Record Fixed-Ops Gross Hides Your Retention Problem

Average parts and service gross profit reached $5 million per dealership in 2025, up from $3.3 million in 2020. That is Kerrigan Advisors' figure, reported by CBT News on 20 August 2026 in a piece about dealers leaning harder on fixed operations as new-vehicle margin softens. If you run a service department, some version of that line is probably on a wall near you, and it should be. It is a real number, earned by real people working real Saturdays.

It also cannot tell you whether you are winning.

Set it next to a second finding. Cox Automotive's study of the service market found that dealerships captured record fixed-operations revenue even as their share of service visits slipped to independent and general repair shops. Both are true at once. Together they say something no profit-and-loss statement is built to say out loud: a service department can post its best year on record while quietly losing the fight for the customer.

That is possible because gross is a mix number, and mix is invisible in dollars.

Three kinds of work, and only one of them is yours

Work that arrives because of who you are. Warranty. Recall campaigns. Manufacturer-directed service. The customer is in your drive because a letter told them to be, or because only a franchise store can do the job. You did not win that customer, you cannot keep them by doing the work well, and the volume ends when the campaign ends. It is weather, not climate.

Work that arrives because the car is old. The average vehicle on US roads is approaching thirteen years. Older cars need more, and what they need costs more, so the average repair order climbs without a single customer deciding anything differently about you. Ticket inflation raises gross and moves no loyalty at all.

Work you actually won. Somebody with a nine-year-old car and an independent shop three miles closer chose your store anyway. That is the only stream that compounds, and the only one that is genuinely yours.

The five-step earned-share audit

Every step below uses data you already own.

One: split the repair-order count by who sent the customer

Four buckets. *Manufacturer-directed* covers warranty, recall and campaign work, anything the customer came in for because a letter or a portal told them to. *Store-directed* covers customers who came because you contacted them: a declined repair you followed up, a first-service reminder, a maintenance interval, a missed appointment you chased. *Customer-initiated* covers the ones where something broke and they chose you. *Comeback* covers work you are doing again because it was not right the first time.

All four already exist in the dealer management system. They are simply never reported side by side.

Two: track the ratio, not the total

Put one number on the wall next to the gross figure: store-directed plus customer-initiated repair orders as a share of all repair orders. Call it earned share. Run it monthly for twenty-four months.

Total gross can climb for eight straight quarters while earned share falls the whole time, and most stores would currently have no way of knowing.

Three: cohort the return, do not count the visits

Take every customer who bought a vehicle twenty-four months ago. What share came back for their second paid service? Not the first, which is often prepaid or promised at delivery and therefore flatters you. The second visit is the honest one, because by then nothing is holding them but the experience.

Run the same logic on service-only customers: of the people whose first visit was a repair, what share returned within a year?

Four: find the moment, not the month

Defection rarely appears as a gradual slope. It clusters at dated events:

• The end of the prepaid maintenance plan.
• The first declined repair that nobody followed up on.
• The appointment somebody could not get inside the window they needed.

Each is a timestamped row in a system you own. Pull the customers who hit each event in the last year and check what share ever came back.

Five: put a real number on the phone

The fourth moment leaves no record at all, because a call that never connects never reaches the scheduler. Most service departments know their answer rate roughly and their abandon rate approximately. Get both exactly, by hour, across ninety days, and lay them over the hour-by-hour curve of when people actually try to book. Where the two curves diverge is the leak in its most literal form, and it is usually not where anyone guesses.

The honest counter-case

A record gross year is not a trick. It pays technicians, funds equipment and buys time to fix what sits underneath.

There are also legitimate reasons earned share can fall in a healthy store: an unusually large campaign year, a newly acquired rooftop still stabilising, or a market where an independent genuinely is the right answer for a twelve-year-old commuter car. A falling earned share is not automatically an emergency.

The argument is narrower. You should know which way it is moving, and today most stores cannot say, because the number that would tell them has never been assembled.

Size it before you decide it is worth a morning

A store writing 900 repair orders a month with 30% manufacturer-directed is carrying roughly 270 orders of cover it did not win. At a $400 average repair order that is about $108,000 a month riding on campaigns with an end date.

Knowing that a quarter early is a completely different conversation from discovering it in the quarter it lands.

For anyone building or buying software for this: the layer that should own the earned-share number is whoever can already see both the phone and the scheduler at the same time. Very little in the current stack can. That gap, not another dashboard, is the interesting part of this market.

So before the next fixed-operations review, do not open with the gross. Open with the ratio, and make somebody say out loud which direction it moved.