Cohort retention

Go to Measure → Analytics → Cohorts to build retention analyses.

How cohorts work

A cohort groups contacts by when they first did something (the entry event). You then track how many of those contacts come back over time (the return event).

Building a cohort

1. Entry event — what defines cohort membership (e.g., signup_completed, first_purchase)
2. Return event — what counts as “retained” (e.g., order_completed, app_opened)
3. Granularity — day, week, or month
4. Date range — how far back to look

Reading the retention table

The output is a triangular table:

Each cell shows the percentage of the cohort that performed the return event in that period.

What to look for

• Flattening curve — retention stabilizes after week N → that’s your “core” retention rate
• Improving cohorts — newer cohorts retain better → your product/marketing is improving
• Drop-off cliff — steep drop between specific weeks → target that window with a re-engagement journey

Segment overlay

Compare cohort retention across segments — “Pro plan users vs Starter” or “WhatsApp-onboarded vs email-onboarded” — to understand which paths lead to better retention.

Outlet narrowing

For multi-outlet brands, the workspace location chip (the ?location= URL state) narrows the cohort to contacts whose entry event fired at a specific outlet. With the chip set to All outlets, the cohort aggregates brand-wide; selecting an outlet rebuilds the triangle using only events stamped with that location_id.

Natural-language outlet phrasing — “Bandra outlet customers who reordered in 30 days” — also resolves through the segment AI; the underlying outlet-name resolver maps the phrase to a location_id before the cohort runs.

What’s next

• Funnels — step-by-step conversion analysis
• Attribution — link campaigns to retention