Your Customer List: Find Who Stopped Coming, and Ask Them Back

Almost nobody tells a restaurant they have stopped coming. They just stop.

The regular who was in every Thursday for two years moved jobs, or had a baby, or went somewhere else once and it was fine. From behind the counter it is invisible — Thursday is still busy, because somebody else filled the seat — and by the time it is noticeable, six months have gone.

This is the least glamorous and most reliable growth available to a food business: you already earned these customers once.

Who it is for

Any venue with repeat custom. Cafés and neighbourhood restaurants most of all, because their whole economics is frequency rather than footfall.

Where the list comes from

Your own activity, and nothing else.

Counter visits, loyalty, orders, bookings and event tickets at your venue build the picture of who your customers are and how often they come. It is not a bought list, it is not scraped from anywhere, and it is not other venues' customers. Nobody appears in your list because they ate somewhere else.

A counter visit counts. That matters for a real reason: a regular who pays cash and never orders online is exactly the customer most likely to look lapsed in a system that only sees online orders, and they are also the customer you can least afford to lose track of.

How it decides somebody has gone quiet

Not by a fixed number of days, because that would be wrong for almost everybody.

A customer who came weekly for a year and has not been in for five weeks has gone quiet. A customer who comes every couple of months and has not been in for five weeks is behaving normally. Treating those as the same thing produces a list full of people who never left, which is how venues stop trusting the list.

What you get is people whose own pattern has changed, which is a far shorter and far more useful list.

Consent, and why you will see people you cannot message

An offer is marketing, and marketing needs permission. You will see customers in the list you cannot send an offer to.

That is the system working. It is worth understanding the distinction because it is easy to read it as a bug:

• A transactional message is about something the customer did — your order is ready, your booking is confirmed, your table is waiting. You are not asking for anything and consent governs what is shown rather than whether you may send it.
• A marketing message is an offer. It needs actual agreement from the customer, and without it the message does not go.

The practical upshot: the most valuable thing you can do for this feature is give customers a reason to opt in, which mostly means running something worth opting in for. A loyalty programme is the usual answer.

What to send

The single biggest determinant of whether a win-back works is whether it contains a reason.

"We miss you" is not a reason. It is a feeling, addressed to somebody who does not feel it. What works is something specific enough to put in a diary:

• A named dish that is back on, or new
• A set amount off, on stated days
• An invitation to a particular night

Be specific about the days, too. Most venues sending a win-back are trying to fill Tuesday, and an open-ended offer fills Saturday, which was full anyway.

How often

Rarely. This is the part venues get wrong.

A win-back message works because it is unusual. Sending one every fortnight converts it into background noise, and worse, it teaches a proportion of your customers to turn notifications off — which removes them from every future message including the ones they would have wanted.

Two or three times a year, aimed at people who have genuinely gone quiet, will outperform monthly sends by a wide margin.

Reading the result honestly

The number that matters is not how many people opened the message. It is how many came in.

Those two come apart more often than you would think. A message with a striking subject line and a weak offer gets read and ignored; a plain message with a genuinely good reason gets acted on. If you only watch opens, you will optimise towards the first one.

The other figure worth tracking is what happens next. Somebody who comes back once because you asked them to is a recovered visit. Somebody who comes back and then returns the following month without being asked is a recovered customer, and that is the whole point of the exercise.

What it costs

Building and reading the list costs nothing. A paid campaign carries its own cost, which you set, and you can see the audience size before you commit to it.

A worked example

A café with 400 identifiable customers finds 60 who have changed their own pattern — people who were weekly and have not been in for over a month.

Of those, 35 have agreed to marketing. They get one message: a named pastry, back on the counter, and €2 off any coffee and pastry on a Tuesday or Wednesday for two weeks.

If nine of them come back and six become regular again, the café has recovered six weekly customers it had already lost. At €5 a visit, that is roughly €1,500 a year, from one message.

What it connects to

• Run a loyalty programme — free, and the main reason a customer agrees to hear from you
• Mosey Tap — how a cash regular gets onto the list at all
• Automatic customer messages — the messages that send themselves
• Analytics and your customer data — the rest of what your own activity can tell you
• What Mosey costs — every rate on one page