Car Dealer SEO Budget: Why National Campaigns Waste It [Study]

I ran 69 real used car buying questions through ChatGPT and Perplexity and logged every source they cited. On national questions, independent dealers took 3% of citations. AutoTrader alone took 15.7% of every citation in the entire study. That gap is not a gap you close with a bigger retainer. It is the structure of the market, and any campaign priced against closing it is a campaign that spends your budget on a result nobody can deliver.

This article is about where that money goes instead. It is the practical follow up to my study of 967 AI citations in UK used car search, and every figure in it comes from that dataset or from named third party research.

01The Number That Ends the Meeting

Three percent, and what it costs to chase it.

Here is what my data shows for questions with no location in them. Questions like “where to buy a used car online UK” or “where to buy a used Ford Fiesta UK”, which are exactly the terms a national campaign targets.

AutoTrader is not a competitor you outrank. Ahrefs data puts AutoTrader’s monthly search traffic at around 12 million visits. It holds more Ford Fiestas at any given moment than your entire forecourt will sell this decade. When a buyer asks a machine where to buy a used Fiesta in Britain, the honest answer is AutoTrader, and the machine gives the honest answer.

Your agency is not incompetent for failing to change that. Nobody changes that.

A campaign that targets national terms is not a campaign with a low success rate. It is a campaign with no mechanism.

02Where the Same Question Becomes Winnable

Add a town and the market you are competing in changes entirely.

I asked the same kind of question at six levels of specificity, from “where to buy a used car online UK” through to “cheap used car dealers in Kidderminster”. The results separate cleanly.

At town level, independent dealers took 26.0% of all citations. AutoTrader, Motors, CarGurus and Cazoo together took 25.7%. The small dealers, collectively, outcited every marketplace in the country.

Sixty seven named independent dealerships appeared across the study. Bells Garage. Dudley Motor Company. Clarks of Kidderminster. Coventry Sterling Motors. Walsall Car Centre. None of them outrank AutoTrader on anything, and all of them get named in AI answers when the question includes their town.

03The Objection: Buyers Travel Now, So Why Go Local?

The counter argument deserves a straight answer, because the evidence genuinely cuts both ways.

Any agency selling you a national campaign will raise this, and they are not making it up. AutoTrader research reported that the average distance a buyer would travel rose to around 50 miles, with 46% willing to buy from a retailer over 50 miles away. Buyers demonstrably travel.

Now the other side. JudgeService, working from 53,585 customer surveys, found that 44% of buyers travelled up to five miles to visit a franchised dealership and a further 28% travelled between five and fifteen miles, while the share prepared to travel over 30 miles fell from 21% to 16.5%.

Carwow’s analysis reconciles the two. Writing in AM-online, its head of trading described the pattern as buyers travelling further than expected when presented with transparent options, while noting that all other things being equal, a consumer picks their closest dealer.

That sentence is your entire strategy. Buyers travel when something specific pulls them: a rare spec, a price gap, a reputation. They do not travel for a generic used hatchback that three dealers within ten miles also hold. Most of your market lives inside fifteen miles, and the ones who travel come for a reason you can name, not for a top ten ranking on “used cars UK”.

04Where the Budget Goes Instead

Four allocations, each tied to a number from the study.

What follows is not a pricing model. Budgets vary too much for that to mean anything. It is an allocation logic, and the proportions come from where the evidence says the return sits.

Every figure traces to the study or to named third party research. Nothing here is an industry benchmark I invented.

Why location pages fail when they fail

Most dealers have location pages already, and most of them do nothing. The reason sits in that 73.3% figure.

A page that swaps one town name for another across fifteen near identical templates gets retrieved and discarded, because it contains no information a buyer in that town cannot get elsewhere. The page exists. It does not answer.

A location page earns its citation when it carries things only you know. Which models move fastest off your forecourt in that area. What the part exchange market looks like locally. Which roads to come in on. Who the buyer speaks to when they arrive. That material takes an afternoon per town to gather and it is the difference between a page that ranks nowhere and a page a machine reads as the local answer.

This is the same discipline I set out in my Pain Point SEO strategy, applied to geography instead of objections. Find the thing only you can say, then say it plainly.

05Five Questions That Test Any Agency Pitch

Ask these in the meeting. The answers sort the proposals quickly.

1. Which towns does this campaign target, by name? A proposal that answers with keyword categories rather than place names is a national campaign wearing local clothes.
2. Show me a page you would build for one of those towns. If the sample reads like a template with the town name dropped in, you are buying the 73.3% discard pile.
3. What happens to my visibility in ChatGPT and Perplexity? Not a trick question. The honest answer describes measurement and uncertainty. A confident guarantee is a warning.
4. Which national terms are we deliberately not chasing, and why? An agency that names what it refuses to do has a strategy. An agency that targets everything has a keyword list.
5. How will we know in ninety days whether this works? The answer needs a measurable thing that is not a ranking screenshot.

06What to Measure Instead of Rankings

Four things, none of which need a rank tracker.

Named appearances in AI answers for your towns. Ask ChatGPT and Perplexity who the best used car dealers are in each town you serve. Record who gets named. Repeat monthly. This takes twenty minutes and it is the closest thing to a direct measure of the opportunity in this study.

Enquiries that mention a town page. If somebody arrives having read your Kidderminster page, that is the loop closing. Ask on the enquiry form.

Share of voice against local competitors, not national ones. The relevant question is whether you appear when the buyer names your town, not whether you appear when they name the country.

Which pages get retrieved and ignored. Harder to see directly, but a page attracting impressions and no clicks or citations is the same signal as a page a machine reads and rejects. Rewrite it rather than adding a sibling.

07Frequently Asked Questions