How to Get Past Customers to Come Back

The cheapest customers you will ever get are the ones you already had. Every local business is sitting on a list of people who bought once or twice and then quietly stopped, and almost nobody works that list. It costs nothing to reach, it converts better than cold traffic, and it is usually worth more per month than the ad budget.

This is the leak I check first, before missed calls and before reviews, because it is the one you can act on this afternoon without buying anything.

Who counts as a past customer?

Anyone who has not come back within roughly twice your normal visit gap. That definition matters more than a fixed number of days, because the gap is different for every business.

A barbershop customer on a four-week cycle who has not been in for ten weeks has drifted. A deck builder's customer at three years is completely normal. If you use a generic ninety-day rule for a business with an annual cycle, you will pester people who were never going anywhere.

Split the list into two buckets rather than one. People who are somewhat overdue need a light nudge. People who are long gone need a different message entirely, because pretending nothing happened reads as odd when it has been a year.

How much is the list actually worth?

Same approach as the missed-call arithmetic in what a missed call actually costs you. Count the people, multiply by your average ticket, then multiply by a realistic win-back rate.

I use 15 percent as the working assumption for a decent message sent to a list that has not been contacted in a while. That is an assumption, not a measured benchmark, and it will be wrong for your business in one direction or the other. If your list is old and cold, halve it. If you have a personal relationship with most of them, it can run higher.

Two hundred dormant salon customers at a $120 ticket and a 15 percent return is $3,600. That is not a rounding error for a business doing a few hundred thousand a year, and it came from a list you already own.

The reason this beats advertising is not that the message is better. It is that you skip the expensive part. You are not paying to introduce yourself to a stranger, because these people already know what you do, where you are, and what it costs.

Why did they actually stop coming?

Almost never because they were angry. Owners assume the worst and it is usually far more boring than that.

People move. Their schedule changed. Their regular stylist left and they never got round to rebooking with someone else. They meant to call in March and then it was August. Life happened and you were not part of the decision.

That matters because it changes the message. If you assume they are upset, you write something apologetic and defensive, which is strange to receive when the customer has not thought about you in six months. The right tone is casual, as though you noticed and thought of them, not as though you are owed an explanation.

A small share did leave unhappy, and you will hear from them when you reach out. That is useful. A complaint you can answer is better than silence you cannot.

What do you say?

Short. Personal. No discount in the first message.

The discount instinct is strong and it is usually a mistake up front, for two reasons. It trains the list to wait for offers, and it prices your regulars down for no reason. Plenty of dormant customers come back for a reminder alone, and you have given away margin you did not need to spend.

Try the plain version first:

"Hey Sarah, it's Devin at Runna. Realised it's been a while since we've seen you. No pressure at all, just wanted to say the door's open if you want to get back in. Here's the calendar if it's easier: [link]"

That is it. It works because it sounds like a person noticed, which is what a small local business can do and a chain cannot.

Save the offer for the second or third touch, aimed only at the people who did not respond to the plain message. That way you are discounting the hard cases rather than everybody.

What if my records are a mess?

They usually are, and it is a smaller obstacle than it looks.

Most booking software will export a customer list with last-visit dates, even if you have never used the reporting. If you do not have booking software, the list is still recoverable from your phone contacts, your payment processor's transaction history, past invoices, or your email. It is tedious rather than impossible.

This is one of the genuinely good uses for an AI assistant, because sorting a messy export is exactly the kind of dull work it handles well. Paste the export in and ask it to group customers by last-visit date, flag your top spenders, and separate the first-timers who never came back. You get a worked list in a couple of minutes instead of an evening in a spreadsheet.

One caution worth stating plainly. That export contains your customers' personal information. Do not paste names, phone numbers and addresses into a tool you have not checked, and strip identifying columns down to what you actually need for the sorting. You can do the grouping on initials and dates, then map it back yourself.

Who should you contact first?

Your former best customers, not the whole list at once.

Sending two hundred messages on day one means two hundred replies arriving while you are working, and a rushed reply is worse than no message. Start with the twenty people who spent the most and stopped. They are the highest value, and they are also the group most likely to respond warmly, which is good for your nerve.

Work down in batches you can actually handle. Twenty a day is fine. The list is not going anywhere.

