Named exits, not blended rates
You learn to split “did not activate” into email bounce, password stall, empty workspace, and delayed human review — then pick one to repair first.
Birdwell studio · GB product teams
Desktop Anchorcore is a teaching practice for Onboarding Drop-Off Analysis. We work with people who already have dashboards and still cannot explain why a verified user never reaches the first useful job.
Flagship
Twelve lessons that follow a single product from invite email to the first saved object. You will leave with a drop-off map, a shortlist of friction you can actually change, and language your support team can reuse.
Fees are listed separately. This site does not take payment.
We ask students to keep one abandoned signup on the table for the whole course. Aggregate charts come later, once that person has a name, a timestamp, and a plausible reason they left.
These notes refer to modules you will find in the studio, not generic praise.
The “quiet hour” exercise in week two made our activation chart look less clever and more honest. We still disagree about the invite copy, which is useful.
I came for the funnel map and stayed for the support-handoff worksheet. It is slower than a template pack. I needed that pace.
Priya’s module on empty states helped, though the homework load in week three collided with our release freeze. I skipped one critique and felt it.
You learn to split “did not activate” into email bounce, password stall, empty workspace, and delayed human review — then pick one to repair first.
Drop-off work dies when only growth can explain it. We write short scripts for the people who already talk to confused new users.
Every map includes a line on what the data cannot show. GB teams tell us that sentence saves more time than another chart.
Observation
Why a green verification tick can still be the last event you ever see.
Method
A caution about weekly aggregates that swallow Thursday-night exits.