Why Google Vehicle Listing Ads Underperform for Car Dealerships
15+ years in performance marketing, including in-house time inside a dealership group. Connect on LinkedIn.
⚡ Quick Answer:
Google Vehicle Listing Ads underperform because Performance Max routes most of your budget to Display, YouTube, and Gmail instead of true inventory search placements. You end up with cheap clicks from low-intent audiences. Fix it by auditing your placement reports, separating VLA from brand campaigns, and tightening your feed so Google can match the right car to the right buyer.
What Are Vehicle Listing Ads?
Vehicle Listing Ads (VLAs) are Google’s inventory-based ad format for car dealerships that shows a specific vehicle’s photo, price, mileage, and location directly in search results. When a buyer searches for a car, Google matches the query against your Merchant Center feed and serves the exact vehicles that fit, linking each ad straight to that car’s Vehicle Description Page (VDP) on your website. A VLA is the ad format. Performance Max is the campaign type that has delivered it since 2023, alongside Display, YouTube, Gmail, and Discover placements, and that distinction drives most of what this post covers. Dealers pay per click, and results depend heavily on feed quality, campaign structure, and how much of the budget actually lands on true vehicle listing placements instead of Performance Max’s broader network.
The format was built on a simple promise: show your exact inventory directly to a buyer searching on Google, and land them on the VDP for that specific car, not a marketplace listing you’re paying a subscription fee for.
The promise is real. The execution, for most dealerships, is not.
The complaint shows up the same way almost every time: high click volume, a cost-per-lead that looks fine on a slide, and a sales floor that swears it never saw any of it. What separates the dealerships getting real sold-unit results from VLAs from the ones generating expensive activity comes down to a handful of root causes, and every one of them is fixable.
The Core Problem: You’re Not Buying What You Think You’re Buying
Since 2023, VLAs only run inside Performance Max. That gets you Google’s full machine learning stack, but it also means your “inventory ads” campaign no longer serves only clean, high-intent vehicle listing placements.
In a lot of accounts, the majority of spend and clicks land on Display, YouTube, Gmail, and Discover instead of the precise inventory unit a buyer searched for. Those placements are often cheaper per click and can pad a conversions column. They rarely match the intent of someone who saw your exact car in a search result and clicked straight through to its VDP.
Based on 6 dealership Google Ads accounts we managed directly, over a 60-day period.
The dashboard looks active. The cost per sold unit, the number that should be driving every budget decision, tells a worse story. We’ve already published the ranges for this:
| Campaign type | Typical cost per contacted lead |
|---|---|
| Performance Max (VLA delivery) | $55–$130 |
| Standard Search | $38–$85 |
That spread is Performance Max’s broader, lower-intent placements diluting what used to be a tightly scoped channel.
Most of Your Lot Never Gets Shown at All
Here’s the part that doesn’t show up in any dashboard summary: pull impression data at the individual VIN level, and most dealers find that a small slice of their inventory absorbs nearly all the impressions while the rest sits at single digits or flat zero.
The reason is structural, not a mistake anyone made. Performance Max allocates impressions toward whatever it predicts will convert, and it predicts based on signals like search volume, price competitiveness, and how similar vehicles have performed in the past. A clean 2022 Camry at a sharp price has all three. A 2019 minivan with 70,000 miles, a less common trim, or a price that’s a notch above market doesn’t. The algorithm isn’t being lazy. It’s doing exactly what it’s built to do, which is chase the highest probability of a click, and that means systematically starving anything that doesn’t look like an easy sell.
That’s a problem, because the vehicles least likely to get algorithmic attention are usually the ones that most need it. A car sitting on a floor plan for 75 days is costing you interest every single day it doesn’t sell, and that cost doesn’t slow down because Google decided the car wasn’t worth showing. Aging units lose value the longer they sit, your aggregator and VLA spend keeps getting allocated to the vehicles that would probably have sold anyway, and the units actually dragging down your turn rate stay invisible to the one channel built to put them in front of a buyer.
Check this yourself before assuming your VLA spend is doing its job. Pull a VIN-level impression report (Merchant Center or your ads platform will show it) and sort by impressions, ascending. If a meaningful share of your active inventory shows zero or near-zero impressions over the past 30 days, that’s not a small optimization gap. That’s most of your floor plan running ads that never reach a single buyer.
The Setup Mistakes That Make It Worse
Even when the format itself is sound, a handful of execution problems do most of the damage.
No segmentation. One or two campaigns covering the entire lot is the fastest way to waste a budget. A $12,000 trade-in and a $45,000 low-mileage truck have nothing in common: different buyers, different search behavior, different margin. Lump them together and Google’s algorithm spends disproportionately on whichever model gets the most volume, while the inventory that actually needs exposure sits starved of budget.
A weak feed. VLA performance is a direct function of feed quality. Missing attributes, stale pricing, absent photos, and bad categorization cause disapprovals and weak matching, and most accounts never fully recover from a Merchant Center feed that was wrong from day one.
