Auto.TV / Journal / Buyer behavior

What Car Shoppers Actually Do Before They Contact You

Roughly 14 hours of research. 4.2 websites. 61% of that time on somebody else's platform. And two thirds of shoppers now contact the dealership before they ever walk in. The journey is almost over by the time you enter it.

Most dealer marketing is designed around a moment: the lead form, the phone call, the walk-in. That moment is treated as the beginning of a relationship.

By then most of the deciding is already done. The lead form is near the end of the process, not the start of it, and the choices that mattered got made somewhere you were not present for.

The shape of the journey

Cox Automotive has tracked this for sixteen years. The 2025 study, released January 2026, surveyed roughly 2,300 buyers who purchased or leased in the previous twelve months, weighted to be representative of the buyer population.

Fourteen hours. For comparison, the entire new-vehicle buying process runs around thirteen and a half hours end to end, with the increase in recent years driven by research and discovery rather than paperwork. Cox found the extra time did not reduce satisfaction, because it went into shopping, test driving, and learning about features rather than into the transaction itself.

People enjoy the research. They do not enjoy the purchase. That distinction should shape where marketing money goes, and mostly it does not.

Where those hours are spent

Not with you.

Sources: Ritner Digital, B2B Autotrader, Camp360 on the Cox study, Invoca

Sixty-one percent of the most valuable attention in your category, spent on platforms you rent access to at roughly $109,487 per year, alongside your competitors, on a page designed to make you comparable.

That last clause is the one worth sitting with. A marketplace listing page is not neutral ground. Its function is comparison. It exists to make the row above yours as easy to click as yours.

The undecided majority

This is the most commercially important sentence in the study and it gets read as a throwaway.

Two thirds of the market has not decided. That means most of the demand in your category is not being captured, it is being created, during those fourteen hours, on platforms where your role is a row in a table.

Last-click search does not create that demand. It bills for it afterward. Which is precisely why last-click search looks so efficient in reporting and why, per the attribution research, budget keeps flowing to the most trackable channels rather than the most influential ones.

What they are watching

The research hours are increasingly video hours, and the direction is unambiguous.

Google reports that watch time for test drive videos on YouTube grew by more than 65% over two years. That is not the general rise in video watching. It is something narrower: people watching a car get driven before they drive it themselves.

The 2025 Cox study adds a new layer. In its first year tracking AI usage, it found 19% of all buyers and 25% of new-vehicle buyers used AI websites or AI-generated overviews during shopping, and those shoppers reported the highest satisfaction levels. 84% of shoppers who leaned into AI-powered online tools reported high satisfaction.

An answer engine now sits between a shopper's question and your inventory. It cites sources. Whether it cites yours is a function of whether you published something worth citing at an address worth naming.

Fourteen hours of research, and the destination is somebody else's brand. Shoppers are going to research regardless. The open question is whose address they have in their head when they do it.

The used-vehicle shopper is a different animal

Worth separating, because the averages hide it. Cox research with NIADA, drawn from nearly 400 used-car buyers surveyed between September 2024 and August 2025, profiled buyers of vehicles aged 11 to 20 years.

The average independent car shopper spends 97 days shopping and conducts more than 16.5 hours of research. The stated reason is blunt: they have a need before they have the means to pay for it. 77% earn less than $75,000, 43% have fair or poor credit, and 56% pay cash.

Ninety-seven days. More than three months of intermittent research, at higher hours than the average buyer, from a shopper with less money and more constraints.

And the loyalty finding is bracing: 60% were satisfied with the shopping experience, but only 21% would return to the dealership. As Cox's Stegall put it, when the car is the star, the dealership is irrelevant as long as it seems fair.

If your brand is not the destination, the vehicle is, and vehicles are fungible.

What the behavior implies for spend

Three things follow, and only the third is about domains.

Speed at the moment of contact is the highest-leverage lever you fully control. Responding within 15 minutes yields 50% more closed deals, and 40% of leads close after day three, so follow-up through day seven matters. Phone shoppers spend 28% more, convert 30% faster, and retain 28% better than digital-only leads, and phone is described as the most underinvested channel relative to its conversion rate.

The gap between what shoppers want online and what they get is a free conversion. Cox found 48% want to apply for credit online but only 33% do; 40% want to select F&I products online but only 16% do; 37% want to finalise price online but only 19% do. 63% say an omnichannel approach would be ideal, against 28% wanting all-online and just 7% actually buying completely online.

The research phase is the only phase where a brand can be built, and it is the phase currently outsourced. Sixty-one percent of research time on rented platforms means the category's formative hours are spent teaching shoppers somebody else's brand.

Where the address fits, honestly

This journal is published by the owner of Auto.TV, who is selling it, so discount accordingly.

The case: two thirds of shoppers are undecided, they spend fourteen hours researching, and 61% of that happens on a platform whose business model is making you comparable. A category address is the one thing on that list a marketplace cannot rent you, because it is the one asset that does not sit in a comparison row.

The counter, which is real: none of this data says shoppers type category addresses. It says they search, they use marketplaces, and increasingly they ask AI. Nobody in that 2,300-person sample typed a category domain, because the destination they would have typed it for does not exist yet. Buying the address does not create the destination. It gives the destination a name that a researcher could plausibly guess, which is a real advantage and a smaller one than a sales page would claim.

What is not arguable is the direction of the money. Fourteen hours of attention, 61% of it rented, at a cost that rose 41% in a year, spent teaching shoppers a brand that is not yours. Whatever the answer is, it involves owning more of the research phase than you currently do.