Retention
Reading cohort retention without the theatre
Retention tiles love a sparkline. Sparklines love a soft landing. In B2B Dashboard App Analytics the first question is not “is the line up?” It is “who is still in the denominator, and when did they enter?”
Week three of Reading the Revenue Grid is built around that question. A snapshot of “logos still paying this month” can look healthy while a cohort that landed in February is quietly failing to expand. The snapshot is not a lie. It is a different claim. Mixing the two on one home-screen tile is theatre: it lets a present-tense number wear a past-tense costume.
We ask operators to split the claims. One tile: current paying population, with a falsifier (“drops if invoices past 21 days exceed X”). Another tile: a named cohort’s expansion or logo survival, with the entry week printed in the title, not in a tooltip. If the dashboard app cannot title a tile with the entry week, the tile is not ready for Monday.
Seasonality in Thailand makes this worse, not better. A February cohort that includes Chinese New Year closures and a later Songkran lull will look “worse” than a June cohort even when the product did nothing wrong. The charter should say so. Pretending every cohort lives in a US fiscal year is how regional teams get scolded for a calendar.
Pimchanok’s reservation from cohort work still stands: the seasonality homework assumed more public-holiday fluency than a mixed ASEAN team had. We now send a one-page calendar with the welcome note. The maths did not change. The courtesy did.
If your retention tile cannot name its people, it is a mood. Moods belong in the appendix. Claims belong on Monday.