1. New: start with the first recorded visit

Define what counts as a visit in your business. An order, an attended session and a completed appointment are different events. Ask for contact permission clearly; a purchase alone is not permission to send marketing.

2. Second visit: make the next step useful

Look for the customer’s next qualifying visit. Choose a reminder window that fits the normal buying cycle, then test a simple reason to return. A gym session and a haircut should not use the same timing.

3. Regular: learn the routine

Use repeated behaviour, not a single large bill, to recognise regulars. Review how often someone usually visits, what they choose and whether that routine is changing.

4. At-risk: look for a change in pattern

Compare the current visit gap with that customer’s usual gap. Treat the signal as a reason to investigate, not a certainty that the customer has left. Seasonal patterns and incomplete records can explain a change.

5. Lapsed: be useful and restrained

Decide how many missed cycles make someone lapsed. Send only where permission and messaging rules allow it. Cap frequency, honour opt-outs and avoid endlessly repeating the same offer.

6. Won back: measure the next visit too

Record a qualifying return inside a defined attribution window. Then follow repeat behaviour after that return. Attributed sales are not proof that a message caused every purchase. These are proposed stages; tune their definitions with your team.

Put the idea into practice.

Try the related workflow