Names and aliases
buyer cohort survival, buyer retention curve
Formula
buyers_active_in_period_n / buyers_in_cohort × 100%, for n = 1, 2, 3 …
Metric card
| Unit | % of cohort per period |
|---|---|
| Data source | Cohort-entry events, completed transactions carrying a buyer identifier, the period definition, and account deduplication rules. |
| Frequency | By first-transaction cohort, with the period sized to the natural purchase frequency of the micro-market. |
| Segments | micro-market, first-purchase category, acquisition channel, geography, cohort |
Interpretation errors
- Collapsing the curve into one number, which erases the difference between rapid churn and slow decay.
- Choosing a period shorter than the natural purchase frequency and labelling an ordinary gap between purchases as churn.
- Plotting cohorts of different maturity together without aligning them on period index.
Example
Illustrative example: a cohort retains 38% in the first period, 24% in the second, and 19% in the third. The numbers describe no real platform and set no benchmark.
Use
Publish the curve rather than a point, and state the period definition beside the value. Size the period from the observed distribution of intervals between purchases in the specific micro-market, not from calendar convenience. Read retention together with demand frequency: an identical curve means different things under different natural demand rhythms.
Anti-example
Higher cohort retention after a repeat-purchase incentive may reflect a timing shift rather than a formed habit: buyers brought a planned purchase forward, and the following period will show the shortfall.
Limitations
- Retention is category-dependent: for infrequent purchases a low curve is normal and is not evidence of a problem.
- The metric observes transactions only and cannot distinguish a departed buyer from one who currently has no need.
- Values are sensitive to identity resolution; without deduplication retention is systematically understated.
Related metrics
buyer-repeat-rate, demand-frequency, seller-retention, transaction-completion-rate
Context sources
These works provide two-sided-market context; the formula above remains an author-defined working definition.
- Geoffrey G. Parker, Marshall W. Van Alstyne. Two-Sided Network Effects: A Theory of Information Product Design (2005).
- Mark Armstrong. Competition in Two-Sided Markets (2006).