About this block

Cohort Retention Grid lays out eight weekly signup cohorts as rows, from Apr 20 at 412 users down to the freshest Jun 8 cohort at 418, with weeks since signup as columns from Week 0 through Week 7. Each cell is shaded by the share of the cohort still active that week, so a strong week reads dark and a thin one reads faint. Newer cohorts trail off into a staircase of muted pending cells, and two dropdown chips (Weekly cohorts, Active users) sit above the grid.

Cohorts are one typed array and cellClass maps each value to a foreground opacity band, so the heatmap needs no chart library. A shade legend runs across the footer next to the methodology note that a user counts as active with one or more sessions in the week, which is the definition every retention debate eventually turns on.

When to use

Reach for this block on a retention or growth dashboard, wired to your cohort query and active-user definition. The two dropdown chips should switch the cohort grain and the retention event without leaving the page.

A natural flow around it on an Application Pro page:

Before

• App shell
• Report filters bar

After

• Charts
• Data table

What you can display

One strong use is signup cohort retention. Other grids of the same shape:

• Feature retention. Cohorts by first use of a feature, cells shaded by weekly return to that feature.
• Revenue retention. Dollar retention by billing cohort, cells shaded by share of starting MRR kept.
• Channel cohorts. Rows split by acquisition source so paid and organic curves sit side by side.

Tip: the active definition in the footer note settles more arguments than any single cell, so keep it visible.