Why trial-start and first-charge cohorts disagree

Open two retention charts that look identical at a glance. One starts the clock when a person taps “start trial.” The other starts when the first charge clears. By week four the curves tell different stories, and the weekly meeting becomes a negotiation about which story is “true.”

In subscription app analytics, both clocks can be valid. The mistake is mixing them without labelling the grain. Trial-start cohorts answer questions about onboarding friction. First-charge cohorts answer questions about paid durability. When a core KPI dashboard shows a single “retention” tile without naming the start event, product and finance will each import their preferred narrative.

A practical labelling habit

Write the start event into the chart title: “Retention from trial start” versus “Retention from first successful charge.” If your tool truncates titles, put the definition in a pinned note above the dashboard. During Core Pulse Point reviews we treat unlabelled retention tiles as incomplete — not because the engineering is wrong, but because the conversation cannot be refereed.

What to retire

Cumulative “users retained” counters that reset definitions mid-year. If you changed trial length from fourteen days to seven, split the cohorts at the change date instead of blending them into one heroic curve.

Takeaway

Pick one primary retention clock for the executive pack. Keep the other in an analyst workspace. Your core KPI dashboard should force a choice, not hide it.