What about people who enquired and never booked?

Different list, same principle, and often a better one. These are people who called, filled in a form, or sent a DM, got a quote or no reply at all, and then vanished. They raised their hand once. Nobody followed up twice.

Quoted work that went quiet is the richest version of this. An owner sends a number, hears nothing, and reads the silence as a no. Usually it is not a no. It is a maybe that got buried under everything else in that person's week. A single follow-up a fortnight later, with no new pitch in it, recovers a real share of those.

"Hey, just floating this back up in case it got buried. No rush at all, happy to answer anything if you're still weighing it up."

That is the whole message. It works because it gives them an easy way back in without making them explain the delay.

How do you keep this from feeling like spam?

By making it feel like it came from a person, and by respecting the rules around contacting people.

The tone half is straightforward. One message at a time, written the way you talk, sent from a number they can reply to. Not a template with their first name slotted in, and not four messages in a fortnight. If you would be irritated to receive it, they will be.

The rules half is worth taking seriously, because marketing texts in the United States are regulated and the penalties are not small. Broadly, you need a lawful basis to send marketing messages, a clear way to opt out, and you need to honour opt-outs promptly. Business texting also runs through a carrier registration process for the number you send from. None of that is a reason to avoid it, but it is a reason to set it up properly rather than blasting a list from your personal phone.

I am not a lawyer and this is not legal advice. If you are contacting a large list, it is worth twenty minutes with someone who is.

The safest starting point is people with an existing relationship who gave you their number for this purpose, contacted individually. That is also, conveniently, the group most likely to come back.

How often should you do this?

Once, properly, and then never again as a campaign.

The first sweep is a one-off cleanup of years of drift, and it produces the big number. After that the job changes shape: you want something that notices when a regular slips past their normal gap and sends the nudge automatically, so the list never builds up again.

That is the difference between a good weekend of work and a system. The weekend gets you the $3,600. The system stops the next $3,600 from ever accumulating, and it is the part most businesses never get to because it requires wiring the reminder into the booking software rather than remembering to do it.

Does the timing matter in Chicago?

Some. Salons and barbershops do well reaching out ahead of the holidays and through the spring wedding run, when people are already thinking about how they look. Gyms have January, though the more interesting window is late February when the new sign-ups have lapsed and are quietly embarrassed about it. A warm, no-guilt message lands well there.

Home services in the south suburbs run on weather. The message that gets ignored in December gets answered in April, because the customer is looking at their yard again.

None of this outweighs just doing it. A message sent at a mediocre time beats a perfect campaign you never send.

Frequently Asked Questions

How long should I wait before calling someone a lapsed customer?

Roughly twice your normal visit gap. For a barbershop on a four-week cycle that is eight to ten weeks. For a home services business with a seasonal cycle it might be two years. A fixed ninety-day rule will misfire on any business with a long natural gap.

Should I offer a discount to win someone back?

Not in the first message. Many dormant customers return for the reminder alone, and leading with an offer gives away margin you did not need to spend while training the list to wait for deals. Save the discount for people who did not respond to the plain message.

Is it better to text or email?

Text gets read faster and suits a short personal nudge. Email suits a longer message or a list large enough that replies would overwhelm you. If you text, make sure you have a lawful basis to do so and an easy way for people to opt out.

What if they stopped coming because of a bad experience?

You will find out, and that is a good outcome. Acknowledge it, do not argue, and fix it if you can. One recovered unhappy customer is worth more than the message cost, and the feedback tells you something your happy customers never will.

Can I use AI to write the messages?

To draft and to sort the list, yes. Give it three messages you have genuinely written so it matches how you talk, rather than letting it write from scratch. Be careful about pasting customer personal data into any tool, and strip the export down to what the task actually needs.

How many should I send at once?

Start with your top twenty former spenders and work in batches you can personally handle. Sending the whole list on day one produces a flood of replies while you are working, and a rushed answer undoes the goodwill the message created.

What is a realistic response rate?

It varies enough that any single number is misleading. Fifteen percent is a reasonable planning assumption for a warm list contacted personally, but treat it as a starting estimate to test rather than a promise, and re-run your own numbers after the first batch.