When to Stop Chasing a Dormant Customer

Why this matters

Reactivation is sold as free money, so shops start a dormant campaign, get a decent first result, and then keep calling the same non-responders for months because stopping feels like leaving something on the table. The office time is real and it comes out of the same hours that answer the phone for customers who are actually trying to reach you. A dormant list has a genuine yield and it is front-loaded and steep. Knowing where the curve flattens is what separates a reactivation program from a shop quietly running a call center at a loss.

Three states, and only one of them earns real effort

"Dormant" gets used for everything from a customer who is one month past due for a seasonal visit to one who moved out of state in 2019. Sort the list first, because the stopping rule is different for each.

Anchor "dormant" to that customer's own median gap, never to a fixed number of days across the whole list. A 200-day silence is alarming on an account that historically called every 6 weeks and completely unremarkable on one that calls every other spring. A single day-count threshold applied to a mixed list generates so many false positives that the office stops trusting the report.

Four conditions that stop the chase immediately

These override any touch budget. None of them are about effort running out.

1. They said no once, clearly. "We are using someone else now" or "please stop contacting us" ends it. Mark do-not-contact and mean it. A second ask after a clear no does not read as persistence, it reads as not listening, and it is the version of this that generates a bad review.
2. There is unresolved money between you. An open dispute or an unpaid balance has to be settled on its own terms before any work solicitation, or you look like you are trading the dispute for the job.
3. The reason for the relationship is gone. Property sold, business closed, they moved outside your area, or the equipment you serviced was replaced by something you do not work on. Confirm the fact rather than assuming it, then retire the record.
4. You let them go, or you should have. Reactivating a customer you deliberately stopped serving is a separate decision with its own criteria, and it is not a list-management question. See the related article on taking back a customer you let go.

The decay to expect, and the one thing that resets it

In a no-rupture dormant sequence carrying essentially the same offer each time, expect each successive touch to return somewhere between a third and a half of the prior touch's response rate. That is a planning assumption, not a law, and it holds specifically when the message is a variation of the same ask. It is what makes touch four almost always a waste.

The exception is worth building your program around: a touch with a genuinely new reason does not decay, because it is not the same offer again. A safety notice affecting equipment you know is on their property, a code change that affects their installation, or a real service-due date derived from your own history of what you installed and when - these get read as information rather than solicitation, and they perform like a first touch no matter how many generic ones preceded them. A shop that has run out of new reasons has run out of touches, regardless of the count.

Worked example: 412 dormant records, three touches

A shop pulled every customer with no ticket in 24 months or more and got 412 records. It ran three touches over ten weeks.

Touch 1 was a mailed and emailed service-history note, personalized with what the shop had actually done at that address and when. Sent to all 412. 34 responded, 19 booked. Against the 412 sent, that is an 8.3% response rate and a 4.6% booking rate.

Touch 2 was a phone pass to the 378 who had not responded. 14 responded, 6 booked. Against the 378 called, a 3.7% response rate and a 1.6% booking rate.

Touch 3 was a second mailing to the 364 still silent. 4 responded, 1 booked. Against the 364 mailed, a 1.1% response rate and a 0.27% booking rate.

Across the whole program, 26 of the 412 records booked, 6.3% of the list. Of those 26 bookings, 19 came from touch 1, which is 73% of all bookings from the first contact. Touches 2 and 3 together produced 7 of the 26 bookings, 27%, and they required 742 additional contact attempts to do it.

Now the office hours, which is where the decision actually lives. Touch 1 was a mail merge off the service history: about 4 hours of list cleanup and preparation for the whole 412, producing 19 bookings, so roughly 0.2 office hours per booking. Touch 2 was 378 phone attempts averaging about 3 minutes each including no-answers and voicemails, about 19 hours for 6 bookings, or roughly 3.2 office hours per booking. That is about 15 times the office hours per booking that touch 1 required, for the same shop, the same week, the same list.

