How to win back lapsed customers, ten names at a time

If you want to win back lapsed customers, the job is smaller than the advice usually makes it sound. Decide what "lapsed" means for your repeat cycle, pull the names that crossed that line, and write to ten of them personally over the next two weeks. Ten messages in your own voice, sent one at a time, each one referencing something true about that specific person. Then count how many walk back in within a month.

That last step is where owners quietly cheat themselves. A blast to 500 addresses feels like progress because you get a send confirmation to look at, and a send confirmation is a very satisfying thing to look at. It's still a worse hour than five careful texts. The people who already trusted you once are the cheapest revenue you will ever find, and nothing burns that goodwill faster than a message that obviously went to everybody.

I'll use one business the whole way through. Dana runs Hillcrest Grooming: two tables, roughly 40 dogs a week, $78 for a full groom. Her Tuesdays and Wednesdays between 10 and 2 are dead. Her Saturdays have a waitlist.

What counts as lapsed depends on your repeat cycle

There is no universal number of days. Lapsed is relative to how often a healthy customer of yours shows up.

Dana's regulars come every 6 to 8 weeks, so a dog she hasn't seen in 12 weeks has gone somewhere else. A café regular who normally comes twice a week is lapsed after three. A landscaping client isn't lapsed until they've skipped a season. A wedding photographer has no lapsed customers at all, which is a different problem entirely and not one this post solves.

Pick your number by taking your repeat cycle and roughly doubling it. If you don't know your cycle, sort your last hundred transactions by customer and eyeball the gaps. Ten minutes of that beats any benchmark you'll read online. Write the number down, because the next sweep has to compare against the same line or the comparison means nothing.

Before you panic at the size of your list, check the calendar. Dana's lapsed count looks alarming in February and completely fine in May. Nobody books a full groom in the dead of winter, and everybody books one the week the shedding starts. Same customers, different month. I have watched owners rebuild an entire pricing model around a seasonal dip that would have corrected itself by April, which is an expensive way to learn what a season is.

Ten names you work by hand beat five hundred addresses

Pull everyone past your threshold from whatever records you keep: booking software, the POS, a shoebox of invoices, your own memory if that's all there is. Dana's booking system gave her 61 names past 12 weeks.

Then split the list. These people are not the same problem.

The first pile is regulars who faded. Three or more visits, then nothing. Dana had 14. A dog on a 7-week cycle is worth about $580 a year, so this pile is real money sitting in a spreadsheet. Each one gets a message from Dana herself, written individually.

The second pile is one-timers who never came back. There were 39 of those, more than half her list. Lower value each, so lower effort each. One short message, then leave them alone.

The third pile is the awkward one: anyone whose record shows why they left. A complaint. A job you botched. Dana had 8, including a woman whose poodle went home with a nicked ear. Handle those personally or decide on purpose to leave them alone. Both answers are defensible. What you cannot do is drop them into a cheerful "we miss you" send and pretend it never happened. They remember the ear. You're the one who forgot.

Work the 14 first. If that pile is the only thing you ever get to, you've captured most of the available money.

Write like a person who remembers them

The first message to a faded regular needs no offer. Being remembered is the offer.

Here's what Dana sent, one at a time, over about three days: "Hi Karen, it's Dana at Hillcrest. Haven't seen Moose since April, hope he's doing alright. I've still got the oatmeal shampoo he likes if you ever want to bring him in."

That's the whole thing. There's no discount in it and nothing that looks designed. It went out as a text, so there wasn't even a subject line to write. It reads like what it is, a person who noticed.

One-timers get something shorter with one concrete reason to come back. Something new since they visited works better than money off. Dana used a nail grinding setup she'd added in the spring, which is exactly the thing owners of squirmy dogs care about and roughly nothing to anybody else.

A few rules I'd hold to. Never write "we miss your business," because that's a sentence about you. Reference the specific animal, order, or job if your records support it, and skip personalization entirely if they don't, since a wrong detail lands worse than no detail. Two touches maximum, about two weeks apart. After that, silence is an answer, and a third message mostly buys you a mute.

Point returning customers at the hours you need filled

Almost everyone skips this part. You're already asking someone to come back, so you may as well say when.

Dana's messages ended with a nudge toward Tuesday or Wednesday midday, because those slots were empty and her Saturdays were already full. "I've got Tuesday and Wednesday mornings open the next few weeks if either works." Nobody reads that as pushy. It reads as helpful, and it moves demand into the window that was costing her money.

If your slow period is January, run the sweep in December. If it's the 2pm lull, say 2pm. A win-back that fills an already packed Saturday earns you close to nothing, and you spent your one good message doing it.

Measure returns instead of sends

Open rates tell you nothing. Count how many lapsed customers actually came back inside your window, which is a month for frequent-visit businesses and a quarter for slower ones.

Track it the dumb way. A spreadsheet with name, date contacted, and whether they showed. Or a code word: "mention this text and I'll throw in the nail trim." Or the owner recognizing a face and making a mark on a notepad.

Dana's numbers after six weeks: 6 of 14 faded regulars rebooked, 4 of 39 one-timers came in. Ten returns, about $780 in the door immediately, and six people back on a cycle worth roughly $3,500 a year if they hold. Look at the split, though. Fourteen personal notes produced six visits. Thirty-nine messages produced four.

Read the list as a diagnosis

Dana got something out of this worth more than the $780.

Thirty-nine of her 61 lapsed names were one-timers. That's a first-visit problem wearing a win-back costume, and no amount of texting fixes it from the back end. People came once, had a perfectly fine time, and never got around to booking again.

So she changed the checkout. Every new client now gets asked to book the next appointment before they leave, and she writes the date on the receipt in pen. Four months later the one-timer pile is noticeably thinner. That change came out of reading the list, not out of sending the messages.

Then write down what you learned: your threshold, the date you ran the sweep, who you contacted, and who came back. Boring, and it's the difference between a one-off sweep and a habit. Next time you're measuring against something instead of starting from zero, and nobody gets the same "haven't seen you in a while" text twice in one year.

Scaffle is an AI advisor for businesses. It learns how yours actually runs and keeps watch through the tools you connect, so a customer list going quiet reaches you while it's still ten names. It's at scaffle.ai.