Targeting set too wide. Broad radii, or “Presence or Interest” instead of “Presence only,” pull in traffic from well outside your primary market area. It’s cheap traffic. It rarely turns into a visit.
No real audience signal. Without first-party data feeding the campaign, Google’s model defaults to optimizing for volume rather than the buyer profile most likely to actually purchase from your store.
Reporting on the wrong number. Low CPC and high click counts feel like a win. They aren’t, if none of it traces back to a sold unit. This is the same trap we cover in evaluating whether a vendor is actually performing: activity metrics and sales outcomes are not the same report.
Third Parties Are Often Running Ads on Your Own Inventory
Autotrader, CarGurus, and Cars.com can run their own VLAs using your inventory data, sending the click to their marketplace page instead of your site, and most dealers never notice until someone points it out. You can sometimes constrain this through feed settings or account structure, but it’s a structural reality of the format, not a bug you can fully turn off.
It also costs more than it looks like on paper. A dealership’s own site converts at roughly four times the rate of an aggregator or endemic listing page, based on the same 6 dealership accounts we managed directly over a 60-day period. Every click a third party redirects away from your VDP is a quieter, costlier version of the same sale.
Meta Faces the Same Problems, for the Same Reasons
Dynamic inventory ads on Meta run into nearly identical issues: high lead costs, lower conversion quality than the dashboard suggests, and the same root causes (feed accuracy, audience quality, optimizing for the wrong outcome) showing up on a different platform. If your Meta and TikTok numbers look soft, check the feed and the conversion event setup before blaming the channel.
What Actually Makes VLAs Work
The dealerships getting real results treat VLAs as a precision instrument, not a campaign you turn on once and leave alone.
Segment aggressively. Separate campaigns or asset groups by new versus used, price band, body style, or days in inventory. That’s how you control budget allocation and bidding by actual demand and margin, instead of letting volume decide for you.
Fix the feed before anything else. Complete, accurate, frequently refreshed data is the foundation everything else sits on. Use the Content API or a scheduled feed push to keep price and availability current, and audit for disapprovals on a regular cadence, not once a quarter.
Tighten geography and audience signals. Start with a realistic radius around the rooftop and feed the campaign real CRM and website data. Widen the net only for specific aging units that need broader exposure to move.
Track outcomes that matter. Move past form-fill counts. Track calls, appointment requests, and tie performance back to sold units wherever your data allows it. Value-based bidding behaves very differently once it has real transaction value to optimize toward.
Push for cleaner VLA delivery. Work with whoever manages the account on exclusions and structure that increase the share of spend landing on true vehicle listing placements instead of broad Performance Max distribution. It’s the difference between renting Google’s whole ecosystem and renting the part of it that actually sells cars.
Report on cost per sold unit, not clicks. This one change, more than any setting inside the campaign, usually reveals which channels are genuinely contributing and which are just generating noise.
When the Platform Isn’t the Problem
VLAs can be one of the most effective channels a dealership runs, because they put a specific car in front of a buyer who is actively looking for it. Performance Max made the format more capable and more complicated at the same time, and results now hinge on structure, data quality, and ongoing management more than they did when the format was its own standalone campaign type.
If your VLA or Meta inventory campaigns are generating activity without generating sold cars, the platform is rarely the actual issue. It’s almost always the setup: segmentation, feed quality, targeting, or measurement pointed at the wrong number. A digital audit is how to find out which one, with a real cost-per-sale baseline instead of a guess.
If you’d rather have the account structure itself fix the placement dilution and coverage gap covered in this post, that’s exactly what Enhanced VLA is built to do: confine delivery to true Search and VLA placements and guarantee every VIN clears a minimum impression floor. For the standard Search RSA side of the account, AutoAutoAds covers a related but different problem: keeping text ad headlines current with your lot, with full control and no third party touching your data.
Frequently Asked Questions
What’s the difference between VLAs and Performance Max?
A VLA is the ad format: a vehicle-specific listing with photo, price, and mileage pulled from your feed. Performance Max is the campaign type that now delivers it, alongside Display, YouTube, Gmail, and Discover placements. Since 2023 you can’t run VLAs outside of Performance Max, which is why the format performs differently than it did as a standalone campaign type.
Should I turn off Performance Max and run pure VLAs instead?
You can’t fully separate them anymore, but you can structure the account, asset groups, and exclusions to push more of the spend toward true vehicle listing placements. That’s a setup and management question, not a simple toggle.
How do I know if my feed is the actual problem?
Check Merchant Center for disapprovals and warnings first. Then verify that price, mileage, and photos in the feed match what’s live on your VDPs right now. A feed that’s even a day or two stale on pricing is enough to tank match quality and trigger disapprovals.
Do third-party sites really run ads on my own inventory?
Yes. Aggregators that receive your feed can serve their own VLAs using your vehicle data, sending the click to their marketplace page instead of your website. It’s a structural feature of how vehicle feeds get distributed, not something unique to one platform or one dealer.