Buyers split into six behavioral segments by recency and frequency. The retention grid then shows how quickly first-time visitors fall away, which is the real constraint on growth here.
Population and per-segment economics from user_rfm_segments.
| Segment | Buyers | Avg recency | Avg spend | Total spend |
|---|---|---|---|---|
| Champions | 2,897 | 8d | $922 | $2,669,933 |
| At Risk | 2,845 | 43d | $608 | $1,730,075 |
| Lost | 2,565 | 45d | $294 | $753,941 |
| Loyal Customers | 2,371 | 18d | $585 | $1,386,860 |
| Promising | 1,812 | 10d | $337 | $610,161 |
| Regular | 1,033 | 21d | $306 | $316,420 |
Every buyer placed by recency and lifetime spend. Clusters correspond to the named segments.
Each row is the week a user first appeared. Cells are the share still active in that later week.
| Cohort start | Users | w0 | w1 | w2 | w3 | w4 | w5 | w6 | w7 | w8 |
|---|---|---|---|---|---|---|---|---|---|---|
| Sep 30 | 20,810 | 100% | 34% | 31% | 26% | 25% | 28% | 32% | 23% | 20% |
| Oct 7 | 19,068 | 100% | 26% | 21% | 19% | 22% | 25% | 17% | 15% | |
| Oct 14 | 15,876 | 100% | 21% | 17% | 20% | 23% | 16% | 14% | ||
| Oct 21 | 13,870 | 100% | 18% | 18% | 20% | 14% | 12% | |||
| Oct 28 | 11,780 | 100% | 21% | 21% | 14% | 12% | ||||
| Nov 4 | 14,609 | 100% | 26% | 15% | 12% | |||||
| Nov 11 | 16,449 | 100% | 20% | 14% | ||||||
| Nov 18 | 13,563 | 100% | 14% | |||||||
| Nov 25 | 9,742 | 100% |
LowerHigher retentionGrid runs to week 8.
Join one user's RFM segment to their propensity score.
The lookup needs two files that are not loaded on any other page: the full segment table and every propensity score. About 4 MB, fetched once and cached.