US dealers spent $9.96 billion on advertising last year. That works out to $739 for every vehicle sold, up $34 from the year before, and it beat the previous record of $9.82 billion set back in 2016. Digital took 74.9% of it; traditional media held 25.1%, with television at just over a billion dollars and radio at $687 million.
I have no argument with any of that. Dealers advertise because it works. What I want to do is put a second number next to it.
The comparison nobody runs
Take the $739. That is the blended cost of producing one sale from all sources — the conquest shopper who found you on a third-party site, the person who saw your Facebook ad, and the customer you sold in 2021 who came back on their own and cost you approximately nothing.
That last group is in the average, quietly making it look better than the conquest number really is. Strip them out and the true cost of a net-new buyer is meaningfully higher than $739. Every dealer knows this instinctively. Very few have the number, because sales attribution and service data usually live in systems that don't speak to each other.
Now price the other side. What does it cost to send a segmented, personalized message to a customer already in your DMS — someone whose vehicle, purchase date, service history and equity position you already own? Direct mail is the expensive end of that. Email and SMS are rounding errors by comparison. Even at the expensive end you are not in the same postcode as $739.
The point isn't that advertising is wasteful. It's that most stores are running a sophisticated, expensive, well-measured operation on the acquisition side and something close to nothing on the retention side — and then describing the result as a marketing budget.
Why the retention side stays empty
Three reasons, in the order I hear them:
"Our data's a mess." It usually is. Duplicate records, dead emails, phone numbers that belonged to somebody in 2019. That is an argument for cleaning it, not for ignoring it. Segmenting the addressable sales, service and unsold-prospect database is the first thing Bumper Retention does, because you cannot micro-target a list you don't trust.
"The OEM already emails them." The OEM emails everybody the same thing on the same day. That is not your retention programme. It's weather.
"We tried it and it didn't work." Nearly always this means a single blast to an unsegmented list through one channel. If a customer has ignored email for two years and answers texts, sending them email is not a test of retention marketing. It's a test of email.
The part that actually moves the number
Channel discipline is where most of the lift lives, and it's the least glamorous thing to talk about. Our platform tracks which medium each customer actually engages with and concentrates future messaging there. Not because it's clever, but because the alternative — everyone gets everything — trains people to ignore you across every channel at once.
The second piece is timing. Reaching a customer in the two months before they're in-market is worth more than reaching them ten times in the eighteen months before that. Sorting and analyzing existing customer data to surface those signals is what we mean by intent mining, and it's the difference between a database and a mailing list.
A number to bring to your next meeting
Divide your total ad spend by units retailed. You'll get something near $739 if you're an average store.
Then ask what percentage of that budget went to people who had already bought from you. If the answer is under 10%, you are spending like the market is growing.
It isn't.
Figures: Borrell Associates dealer ad spend data as reported by Dealership Guy.
