The short answer. Winning back former customers means contacting people who already bought and stopped coming. It's the cheapest sale in a service business: they know who you are and they already decided once that you're worth the price. It isn't hard. It's invisible. A new customer lands in your inbox. Someone leaving lands nowhere.
Old figure, and it's a thesis, not a promise. We cite it because you can open the authors' own summary and read the sentence yourself. The rest of this works whether you believe the number or not.
Losing a customer and having one drift off are different events. Losing one has a moment attached: a complaint, a job gone wrong. You see it, you deal with it. Drifting has no moment. They were due back, they weren't, and nobody mails a letter to announce it. Picture a gap in a line. Someone steps out, the line closes, and from the front nothing looks different.
1. Where the list actually lives
Owners who say "I don't have a database" usually do. They're looking in the wrong drawer.
They look in contacts, where the supplier, the wrong number and your cousin all sit at the same rank. A list where everyone is equal can't settle a single decision.
Go where the money was recorded. A payment leaves the two things a contact record never has: when and how much. That's the whole question answered. Who paid me and hasn't come back?
The reason it stays buried isn't laziness. It's that we open payment records to collect, not to sell. Which is how a business ends up buying ads to meet strangers while several hundred people who already paid sit two clicks away. That's the flip side of you rent the tool, you own the asset, which maps where that data lives and what happens when the software holding it stops answering. Today we're just pulling a list out of it.
2. Not follow-up, and the difference is the whole message
These get mixed up constantly, and it shows in the tone.
Follow-up is for the person who raised a hand and never bought. They don't know yet whether you're worth it, so the job is proof, objection handling, patience. That one's already built: the follow-up sequence template.
Reactivation is a different animal. They paid. They know how you work. They said yes once. There's nothing left to prove. All that exists is an absence neither of you mentioned.
So: reasons for the one who never bought, a reason to come back today for the one who did. Convincing versus reminding.
Take one distinction from this piece and take that one. Run the sales script at a former customer and it reads exactly like what it is. Two years of silence, then a pitch. It fails because you addressed someone who knows you as if they didn't.
3. Why they don't come back
The two answers everyone reaches for are a competitor and price. Sometimes true. Four duller reasons come first, and dull is good news, because dull is fixable.
None of the four contains a decision. That's the point. An absent decision gets corrected with a message, not a discount, and discounting is the most expensive move on the board: margin handed to someone who left by accident, plus a lesson to wait for it next time.
4. The three cuts, and who goes first
Nobody writes to everyone on the same day. You couldn't serve them if they all answered, and they don't all deserve the same message.
Cut 1. Had a rhythm, broke it. This is the idea that changes the exercise. The signal isn't elapsed time, it's a broken rhythm. Every three months and now it's been eight: signal. Every two years and now it's been eight: nothing. Same number, opposite meanings. Easiest cut, fastest return.
Cut 2. Spent big, once. Paid above your normal, never came back. Most valuable, hardest to write, because there's no rhythm to point at. Your reason has to be something new.
Cut 3. Long gone. Past two years. Most won't answer. Do it anyway, last, because typing it costs the same.
And the category nobody says out loud: take off the people you don't want back. Every business has them, no system records them, you know the names. Other exclusions belong in the guide, because you make those with the list in front of you.
Running one? Run Cut 1. It's the only one where honesty comes free, because you're stating something checkable: they haven't been in. The other two ask you to bring news.
5. The message that doesn't read like you need the money
Three things decide this, and none of them is phrasing.
Never: a discount in the first message, guilt dressed as affection, a hundred identical sends. Three lines naming the specific job they hired you for beat a polished paragraph that fits anyone.
This matters more than it did two years ago, because showing up where people search got expensive. Cited or bought prices both routes, and SEO vs. GEO vs. AEO sorts out which channel you actually need. People who already paid you are the one audience you never have to buy twice.
Where this returns little: a service bought once in a lifetime (that's referrals, different job); a business with no payment trail at all (start recording, that's a month); someone who left over something still broken (fix it first, or you hand them a stage). And the warning we care about most: if you can't take on more work right now, don't send. The guide covers batch sizes and why.
Naty and Todd's corner
Naty: Back when I had the stores, I learned the list is almost never the problem. The problem is that writing to someone who left feels embarrassing. You feel like you're going to sound like you're asking them for something. So the list gets built, saved, and not a single message goes out.
Todd: It helps to rename what you're doing. That first message isn't a sale, it's an update to your own records. If they reply that they don't need it anymore, you won. That name comes off the list. If they reply that it slipped their mind, you also won. The only outcome that's no use to you is continuing not to know.
Naty: That's what I'd say to anyone who's holding back. You're not asking them to come back. You're asking whether they're still your customer, which is a question you have every right to ask, and almost everyone answers it well when it's put that way.
Todd: And it helps to look at the bet straight on, because that's what takes the fear out: if you write, the worst that happens is they don't answer. That is exactly where you are with that person today. There's nothing to lose that you haven't already lost.
What you're doing today
Not elapsed time. Broken rhythm. That changes who goes first.
- Pull twenty, with a spread. Some gone a short while, some a long while. Take only the recent ones and you'll find no broken rhythms and conclude the wrong thing.
- Two columns beside each name: how often that person came back, and how long it's been. Wherever your payments live. Every three months, been eight. Every two years, been eight.
- Mark anyone past double their own interval. That's Cut 1, and it's a different set of people than a blanket threshold would hand you.
Eight months is loud in the first case and silent in the second. One threshold for everyone writes to the wrong people and leaves the right ones alone.
The full guide has the four places to look, the three cuts, and the three messages already written.
Keep reading
- SEO vs. GEO vs. AEO: which one your business needs in 2026 - where this sits in the wider picture
- The 7-Day Follow-Up Sequence Template - the other case, the person who asked and never bought
- You rent the tool. You own the asset - where your customer data actually lives
- Cited or bought - what each way of finding new people costs
- Zero Defections: Quality Comes to Services, Harvard Business Review, 1990 - the source for the figure above
Sources
- Reichheld, F. and Sasser, W. E. Jr., "Zero Defections: Quality Comes to Services," Harvard Business Review, September 1, 1990. Open summary by the author at Bain & Company, verified 2026-09-23: "Companies can boost profits by almost 100% by retaining just 5% more of their customers." https://www.bain.com/insights/zero-defections-quality-comes-to-services-harvard-business-review-hbr/
