CNBC published the itemized version in September. The average American driver spends $5,851 a year owning a vehicle before a single loan payment: $2,237 on insurance, $2,126 on fuel, $1,488 on maintenance and repair. Add Experian’s average new-car payment of $765 a month and the year runs about $15,031.
Every article written off that data stops there, because it’s written for the person paying the bill.
For a dealer the interesting question isn’t what the year costs. It’s what share of it lands at your store.
That’s share of wallet. It’s a standard metric in almost every other retail category, and almost nobody in automotive has actually measured theirs.
So we measured ours.
The answer is about 28%
Across our dealer network, a customer returns roughly $421 of that $1,488 to the dealership over a year in customer-pay service work.
That’s 28%. Two dollars in seven.
The rest — call it $1,067 a year, every year, for every customer in your database — goes to an independent shop, a tire store, a quick-lube, or a brother-in-law with a lift.
Nobody at your store ever sees the invoice.
Read the other way, that’s the opportunity: roughly seventy cents of every maintenance dollar your existing customers already spend is still on the table, with people who have your name in their glovebox.
Why your own number may look friendlier
Pull this at your store and you’ll probably get a cheerier answer, for two reasons worth knowing about.
The first is who you count. Measure spend per visiting customer and you have quietly deleted everyone who didn’t come in this year. Your share of wallet roughly doubles. It’s the most common way this metric gets flattered, and it’s the same move that makes retention look better than it is.
The second is who you don’t exclude. Commercial accounts wreck an average. In our data, eight-tenths of one percent of customers hold 48.6% of all deals — fleet accounts, commercial accounts, dealer accounts. On the service side, a tenth of a percent of customers account for 6.9% of all service revenue. Customers with one to four visits a year are 97.3% of the base.
Leave those accounts in and you are describing a customer you almost never see, not the one sitting in your service drive this morning.
This is also, we suspect, why published estimates of dealership customer lifetime value range from $3,000 to $175,000 depending on which vendor blog you read. A 58x spread isn’t disagreement about the business. It’s a handful of fleet accounts nobody remembered to exclude.
The number the industry publishes is not the number it says it is
If you’ve read much fixed-ops content you’ll know the number: a “63.9% national average service retention rate,” quoted everywhere — vendor guides, conference decks, board presentations.
We can tell you exactly what that number measures, because we’ve been running it every month for two years.
Across our dealer network the same calculation has held between 70% and 72% every single month for twenty-two consecutive months, averaging 71%. No seasonal swing, no drift. August looks like February.
So on the industry’s own preferred metric, our dealers sit several points ahead of the published average. We’d still rather tell you why it’s the wrong metric.
That calculation answers “how many of today’s customers are familiar faces?” It does not answer “how many of last year’s customers still belong to us?”
Those are different questions about different populations, and they do not produce the same number. The second one — the one everybody thinks they’re quoting — comes out substantially lower, because it counts every customer who quietly stopped coming. The first one never sees them; they aren’t in the service drive to be counted.
A store can watch it hold steady for two years while its actual book of customers erodes underneath.
So the industry’s headline retention statistic is a traffic statistic. It describes the mix of people who showed up, not the share of a customer base you held onto.
The other number you’ll see quoted — NADA’s 72% to 75% — isn’t a measurement at all. It’s a benchmark. A target. It was never a description of what stores achieve.
Which means the two figures anchoring most retention conversations in this industry are a traffic mix and an aspiration, being compared to each other as though they measured the same thing.
Before you benchmark your store against either one, make sure you know which question you’re answering.
What actually moves it
Two things showed up clearly, and neither is a threshold.
Frequency, which dominates everything. A customer who visits five or more times a year is roughly 2.3x more likely to still be with you a year later than a customer who comes in once. Nothing else we measured moves the needle that far.
It’s worth being precise about the shape: it’s a gradient, not a cliff. There’s no magic visit number where a customer converts from at-risk to safe — the curve climbs smoothly. This matches what we found when we went looking for a repair-bill threshold and came back with a flat line. Nobody leaves over a repair bill, and nobody becomes loyal on their third oil change.
Free work, which is more interesting than it sounds. In 2026, 22.6% of our closed repair orders billed nothing — prepaid maintenance and complimentary service. Read the op codes and they’re exactly what you’d guess: DEALER FOR LIFE LOF, TOYOTA CARE 5K SYN, NEW OR USED DELIVERY, RECALL — COMPLIMENTARY INSPECT.
Nearly a quarter of your bays are running work that bills nothing.
We split customers by whether any of their prior-year visits was one of those complimentary visits, and controlled for visit count so we weren’t just re-measuring frequency:
| Prior-year visits | Difference in return rate |
|---|---|
| 1 | +3.8 pts |
| 2 | +1.9 pts |
| 3–4 | +3.8 pts |
| 5+ | +6.2 pts |
Better retention in every band, and the effect is largest among your best customers.