Touch 3 cost about 2 hours of preparation plus print and postage for 364 pieces, and returned 1 booking. Per booking it looks cheaper than the phone pass on office hours alone, which is exactly the trap: the material and postage on 364 pieces is a real outlay that hours do not capture, and a single booking is inside the range of random noise. One result is not a measurement.

The segment cut that changes the answer

The same shop split the 412 by service history before drawing conclusions, and this is the number that reshaped the program.

96 of the 412 records had 3 or more prior tickets. Within that group of 96, touch 1 got 17 responses and 11 bookings, an 11.5% booking rate against those 96. The remaining 316 records had one or two tickets. Within that group, touch 1 got 17 responses and 8 bookings, a 2.5% booking rate against those 316.

So a record with three or more prior tickets booked at roughly 4.6 times the rate of a record with one or two, on the identical message in the identical week.

That reorders everything. The correct program is not three touches to the whole list. It is:

• The 3-or-more-ticket segment gets the expensive channel first. Phone them at touch 1, when response is highest, rather than at touch 2 when it has already dropped by more than half. This is the group where 19 office hours might be worth spending.
• The 1-or-2-ticket segment gets one low-cost touch, then goes to the annual list. On the numbers above, the entire phone-pass and second-mailing effort applied to that segment cannot pay for itself.

The general rule underneath: stop deciding how many touches a list gets and start deciding how many touches a segment gets. A flat sequence spends the same effort on a customer who used you six times and one who used you once, and those two are not the same asset.

Retiring a record without deleting it

Stopping the chase is not the same as ending the relationship, and deleting the record is almost always wrong. Dormant customers come back through the door you cannot predict: the same equipment fails again, and when it does, the shop with the documented history of that exact installation is the obvious call. Throw the history away and you throw away the only advantage you had over the competitor whose number they will otherwise search for.

Retirement means: mark the record dormant with the date and the last touch, remove it from active campaign lists so the office is not working it, and put it on one touch a year, attached to a real event - a seasonal service window on equipment you know is there, a notice that genuinely affects them, or nothing at all if you have nothing real to say. Silence is an acceptable annual touch. A manufactured one is not, and it is how a retired list turns into a spam complaint.

Keep whatever data-retention and contact-consent rules apply to you in view when you do this; holding a record indefinitely because it might be useful is a different question from contacting it, and the two decisions should be made separately.

What changes the answer

If your service has a long natural cycle, three or more years between normal calls, the entire framework stretches and a 24-month silence is not dormancy at all. Anchor to the service cycle, not to a generic two-year mark, or you will run reactivation campaigns at customers who were never gone.

If your shop is at capacity, stop the program entirely rather than tuning it. Reactivated customers arrive with an expectation formed by the outreach that invited them, and a shop that cannot schedule them for three weeks converts a dormant customer into an actively annoyed one. Reactivation is a demand tool for soft periods.

If you have changed something real - new service line, new coverage area, a genuine capability you did not have when they left - the decay assumption resets, because you now have a new reason rather than a repeat of the old ask. That is a legitimate restart of the sequence, and it is the only legitimate one.

If the list is small, under about 50 records, the percentages above are not meaningful and neither is any conclusion you draw from a single campaign. Work a small dormant list by hand, account by account, on what you actually remember about each one. Statistics are for lists big enough to have them.

How to verify you got this right

Track two numbers per touch, not one: bookings, and office hours consumed. Bookings alone will always argue for one more touch, because one more touch always produces a booking or two. The ratio is what tells you when to stop, and the moment a touch's office-hours-per-booking exceeds what you would spend to acquire a customer through your normal channels, that touch is a loss you are recording as a win.

The second check is a year later. Of the customers a campaign reactivated, count how many came back a second time without being chased. A reactivation program that produces one ticket and then silence has bought a job, not a customer, and it should be budgeted as advertising rather than as retention.

References

• U.S. Small Business Administration (SBA), small-business marketing and customer retention guidance
• Trade-standard practice for service-interval tracking and equipment-based recall notification
• See related: The Lost Customer Win-Back Checklist, Deciding Whether to Take Back a Customer You Let Go, Customer Churn Reasons + Recovery Strategy, The Customer Data You Store: Protecting It