Auto.TV / Journal / Channel economics

Third-Party Leads vs. Owning the Channel: The Real Cost Per Sale

A $45 lead that closes at 15% and arrives at three dealerships at once is not a $45 lead. Industry benchmarks put the true cost at roughly $8,000 to $12,000 per vehicle sold. Owned channels run a fraction of that. Here is the arithmetic nobody puts in the renewal deck.

The invoice says $2,000 a month. That number is comfortable. It fits in a budget line, it gets approved without discussion, and it renews on autopilot.

The number that matters is not on the invoice.

Start with the headline figures

Third-party listing platforms are the largest single line item in the average dealership advertising budget, and they have a specific claim on that position: 79% of shoppers visit third-party sites during research, spending 61% of their research time on those platforms. That is a real audience, and pretending otherwise would be dishonest.

The cost of reaching it has moved sharply.

Third-party listing sites command roughly $109,487 annually per dealership, typically at $1,500 to $5,000 per platform per month, and most dealers subscribe to several simultaneously.

And in 2026 the pricing moved again. AutoTrader restructured its packages from four tiers to three, moving features that were previously bundled in the middle tier up into the top tier. The reported effect: roughly 60% of franchise dealers pushed into a higher-cost tier, with the new top tier running approximately 22% to 28% higher per month than the 2025 package it replaced.

That is what pricing power looks like from the receiving end. You do not negotiate it. You absorb it.

Now the arithmetic that is not on the invoice

A lead is not a sale. The gap between them is where third-party economics stop being comfortable.

Sources: Rework, automotive lead providers analysis; Foundry CRO, 2026 automotive marketing benchmarks

Read the bottom row again. Foundry CRO frames the gap plainly: Google Search at roughly $765 per vehicle sold against Cars.com at roughly $12,000 per sale is a 16x difference. Their explanation of why it stays hidden is the important part, and it is worth stating in full because it is the mechanism, not the complaint: third-party providers charge monthly subscriptions rather than per-sale fees. A dealer paying $2,000 a month for ten leads that close at 2.5% is paying $8,000 per sale without ever seeing that number written down.

The subscription model does not hide the cost by accident. A per-sale invoice would end the conversation. A monthly invoice renews.

Why the leads convert worse

Not because the platforms are bad at their job. Because of what the product structurally is.

The lead is shared. The same inquiry goes to several dealerships. The shopper is not choosing you, they are comparison shopping across you. You are in a bidding war that started before anyone picked up a phone. Foundry CRO attributes the 4x to 5x higher close rate on self-generated leads to exactly this: exclusivity.

The lead is duplicated. At a 30% to 40% duplicate rate, a customer submitting on AutoTrader, Cars.com, and your own site within a day generates three billable leads and one actual person. Without CRM de-duplication you are paying three times for one shopper and counting them three times in your reporting.

The lead is price-anchored. A shopper who arrives via a marketplace comparison page has been trained by the page itself to treat you as a row in a table. A shopper who arrives at your address has already selected you.

The dependency problem

Roughly 65% of dealerships get most of their leads from third-party providers. The widely cited rule among people who model this seriously: no single source should exceed 30% of lead volume, and past that you no longer have a lead strategy, you have a dependency.

Dependency has a price that never appears as a line item. It is the reason the 2026 repricing worked. When a platform supplies 70% of your internet leads, the renewal is not a negotiation. You can accept the increase or you can accept the cliff, and the platform knows which one you will choose.

The leverage runs entirely one direction, and it compounds. Every year of dependency makes the next renewal harder to refuse.

The honest case for third-party platforms

They are not a scam, and any argument that treats them as one is not worth reading.

They put your inventory in front of an audience you could not assemble yourself, and that audience is real: 79% of shoppers visit them, spending 61% of research time there. They deliver volume immediately, with no build period. They are genuinely useful for testing a new franchise without major marketing investment, moving aged inventory when you need velocity now, and covering seasonal peaks.

The mistake is not using them. The mistake is having no alternative when the price moves. As one benchmark analysis puts it, the dealerships in the strongest position in 2027 will not be the ones who dropped portals entirely.

A realistic mature mix, per the same analyses: roughly 35% owned, 30% paid digital, 25% third-party, 10% other. Note that this is not zero third-party. It is third-party from a position where you could walk.

What building owned actually takes

Here is where most arguments for owned channels get dishonest, so let us not.

Owned channels cost the most upfront and deliver the lowest cost per sale long term. The upfront part is real. Those $50 to $150 per-sale figures are the output of a machine that took time and money to build, not a switch anyone flipped. Documented transitions run on the order of a year: audit and renegotiate first, build in parallel, shift spend only once owned volume is real, and land somewhere near a 40/60 owned-to-paid split from a starting point of 20/80. Reported savings for an average store land in the $4,000 to $8,000 per month range, with better lead quality attached.

A year. Not a quarter, and not a purchase.

Where an address fits, and where it does not

An owned channel needs somewhere to send people. That is the part this journal has an interest in, so weigh it accordingly.

Direct traffic is the only lead source with no auction, no platform margin, no shared inquiry, and no annual repricing. A person who types your address has already chosen you, which is the same property that makes owned leads close at 25% to 35% instead of 12% to 18%.

But an address does not generate demand by existing. A category domain pointed at an unchanged website is a redirect. It reduces cost per sale only if it sits on top of something worth visiting and is promoted until people remember it. That is a program, not a purchase, and anyone selling it as a purchase is selling you something.

What the address changes is what your spend accumulates into. Every dollar spent driving traffic to a marketplace listing builds the marketplace's position. Every dollar spent teaching people an address you own builds yours. Same dollar, different owner at the end of it.

That is the whole argument, and it is worth exactly as much as your planning horizon. Priced against one year of portal spend, a category domain looks expensive. Priced against ten years of renewals you do not control, at a platform that just moved 60% of franchise dealers into a higher tier, the arithmetic reads differently.

The calculation to run before your next renewal

1. Effective cost per lead, not invoiced. Total portal spend divided by workable leads: real contact info, in your area, responded at least once. Audits routinely find this runs 30% to 60% above the rep's number.
2. Cost per sale by source. Not per lead. Track leads to appointments to shows to units, per provider, monthly.
3. Concentration. What share of volume comes from your largest source? Above 30%, price increases are not negotiable.
4. The ten-year line. Portal spend at current trajectory over a decade, against a one-time asset. Then decide which number is the expensive one.