Most service businesses are sitting on a list of people who have already paid them once, already know the work is good, and have not heard from them in two years. That list is the cheapest source of booked work available, and almost nobody works it.

The reason is not laziness. It is that reactivation feels like begging. Emailing someone from 2024 to ask whether they need anything is uncomfortable in a way that running an ad is not, so it never quite gets done.

That discomfort is a good instinct pointed at the wrong thing. The awkward version of this — "we miss you, here's 10% off" — deserves to feel awkward. There is a version that does not.

Why almost everything written about this does not apply to you

Search for customer reactivation and you will find material about subscription churn, dormant SaaS accounts and abandoned ecommerce carts. Win-back flows, RFM segmentation, reactivation emails with a discount code.

Very little of it transfers. A lapsed subscriber cancelled something. Your past customer did not cancel anything — they had a water heater replaced, it worked, and there was no further reason to speak. Nothing is wrong. There is no lapse to win back, and treating them as if there were is what makes the message read as desperate.

Segment by what happened, not by how long ago

The instinct is to sort by date and start with the oldest. That is backwards — the oldest are the coldest and the most likely to have moved.

Sort by what the job implies instead:

Things with a service interval. HVAC systems, water heaters, gutters, alarm systems, anything with a filter. If you installed it, you know roughly when it needs attention, and a message that says so is a service reminder rather than marketing. This is the highest-converting segment by a distance because the timing is genuinely useful.

Jobs that imply a next job. You replaced one window; there are eleven others the same age. You did the downstairs bathroom. Someone who has already spent money with you on a house that has an obvious next phase is the warmest possible prospect.

High-value customers, regardless of date. If someone spent a lot once, they are the kind of person who spends a lot. Worth a personal call rather than a campaign.

Everyone else. A single, low-key, genuinely useful message. Do not build a sequence for this group.

The first three are worth real effort. Most reactivation programmes fail because they treat all four the same and send everyone the same "we miss you" email, which the first three deserve better than and the fourth ignores.

What to actually send

The content test is simple: would this message be worth sending if you were not trying to sell anything? If yes, send it. If no, rewrite it.

Things that pass:

  • "Your system is coming up on the age where the heat exchanger is worth looking at. Not urgent. Here's what to watch for."
  • "We're in your area on the 14th. If you wanted the gutters looked at while we're there, it's cheaper than a separate trip."
  • "The part we used on your job was recalled. Yours isn't affected, but here's how to check."

Things that do not:

  • "We miss you!"
  • "It's been a while — here's 15% off."
  • Anything with a countdown.

The discount is the tell. Offering money off to someone who has no current need does not create the need; it just tells them their loyalty was worth less than a new customer's attention. If a discount is the only reason to send the message, there is no reason to send the message.

This is where reactivation goes wrong legally, and it is worth being careful because the list is large and the temptation is to text all of it.

A past customer's number was given to you for their job. That supports messaging about their job, and it does not automatically support marketing. A service reminder about equipment you installed sits comfortably on the responsive side of the line. A seasonal promotion does not, and needs consent you actually collected.

The practical consequence: your reactivation list and your marketing list are not the same list, and the difference is which consent each contact gave. We covered the general rule in what consent you need to text a lead who called you — the reactivation case is the same principle applied to a much bigger list, which is exactly why getting it wrong here is more expensive.

Email is the safer default for the whole list. Text is better for the segments where the message is genuinely about their equipment, and where you have the consent to use it.

Make it continuous, not a campaign

The most common shape of this is a one-off: someone exports the customer list, writes an email, sends it to everyone, gets a few jobs, and never does it again. That works once and then the list is burnt for a year.

The version that keeps working is a small, permanent trigger. Every month, the customers whose service interval has come due get a message about it. Nobody is contacted twice about the same thing. New customers roll into the pool as their own intervals mature.

That is duller and much better. It converts steadily instead of once, it never sends anyone the same thing twice, and it never has that "the whole list just got emailed" quality that makes people unsubscribe.

What it is reasonable to expect

Honestly: it depends entirely on the quality of your records.

If you know what you installed, when, and for whom, this is the highest-return thing you can do with an afternoon. If your customer history is a pile of invoices with no equipment detail, the first job is not reactivation — it is recording enough to make reactivation possible next year.

We are not going to quote you a return multiple. The numbers that circulate about win-back campaigns come mostly from subscription businesses and do not transfer to a plumber's customer list, and we would rather say that than repeat them. What we will say is that it is the only lead source where the acquisition cost is already sunk, and that is a structural advantage no ad channel can match.

How far back is worth contacting?

For interval-driven work, as far back as the equipment is still in service. For everything else, two to three years is a reasonable edge — beyond that, enough of the list has moved that you are writing to strangers about someone else's house. The Census tracks how many people move each year in Table A-1 of its historical geographic mobility series; it is worth a look before assuming a five-year-old list is still addressable.

Won't we annoy people?

You will annoy some. The way to keep that small is to make every message useful on its own terms and to honour opt-outs instantly — the carriers' own messaging principles treat a plain-language opt-out as a baseline rather than a courtesy, and it is the one thing in this whole exercise with no judgement call in it. A useful message to a past customer is far less annoying than a cold ad to a stranger, which is what most businesses spend money on instead.

Should we call or message?

Call the high-value segment. Message everyone else. A personal call to someone who spent a lot with you is not a marketing activity, and it converts at a rate campaigns do not.

What if we do not have good records?

Start recording now, and reactivate on what you do have. The businesses that get most from this are the ones that decided two years ago to write down what they installed.

More in lead generation.