What is RFM?

RFM groups customers by three dimensions: time since the last purchase, number of purchases and total purchase value within a defined period. It helps prioritize work but cannot by itself explain why someone stopped buying or predict their intent. Results depend on complete order history and consistent customer definitions.

Clean the data before scoring

Align customer identifiers where there is an appropriate basis to use them, and exclude cancelled orders and tests. Choose a period that suits the buying cycle: restaurant visits differ from durable-goods purchases. Record when segments were refreshed so an active segment does not become a stale list reused without review.

Turn segments into hypotheses

New customers may benefit from product and ordering guidance. Frequent customers may need a relevant reason to return. For inactive customers, understanding a previous experience may be more useful than immediately offering a discount. These are hypotheses to test, not certain descriptions of every individual.

Avoid repetitive reactivation messages

Define the purpose and frequency of contact. Respect customer preferences, channel requirements and the appropriate legal basis. Do not use sensitive data to infer a marketing need. Review the offer, timing and opt-out process before sending, and handle requests to stop contact through the business’s procedures.

Test the actual effect

Compare repeat purchasing before and after the experiment and use a suitable comparison group where possible. Include discount and messaging costs rather than attributing every returning customer to the campaign. Business Brain’s loyalty section provides data-dependent segments and estimates and can support follow-up plans. LTV and churn estimates are not guaranteed outcomes.