Giving work away looks like the least profitable thing your service department does. On this data it’s one of the few levers that reliably keeps the customer.
Why this is an automation problem
Look again at the two things that actually moved the number. Neither is something you do once.
Frequency isn’t a campaign. It’s a customer having a reason to come back four or five times a year, for years. And complimentary maintenance isn’t generosity — it’s a scheduled reason to return that you already paid for.
Both are the same mechanism: being in front of the customer, repeatedly, with something specific to their vehicle.
Now cost that out by hand. A store with eight thousand active service customers has eight thousand vehicles on their own maintenance clocks, their own warranty expiries, their own equity positions. Most of those conversations should happen in a month when the customer isn’t in your drive and nobody at the store is thinking about them. No BDC staffs that.
So it doesn’t happen — not because anyone decided against it, but because the only version anyone has ever been able to execute is the one where the customer initiates.
That is the 72%.
VehicleLyfe is the other side of it. Every owner gets a Horizon Owner Report: a personalized view of their own vehicle — what it needs next, what it’s worth, what’s expiring, what’s covered — sent on the vehicle’s schedule rather than the store’s campaign calendar. It goes out whether or not anyone at the dealership thought about that customer this month. Then the Dealership Horizon Loyalty Report tells you what share you’re actually holding, and which way it’s moving.
No store captures the whole wallet. But the distance between 28% and what your store could hold isn’t a staffing problem, and it isn’t a pricing problem. It’s a presence problem — and presence at eight-thousand-customer scale is the one thing software is genuinely good at.
Four of the five costs are yours
Go back to where the customer’s money goes, and set it against how a dealership is organized.
| What your customer pays | Who at your store owns it |
|---|---|
| $2,237 — insurance | Insurance |
| $1,488 — maintenance & repair | Service |
| $765/mo — the payment, and the next one | Sales & F&I |
| The repair they didn’t plan for | Warranty |
| $2,126 — fuel | Nobody |
Four out of five.
That isn’t a clever framing — it’s what a dealership is. A business that sells a vehicle, insures it, services it, protects it, and eventually sells the next one. The five action gears in a VehicleLyfe Horizon Owner Report — Service, Buy, Plan, Warranty, Insurance — were built against that structure. A cost-of-ownership study just happened to itemize the customer’s year in the same order.
Your share of wallet is the distance between those two columns.
How to calculate yours
You can do this on Monday with numbers you already have.
- Count everyone who had at least one repair order with you in the twelve months before last.
- Total the customer-pay service revenue those same people generated over the twelve months since — labor and parts, excluding warranty and internal.
- Divide the second by the first.
- Divide that by $1,488.
The discipline is in step one. Take the list of who your customers were a year ago and hold it fixed — don’t rebuild it from whoever happened to show up. And strip out any account with more than a dozen visits a year first, or your fleet business will tell you a comfortable lie.
The gap is not lost business. It’s business that was never asked for.
The road ahead
Your customer is not going to call and tell you they took the brakes somewhere cheaper. They’ll just absorb the year — about $15,031 of it — and form an opinion about vehicle ownership that you had no part in shaping.
Today they give you roughly two dollars in seven.
Every one of those costs is a conversation. The ones you don’t own, somebody else does.
Loyalty is on the horizon.
Methodology
Vehicle ownership costs are from CNBC’s September 14, 2026 report of an Insurify analysis (using AAA estimates for fuel and maintenance), Experian Q2 2026 loan data, and the Bureau of Labor Statistics.
VehicleLyfe figures are measured from production DMS data across our dealer network, using closed repair orders. Share of wallet is measured over a full twelve-month window, September 2025 through August 2026, against a customer list fixed as of the preceding twelve months. Spend is customer-pay dollars — total repair order less manufacturer-paid warranty. Return-rate comparisons hold visit frequency constant and are reported as differences rather than levels, because the level depends on which of the two questions above you are asking.
Two limitations bias our reported share of wallet upward, meaning the real figure is likely lower. AAA’s $1,488 is per vehicle while ours is per customer, who may own more than one. And defining “your customer” as someone who visited in the prior twelve months excludes longer-cycle owners who are still customers.
One limitation worth stating plainly because it shaped the analysis: roughly a third of repair orders before September 2024 carry no dollar amount at all, against 6% in the most recent year. Treating those blanks as zeroes manufactures a convincing but entirely false trend of rising customer spend. Older data is reliable for visit counts and retention; it is not reliable for dollars. That is why the spend figures here use a single recent twelve-month window rather than a longer series.
