Your Rankings Held. Your Traffic Didn't. Here's What Changed.
A dealer called us last quarter convinced his SEO vendor had quietly stopped working. His rankings report looked the same as it had all year — same keywords, same positions, a few even improved. His organic traffic was down by roughly a third.
Nobody had stopped working. The results page had changed underneath him.
What actually happened
Google AI Overviews now appear on roughly 47% of searches. When one appears, the answer sits above every organic result, and studies have measured organic click-through falling between 34% and 65% on those queries.
Read that carefully, because the mechanism matters. Your position did not move. The page above your position got bigger. You still rank third; third is now further down a longer page, underneath an answer that may have already satisfied the person searching.
This breaks the assumption most dealership SEO reporting is built on — that rank is a reliable proxy for traffic. For twenty years it mostly was. It isn't anymore, and a report showing only positions will show you a healthy business right up until the month the numbers stop making sense.
Which searches lose the most
The impact isn't evenly distributed. AI Overviews trigger most on informational queries — the questions people ask while researching, not while buying:
- "How much is a [model] lease payment" and similar payment questions
- "Best midsize SUV for families" and comparison research
- "What does certified pre-owned mean" and terminology explainers
- "When should I replace my brakes" and service questions your fixed ops content targets
Transactional and local searches hold up better. "Toyota dealer near me" still produces a map pack and a click, because the searcher needs a destination, not an explanation.
The practical read: if your content strategy leaned on informational blog posts to pull top-of-funnel traffic, that's the part of your program under pressure. If you invested in local presence and inventory pages, you're more insulated.
What to do about it
Not panic, and not abandon SEO — the shoppers still exist and still buy cars. But the measurement has to change before the strategy can.
Stop reading rankings alone. Rankings and organic traffic have to be read together. Flat rankings with falling traffic is a specific, diagnosable pattern, and it means something entirely different from falling rankings with falling traffic. One is the SERP changing around you. The other is your site losing ground.
Get your technical house in order. Being cited in an AI answer depends on the same fundamentals as ranking: clean structured data, sensible headings, crawlable pages, accurate business information. Sites that are hard for a crawler to parse are hard for an answer engine to quote.
Take local seriously. Google Business Profile is one of the surfaces AI answers pull from for local intent, and it's free. In our own cohort analysis of one dealership's traffic, visitors arriving from Google Business Profile over-indexed at 2.8x among the most engaged shoppers on the site.
Start with an honest baseline
You can't tell whether you're losing clicks to AI Overviews or to your own site's problems unless you're looking at both at once. Our free dealership SEO audit scores organic traffic, keyword rankings, backlink profile, and technical SEO side by side, then ranks your site against other dealership websites we monitor — so a soft number tells you whether it's soft for you specifically or soft across the market.
If your rankings look fine and your traffic doesn't, that gap is the story. Better to find it in an audit than in a year-over-year report next January.
A note on the numbers
The 47% figure and the 34–65% click-through range come from third-party industry research measuring AI Overview prevalence and impact; methodologies differ between studies and the numbers are moving quickly as Google iterates. The Google Business Profile figure is first-party, drawn from a single dealership's analytics, and reflects correlation rather than proven causation. Use them to know what to look for in your own data, not as benchmarks to hit.
