There’s a model everyone in fixed ops carries around, usually without examining it.
It goes: repairs get expensive, expense accumulates, and somewhere there’s a number — a repair estimate big enough that the customer stops fixing and starts shopping. Find the number, watch for it, and you’ll know when to have the trade conversation.
It’s a tidy model. We went looking for the number across two years of our own dealer network’s repair orders, sales and retention data — tens of thousands of vehicles.
The number isn’t there.
Trade probability was essentially flat across every repair amount from a few hundred dollars to well past seven thousand. No threshold. No inflection. We ran it by single repair order and by cumulative annual spend, and got the same flat line both ways.
That result is either a dead end or a clue, depending on what you do next.
The question was wrong
Here’s what we’d been measuring: did this customer trade their vehicle with us?
That turns out to be a poor proxy for did this customer give up on their vehicle, for a reason that’s obvious in hindsight. Trading with us requires a sales relationship, an equity position that permits it, and a customer who wasn’t going to trade on a two-year cycle anyway. Strip those out and you’re measuring our commercial relationship, not the customer’s confidence.
So we changed the question to one we can actually observe for everybody: did this customer stop coming?
Every vehicle in the service base can answer that. No sales relationship required. No equity gate. And when we asked it, the flat line turned into two very clear signals pointing in opposite directions.
Two exits, not one
The number of trips predicts leaving the vehicle. More visits, holding spend constant, meant a substantially higher chance the customer traded the car.
The size of the bill predicts leaving the relationship. More repair spend, holding visits constant, meant a substantially higher chance the customer simply never came back — and lower odds they traded with us.
Read those together and the picture is uncomfortable. A customer who absorbs a large repair bill doesn’t become more committed to the vehicle. They disappear. They don’t come back to service, and they don’t buy their next car from us. We collected once and lost the relationship.
That’s why the threshold model fails. The dollar figure everyone is hunting for is real — it just predicts the wrong exit. It’s not the moment a customer decides to replace the car. It’s the moment they decide to stop dealing with the dealership. Nobody describes their own behavior that way on a survey, which is exactly why survey-based models miss it.
And once you’re looking for quiet departures instead of loud trades, three signals show up. All three are early. All three are computable from data every dealership already has.
Signal one: the first visit that never happened
Of the vehicles our network sold in a single year, roughly two out of three never came back for a paid service visit in the first twelve months. Not one oil change.
That population is effectively already gone. Vehicles that skipped year one were about ten times less likely to still be servicing with the dealership in year two than vehicles that came in even once.
This lines up with what the industry has been told from the other direction. Cox Automotive’s 2026 study found 80% of new-vehicle buyers intend to service with the selling dealership, while only 23% leave with a first appointment booked. Two completely different methods, the same finding: the relationship is lost at the handoff, not at the repair.
And here’s the part that should change how you chase it. Vehicles whose first service came within four months of purchase retained worse in year two than vehicles whose first visit came at eight to twelve months. An early first visit usually isn’t enthusiasm — it’s a problem. The goal isn’t to get them in fast. It’s to get them in on the interval.
Signal two: off their pattern
Every vehicle develops a rhythm. Some customers appear every three months, some once a year. That’s their normal, and it’s specific to them.
Now consider two customers, each last seen 200 days ago. One normally comes every 120 days. One normally comes every 365. The first has broken a four-month habit — something changed. The second isn’t due yet.
Same 200 days. One is a warning, one is nothing. A mileage reminder can’t tell them apart — it nags the second customer for no reason and treats the first customer’s broken habit as routine.
Measured against their own history, the difference is stark: when a customer falls behind their own rhythm, the odds they never return climb by roughly thirty points — and the effect held at nearly identical magnitude across every level of repair spend.
They haven’t stopped needing service. They’re getting it somewhere else.
The useful thing about this signal is that it’s gradual. It isn’t an event that fires after the damage; it’s a slope you can watch. Which means there’s a long window in which a customer is only slipping, and can still be caught.
Signal three: the stranger repair
This is the one that reframes the repair bill entirely.
Take the same expensive repair and two different customers. One has been coming to you for routine maintenance. The other you’ve never seen except when something breaks.
The customer with no maintenance history mostly disappeared. The customer already in a maintenance rhythm mostly stayed — a difference of ten to twenty points of retention, replicated across two separate cohort years and larger in the second.
Same repair. Same dollars. Radically different outcome, decided by something that happened before the repair ever arrived.
It makes sense once you say it plainly. If the only time a customer sees you is when their car is broken, then you are the bad news — you’re the building associated with the worst days of ownership. A customer in a maintenance rhythm has a routine relationship, and the repair is an episode inside it rather than the whole of it.
A big repair bill doesn’t lose the customer. A big repair bill from a stranger does.
Which means the intervention isn’t discounting the repair, and it isn’t reacting faster when a big estimate lands. By then the outcome is largely set. The work happens months earlier, at an oil change.
What the three have in common
None of them are loud.
Not one of these signals shows up as a complaint, a bad survey, or a customer asking for their keys back. A customer who never books their first service isn’t angry — they just never got around to it. A customer drifting off their own rhythm isn’t upset — they went somewhere closer, once, and it was fine. A customer who pays a big repair bill and vanishes may have been perfectly polite at the counter.
Every one of them, though, is visible in advance. First-service status is knowable the day a vehicle is delivered. A cadence break is measurable the week it happens. A repair landing on a customer with no maintenance history is knowable before the estimate is written.
That’s the whole opportunity: every customer you lose gives you months of warning. The warning just doesn’t arrive as a phone call. It arrives as an absence, and absences don’t page anybody.
One aside worth its own conversation
While we were in the data we found that roughly one in ten records in the service feed isn’t a customer visit at all — it’s dealer-internal work: pre-delivery inspections, used-car reconditioning, detail, carwash on inventory.
If those are sitting inside your repair-order counts, your retention reporting, or your per-advisor numbers, every one of those figures is flattering you by about ten percent. Worth an afternoon to check.
Where this leads
Rising ownership costs are real. Insurance is up, repair costs are up, loan terms have stretched, and more customers are underwater than at any point in years. All of that is true and none of it is the mechanism.
The mechanism is quieter. Customers accumulate small experiences — a first service nobody asked them to book, a rhythm that slipped and nobody noticed, a big bill from a building they only visit when something’s wrong — and at some point they stop. Not dramatically. They just stop.
Every one of those moments is a place where a personalized roadmap — showing what’s coming, when, and what to do about it — is worth more than a reminder blast. Not because it’s cleverer, but because it’s watching the right thing: the individual customer’s own pattern rather than a mileage table.
The road ahead for vehicle ownership is clear.
