The Cheapest Lead Is Already in Your Files: A Back-of-Envelope Look at Past Customers

Hi folks. My neighbor has been remodeling houses around DC for a long time, and nearly all of his work comes from referrals. Over the fence I have picked up a few things about how he thinks. He would rather not field calls about work that was never in the contract, but he takes them anyway, because in a referral business every customer is also a doorway to the next job.

That got me wondering about something I do not know the answer to: how does a business that lives on past customers stay in touch with them? A friendly "how is the deck holding up?" call sounds easy. It also sounds like something that slides to Thursday, then to next month, then to the great pile of things we meant to do. (I have a folder of those myself. It has its own zip code.)

So I spent the week on a question that sounds like a spreadsheet and feels more like a shoebox: what is a past customer worth, and what would it take to remember to ask?

What the repeat-revenue data says (and does not say)

The best-known number I could find comes from ServiceTitan. In a survey of 1,015 residential service contractors, repeat customers drove 58% of the work and 39% of revenue in 2022. The same report put word-of-mouth referrals at 71% of business volume. The research was run by Thrive Analytics on ServiceTitan's behalf and announced in November 2023.

A few honest caveats. ServiceTitan sells software to these exact businesses, so it has a stake in the story. The data is from 2022. The report does not say when the survey was fielded. And the press release and the report phrase the 71% figure a little differently, so I am going with the report's own wording (word-of-mouth referrals).

One thing I find interesting in the two repeat numbers: 58% of the work but 39% of the revenue. My reading, and it is only a reading, is that repeat jobs tend to be smaller than first-time jobs. That would fit a kitchen remodel followed by a faucet, a gutter, a punch-list fix. Smaller tickets, but with no ad spend attached.

Now compare that with a rule of thumb. In a February 2026 ACHR News guest column, Tracy Paul of Cornerstone Advertising wrote that a healthy home service business gets roughly 80% of its revenue from existing customers. She gives no source for it, and she is writing about HVAC, where service agreements pull the number up. So we have a measured 39% and a rule of thumb of 80%. They disagree by about a factor of two. I would not grade myself against either one. I would go count my own.

You may also have seen the claim that a 5% bump in retention lifts profit by 25% to 95%. I tried to trace it. The Reichheld and Detrick article behind it ("Loyalty: A Prescription for Cutting Costs," 2003) says a 5% retention increase produces more than a 25% profit increase in financial services, with no source for the figure. The 95% in that piece is Chick-fil-A's annual operator retention rate, which is a different thing entirely. So that one stays on the shelf for now. It is a good reminder that a number repeated often enough starts to look like a fact.

The back-of-envelope math

Here is a worked example. Every input is a placeholder, not a benchmark. I looked for a trustworthy industry response rate for past-customer outreach and found mostly vendor blog posts that do not explain their method, so I would rather you plug in your own.

The cost side is an afternoon to sort the list, a few short notes, and whatever time you spend on the phone. Now put your own cost per lead next to it. If 25 conversations from ads costs you more than an afternoon, that is the gap worth knowing about. If it costs less, you have learned something too. (And if your response rate is 1% instead of 5%, the same list still gives you around 5 conversations. Try it with your number.)

The other reason I like this math is what it does not need: no new tool, no new budget, and no one has to learn anything. The customers are already in your system, or your invoices, or your phone.

Where AI fits, and where it does not

The ACHR News and BNP Media homeowner and contractor survey (782 homeowners and 205 contractors, fieldwork April 24 to May 19, 2026) has two numbers that fit here. Half of the contractors said they do not currently use AI. Among those who do (102 of the 205), the most common use was follow-up communications, such as reminders and maintenance alerts, at 30%. It is an HVAC trade publication's own survey with a small contractor sample, so I read it as a hint about where people start, not a verdict.

Follow-up is a sensible place to start because it is mostly sorting and drafting, and AI is decent at both. Here is the division of labor I would try:

1. Sort. Export your customers with the last job date, job type, and amount. Ask an assistant to rank them: oldest happy customers first, grouped by job type.
2. Draft. Have it write one short note per job type. "It has been two years since the bathroom. How is everything holding up?" is about the right level of ambition.
3. Edit. You read every note. These tools have been confidently wrong about my own spreadsheets more than once, so a human checks the name, the job, and the year.
4. Send, in small batches. Five a day, from your own phone or email, so replies come to a real person and the conversation can be real.
5. Tally. Replies, quotes, jobs. A pencil works.

One practical note. Rules around bulk texting and marketing email differ from a personal call or a one-to-one note, so it is worth a quick check with your attorney or whoever handles your compliance before anything automated goes out. I am not a lawyer, and I would rather you hear that from me than find out later.

What I am not claiming

I am not claiming any particular response rate, because I could not find a trustworthy one. The ServiceTitan numbers are from 2022 and come from a vendor survey. The table above is arithmetic, not evidence. And I do not know how my neighbor, or any particular contractor, keeps in touch with past customers. That is a question I am curious about, not a conclusion.

One hour, this week

If you have an hour, here is the smallest version: pull your last 50 completed jobs, circle the 10 customers you would be glad to hear from, and reach out to two of them today. Write down what happens. After a week you will have your own response rate, which beats anyone else's.

Thanks to the folks whose work I leaned on: Thrive Analytics and ServiceTitan for the survey, Tracy Paul for the ACHR News column, Fred Reichheld and Christine Detrick for the loyalty article, and the ACHR News and BNP Media team for the homeowner and contractor survey.

I would love to hear from owners: how do you stay in touch with customers after the final invoice, and what do you wish happened more easily? If you would like a second pair of eyes on which past customers are sitting quietly in your system, that is the kind of thing we look at in a Massively Useful Revenue Audit. No pressure, and no pitch deck.