How to win back customers who stopped coming

The distinction that decides everything

“Lapsed” is not one group. It is two, and they need opposite treatment.

Drifted. They still like you. Nothing went wrong. Life reorganized itself — a job moved, a routine changed, they walked a different way for three weeks and the habit dissolved. They would come back on a reminder alone.

Gone. They made a decision, or moved, or found somewhere better. A reminder is noise. Only a genuinely good offer has a chance, and most of the time not even that.

The mistake that wastes most win-back effort is treating both as one list and sending both the same thing. The drifted group did not need a discount and you just taught them to wait for one. The gone group needed more than a nudge and got less.

Defining “went quiet” honestly

There is no universal number of days. The only definition that means anything is relative to how often that customer used to come.

Start from your own pattern:

1. Find your normal gap. For a café it might be three days; for a barber, five weeks; for a salon, eight. Look at your regulars, not at an average that includes one-off visitors.
2. Set “quiet” at roughly twice that gap. Long enough that it is genuinely unusual, short enough that the habit has not fully dissolved.
3. Set “gone” at roughly four times it. Beyond that you are writing to someone who has stopped thinking about you.

A café customer missing for a week is not lapsed, they are on holiday. A salon customer missing for a week is not lapsed either — they are not due. Any tool that applies one fixed number to both is telling you about its own defaults rather than about your customers.

What to send each group

To the drifted: a reminder with no discount.

They left mid-card. The most effective thing you can tell them is where they stand — “you are three stamps from your reward” is a reason to come back that costs you nothing and does not devalue anything. If they were near the reward when they drifted, say so plainly.

To the gone: one good offer, once.

Something clearly worth a trip, with an end date, sent a single time. Not a series. A sequence of escalating offers to someone who has already decided reads as desperation, and it trains the people who are still coming to wait.

The distinction is not about generosity. It is that a reminder answers “did I forget?” and an offer answers “is it worth going back?”, and those are different questions.

What a win-back is actually worth

Do this arithmetic before deciding what to spend, because it is usually smaller than it feels.

Take your average transaction. Multiply by how many visits a recovered customer typically makes before drifting again — for most local businesses this is a handful, not a lifetime. That is the ceiling on what the recovery is worth.

Then subtract the cost of the offer, and subtract the cost of everyone who would have come back anyway and took the discount on the way in. That second number is real and easy to forget: send a 20% offer to a hundred lapsed customers, and a meaningful share of the redemptions come from people already planning to visit.

The result is often that a plain reminder is more profitable than a good offer, even though it recovers fewer people.

Timing

Send when they went quiet, not on the first of the month. A batch sent to everyone on a Monday reaches each person at a random point in their own cycle. A message triggered when that specific customer crosses their own gap arrives when it means something.

This is the practical case for a standing rule over a scheduled campaign: not that it saves you time, but that it is the only way to be timely for each person individually.

Do not send twice in quick succession. The single fastest way to make a lapsed customer permanently gone is to become the business that emails them. If they did not respond to a well-timed reminder and one honest offer, they have answered.

Doing this in NeoLoyal

The Went quiet and Lost audiences in Campaigns are these two groups. Selecting one shows how many cardholders are currently in it, so you can see whether the segment is large enough to be worth a send before writing anything.

The two win-back rules in Autopilot run the same split as standing rules, timed per customer rather than per batch. In suggestion mode they draft and wait for approval, which is worth using for at least a few rounds — reading what the rules produced for people whose names you recognize is the fastest way to find out whether your definition of “quiet” matches reality.

Both draw on the same per-customer contact budget, so a manual campaign sent yesterday quiets the automatic rules today. That is deliberate: the two paths cannot double up on the same person.

What not to do

• Do not send to “everyone” and call it a win-back. Your active regulars receive an offer they did not need and learn that waiting pays.
• Do not escalate. A second, better offer to a non-responder teaches everyone that the first offer was not the real one.
• Do not win back at any price. A customer recovered with a discount steep enough to lose money on is not a recovered customer, they are a subsidized one.
• Do not measure it by opens. The only number that matters is visits inside the following few weeks, which is on the dashboard already.

The honest summary

Most lapsed customers do not come back, and no message changes that. The ones you can recover are mostly the ones who drifted, and they are recoverable with a reminder rather than a discount.

That makes win-back a cheap thing to do well and an expensive thing to do enthusiastically.