Notes from people who sat with the leaving

These notes name the work that happened. They are not star widgets.

The diagnostic spent two days on our export-failure tickets before it touched the cohort table. That order annoyed me at first. Then the memo showed that paid teachers who hit a failed PDF in week two never came back, cancel or not. We changed the retry copy, not the “onboarding.”

Lina V., operations lead, classroom app — diagnostic, 2025

Owen sat with our billing calendar and refused to call a failed Direct Debit “churn” until the retry window closed. Finance had been mixing the two for a year. The sitting was dry, and it was the first time the numbers matched the bank.

Chris P., finance partner, B2B field-service app — cohort sitting, Manchester team visiting Birmingham

I wanted a full diagnostic. Priya told us the product was still in closed beta and booked a readiness review instead. That felt like a downgrade until the punch list named three leave-screen phrases that apologised for the product. We rewrote those before charging anyone.

Dana R., founding designer — retention readiness, 2026

The quarterly watch is useful only because they keep the original ranking on the table. If a pattern has not moved, they say so in a short addendum. Last quarter they wrote that our win-back email was noise. I still send it; I just stopped pretending it was a reading of risk.

Sam K., product, audio subscription — mild reservation included on purpose

A longer visit

A Midlands grocery-list app came in after a winter dip they blamed on “seasonality.” Mara asked for the leave-screen recordings and the tickets tagged “too many reminders.” The recordings showed people hunting for a pause, not a cancel. The tickets repeated the same hour: Sunday evening after a busy shop. The memo ranked “no pause, only cancel” above weather. They shipped a pause for two billing cycles. We were not in the room for the ship; we were in the room for the ranking.

If you want a similar sitting, write to the office.