THE SHORT ANSWER
Retention describes continued customer activity under a stated rule and time window. Repeat purchase and purchase frequency are observable behaviours. LTV is a model of value across the relationship; historical revenue alone is not a complete economic forecast.
Separate behaviours from models
| Concept | Working meaning | Decision use |
|---|---|---|
| Repeat purchase rate | Share of an eligible customer cohort that buys again in the window | Whether another purchase occurred |
| Purchase frequency | Orders per customer over a stated period | How often active customers order |
| Retention | Share still active under a business-relevant rule | Continuity of the relationship |
| Churn | Share becoming inactive under the stated rule | Loss or lapse requiring diagnosis |
| Historical customer revenue | Observed revenue associated with a customer | Description of recorded behaviour |
| LTV | Modelled value or contribution over a defined horizon | Acquisition, service and investment decisions |
Evidence & context: Google Analytics Help · Google Analytics Help
The purchase cycle determines the window
Replenishable products can support short repeat windows; durable goods may not. A furniture customer can be valuable without buying another sofa in 90 days. Define the expected need cycle and consider cross-category behaviour before labelling inactivity as churn.
Use acquisition cohorts so older customers have comparable opportunity to repeat. Comparing a twelve-month-old cohort with customers acquired last week creates a maturity bias.
Retention begins before the CRM message
- Product quality and fit
- Accurate promise and delivery
- Useful onboarding or care guidance
- Responsive support and fair recovery
- Replenishment reminders timed to plausible need
- Loyalty benefits that customers can understand and use
- Relevant email or messaging with clear consent and frequency controls
Build LTV from contribution, time and uncertainty
A simple historical view can total revenue per customer. An economic model should specify margin or contribution, repeat probability, horizon, servicing cost, returns and possibly the time value of money. The appropriate complexity depends on the decision.
Avoid using a speculative long-horizon LTV to justify today's high CAC. Compare forecast and realized cohort value, and shorten the horizon when uncertainty is high.
Evidence & context: Google Analytics Help
Turn cohorts into action
Segment by acquisition source, first product, offer, geography and customer type. Look for differences that suggest a decision, then test the experience rather than merely increasing message frequency. Connect the result to CAC and payback and the commerce dashboard.
Sources & further reading
- BigQuery Export user-data schema
Google Analytics Help. Documents observed lifetime revenue, purchases and sessions in one analytics system. Historical revenue is not the same as a forward-looking customer-profit model.
- Understand user metrics
Google Analytics Help. Official definitions for total, active, new and returning users. Identity limits and configuration can affect interpretation.
- Ecommerce in Google Analytics
Google Analytics Help. Official documentation for ecommerce events and reports. A measurement implementation does not by itself establish causality or profitability.
Examples and exercises are illustrative unless attributed to a source. No independent expert review is claimed.
A correction, a counterexample or an experience worth sharing?
Join the conversation ↗