What retention rate is and how to compute it
Retention rate is the share of customers who returned for a repeat action within a defined period.
In e-commerce the action usually means a purchase, but it can be a visit, an add to cart or any
other event — what matters is fixing the definition and keeping it stable.
The formula:
Retention rate = (customers at the end − new customers) / customers at the start × 100%
An example: 5,000 active buyers at the start of a quarter. 800 new buyers acquired during it.
4,600 active buyers at the end.
Retention = (4,600 − 800) / 5,000 × 100% = 76%
Time windows: D1, D7, D30, D90
Retention rate is always tied to a window. The e-commerce standard:
| Window | What it measures | Where it matters |
|---|---|---|
| D7 | Short-term engagement after the first purchase | FMCG, subscriptions |
| D30 | The monthly return cycle | Fashion, beauty |
| D90 | The seasonal cycle | Electronics, DIY |
| D365 | Loyalty of the base | Any vertical |
Grocery is best read at D7 or D14. For electronics, D30 can be too short — the purchase cycle is
longer.
Benchmarks by vertical
| Vertical | 30-day retention | 90-day retention |
|---|---|---|
| Grocery / FMCG | 40–70% | 60–80% |
| Pharmacy | 30–50% | 45–65% |
| Beauty | 20–35% | 30–50% |
| Fashion | 10–20% | 20–35% |
| Electronics | 5–15% | 15–25% |
Important: a benchmark is a reference point, not a verdict. Comparison with your own earlier
cohorts matters more than comparison with the market. Watch the trend: is retention improving for
cohorts acquired after personalization went live?
What moves retention in e-commerce
The relevance of the first experience. A shopper who got what they expected — on assortment,
delivery and service — is far more likely to return. The first visit forms the mental model of the
store.
Personalization on the repeat visit. A returning shopper should see content matching their
profile and history; otherwise every visit feels like the first. Recommendations on the homepage
and a personalized catalogue shorten the time to finding what they came for.
Triggered communications. The right moment for a reminder — at the end of a consumption cycle,
or at the first signs of churn — moves retention directly.
Cohort analysis versus an aggregate number
An aggregate retention rate hides differences between cohorts. If the January cohort retains at 25%
and the December one at 15%, something changed. Cohort analysis makes it possible to:
- Assess the impact of a specific change — new personalization, a redesign, an acquisition season
- Compare acquisition channels by long-term retention rather than by first order
- Forecast lifetime value from the early points of the retention curve