The cheapest leads you will get this year are in a folder nobody has opened.

Every local business we talk to wants more leads, and most of them are sitting on a list of people who have already paid them money at least once. Job records going back years. Invoices. A CRM somebody stopped updating in 2023.

It is the most valuable marketing asset in the business and the one nobody has a plan for.

Service history beats marketing data

What makes that folder different from a list you could buy is that it is not a list of people who might want something. It is a record of what specific people already bought, when, and for which property.

That is a much better starting point than any targeting you could pay for. You know the boiler you installed is five years old. You know that gutter clean was two autumns ago. You know which customers have a warranty about to lapse. None of that requires guessing at intent, because the intent is written into the job you already did.

Let the work set the timing

The failure mode of reactivation is treating it as a campaign, which means it happens when marketing feels like it, rather than when the customer has a reason to care.

The better trigger comes from the job itself. Service intervals, seasonal work that recurs, a warranty ending, a part with a known replacement window. Those messages do not read as marketing because they are not marketing; they are the maintenance reminder somebody would thank you for, which happens to produce work.

It also solves the awkwardness problem. "It has been about two years since we cleaned your gutters" needs no justification for arriving. "We miss you" needs quite a lot.

Consent does not transfer, and it does not age well

Here is where reactivation advice usually goes quiet. Somebody who gave you a mobile number in 2021 so a technician could text on the way did not thereby agree to receive marketing messages in 2026.

The rules governing business texting require consent obtained in accordance with applicable law, and they require you to be able to produce evidence of it on request. A service transaction is not that evidence. The consent and registration side is worth reading before anyone builds a sequence aimed at a five-year-old contact list, because the sequence is the easy part and the list is where the exposure is.

In practice this makes email the sensible first channel for older contacts, and makes the phone the sensible channel for the ones you have a genuine relationship with. Calling twelve people you actually remember is not a worse campaign than texting eight hundred you do not. It is frequently a better one.

Why the mass email costs more than it looks

The instinct is to export everything and send one announcement. Worth understanding the arithmetic before you do.

Gmail asks every sender, whatever their volume, to keep spam complaints below 0.3%. That is three complaints per thousand delivered messages. Now consider who is most likely to hit the spam button: not your regulars, but the people who have not heard from you in four years, who do not recognise the sender name, and who never expected to hear from you again.

A single enthusiastic blast to a stale list is the fastest way to teach a mail provider that your domain sends unwanted mail, and the customers who pay for that are the ones who did want to hear from you. The deliverability rules are worth understanding first, because this is the one marketing mistake that degrades everything you send afterwards.

What a sensible first pass looks like

Start narrow enough to learn something. Pull the customers from one service type in one time window, ideally the group with the clearest reason to hear from you, and send to a few dozen rather than to everyone.

Give the message one job and make the reason for it obvious in the first line. Then watch what comes back before you scale. Replies tell you the list is warm; unsubscribes tell you it is cooling; complaints tell you to stop immediately, because that number is the one with a ceiling attached and you have just found out where you stand on it.

If that batch goes well, do the next segment. If it does not, you have learned it on forty people rather than on eight hundred, and the asset survives to be used properly later.

That asset only exists if you can actually get at it, incidentally, which is worth confirming while you are in there. A customer list you cannot export is somebody else's.

If you have years of job history and no idea what is safe to send to it, book a strategy call. Working out which slice of the list has both a reason and a right to be contacted is the whole job, and it is considerably cheaper than buying the same customers a second time through advertising.