here a month later?
Early behaviour matched to who stayed, so onboarding can push the right habits.
From a finding to a plan
Every finding that holds up comes with a recommendation: who it targets, what it is worth and the steps to get there.
Review scores hold at 4.27 until day 6 of shipping, then fall to 2.95 and 1.81 by day 10. A boundary, not a gradient
Set a five-day shipping limit and put the 51 slowest sellers on improvement plans
Active players see 17.5 pop-up offers per session. The more offers they see, the less they play
Cap rewarded-ad and booster offers at three per session
Players who finish a first session come back only 3.8 times a month. Most never return in any depth
Launch a seven-day welcome calendar shown on the first session of each day
Every pair is from a real DecisionBox run. Source datasets are anonymised.
What it does, and what it doesn't
- Follows what users do step by step, by group, not just top-line retention.
- Works with the product data you already collect.
- Surfaces quiet problems: groups slowly under-performing, onboarding steps that got worse.
- Replay individual sessions or debug single users. It looks at groups.
- Fix tracking that is broken. Poor data gives poor findings.
- Change anything in your product analytics tools. It only reads.
See DecisionBox at work
Your data has all the answers.
DecisionBox asks the questions.
Let's run it on your data.