DTC retention KPIs that matter
A DTC leadership team needs a short retention KPI set read at cohort level: repeat purchase rate, median time to second order, cohort revenue per acquired customer, third-order rate, reactivation rate and program participation. Open rate, list size and platform-attributed revenue are diagnostics at best, not retention KPIs.
Most retention dashboards track what the tooling makes easy to track. This is the shorter list we actually run against, with the definition we use for each one.
The core set
| KPI | Definition | What it tells you |
|---|---|---|
| Repeat purchase rate (90d) | Share of a first-order cohort with a second order within 90 days | Whether the early lifecycle is working |
| Median time to second order | Days between first and second purchase | Moves before LTV does. Best early signal |
| Cohort revenue per customer | Cumulative revenue per acquired customer at day 90, 180, 365 | Whether value is compounding |
| Third-order rate | Share of cohort reaching three orders | The threshold where habit forms in most categories |
| Reactivation rate | Share of lapsed customers returning within a defined window | Whether winback is real or coincidental |
| Program participation | Share of revenue from enrolled members | Whether loyalty is a mechanism or a badge |
Metrics we deliberately deprioritise
- Open rate. Unreliable since privacy proxies, and never the objective.
- List size. Grows without retention improving.
- Platform-attributed revenue reported as incremental revenue.
- Flow-level revenue with no holdout and no cohort context.
How to instrument these
Every KPI above needs order history joined to customer identity, which is exactly where most stacks break. If the identity spine is not solid, the numbers will be confidently wrong.
Target ranges
We do not publish target ranges without a disclosed dataset behind them.
SOURCE REQUIRED · Category-level target ranges require the benchmarks dataset described in the research methodology.
Have us run these against your data
We build and run lifecycle, loyalty and customer-data systems inside your existing stack. Tool-agnostic. Outcome-owned.
Request a call