Ask most dealers how Facebook targeting works and you’ll hear a version of the same answer: pick an area, choose “interested in cars,” set an age range and hope for the best.

That’s not how it works anymore, and it hasn’t been for a while. Meta doesn’t need you to tell it who’s in the market for a car. It works it out from what people do. The dealers getting the most from Meta understand those signals, and they understand the one signal most dealers never send.

Meta reads behaviour, not boxes

Every day, people leave a trail of small decisions across Facebook and Instagram. On their own, none of them mean much. Together, they paint a remarkably clear picture of who is about to buy a car.

What they stop scrolling for. Someone who starts watching vehicle reviews, comparison videos and walkarounds, and watching them to the end, is telling Meta something. So is someone who suddenly engages with towing, camping or family travel content when they never did before.

What they browse on Marketplace. Searching and saving cars on Marketplace, comparing prices and messaging private sellers are some of the strongest in-market signals on the platform, even though that buyer may end up buying from a dealer.

The groups they join. Owners clubs, 4WD and touring groups, EV communities and model specific forums all signal intent and interest in a particular kind of vehicle.

What they do on your website. With the Meta pixel installed, Meta can see which vehicles someone viewed, whether they used the finance calculator, how often they came back and whether they started an enquiry.

Meta’s catalogue based Automotive Inventory Ads bring these signals together. They draw on your vehicle catalogue (make, model, year and pricing), the Meta pixel on your website, and audience interests and engagement on Meta to show the right vehicle to the right person. As your inventory changes, the ads update to match, and when a vehicle sells it comes out of rotation immediately.

It isn’t a list. It’s a pattern.

There’s a common misconception that Meta has a list of people “likely to buy a car” that dealers can buy access to. It doesn’t work that way.

What Meta does is pattern matching. It looks at the people who have taken the action you’re optimising for, studies what they did beforehand, and goes looking for more people behaving the same way. Targeting settings are increasingly treated as suggestions, while the system does the real work.

Which raises the most important question in this whole article: what action is Meta actually learning from?

The half of the story most dealers never tell

For most dealerships, the answer is a lead.

Someone fills in a form, Meta records a conversion, and the campaign is told it did its job. Meta then goes looking for more people who fill in forms.

But a form isn’t a sale. The sale happens days or weeks later, in the showroom, recorded in the CRM and the DMS. Meta never hears about it. It never learns which leads bought, which never answered the phone, and which were tyre kickers who fill in forms on every dealer site in the city.

So the algorithm keeps optimising for people who submit enquiries, which is not the same as people who buy cars. Dealers end up with cheap leads, frustrated sales teams, and a cost per lead report that looks good while the cost per sale stays invisible.

Closing the loop

Meta’s Conversions API, usually called CAPI, fixes this. It lets a business send events directly from its own systems to Meta, including events that happen offline.

When CAPI is connected to the dealership CRM, the journey stops at the lead no longer. Meta can be told when a lead became a qualified opportunity, when a test drive happened, and when a vehicle was sold. Matching is done using hashed customer details, such as email and phone number, so the personal information itself isn’t shared in readable form.

The result changes how campaigns behave:

• Meta starts optimising for people who look like actual buyers, rather than people who look like form fillers.

• Dealers can finally see a real cost per sale from Meta, rather than estimating it.

• Lookalike audiences get built from real customers, which makes them far more valuable.

• Recent buyers can be removed from vehicle campaigns and moved into service and accessories messaging, so you stop paying to sell a car to someone who just bought one.

• For service departments and independent workshops, customers can be grouped by when their next service is due, so ads reach them at the moment they need to book, rather than months too early or after they’ve gone elsewhere.

What dealers need to get right

None of this works without the right foundations.

The pixel needs to be installed properly across the site, with the right events firing on vehicle pages, finance tools and enquiry forms. The inventory feed needs to be accurate and update quickly, because Meta can only match buyers to vehicles you’ve told it about.

The CRM needs clean, consistent lead statuses. If sales staff don’t update whether a lead was contacted, tested or sold, there’s nothing useful to send back to Meta.

And privacy needs to be handled properly. Customers should know how their information is used, the privacy policy should reflect it, and data should only be shared in line with Australian privacy requirements.

Tell Meta the whole story

This is a big part of what we do at DRVN Digital. We connect Meta to dealer and workshop CRMs so it knows who your customers are, what they’ve spent with your business and where they are in their ownership journey. For service centres, that means we can segment customers and put ads in front of the people who are due for a service right now.

Meta is already very good at finding people who are thinking about buying a car. Most dealers just aren’t telling it what happens next.

The dealers who connect their CRM and send back real sales data give the algorithm the one thing it can’t work out alone: who actually bought. Everyone else keeps paying for leads and wondering why the sales don’t follow.

So here’s the question for every dealer running Meta ads: is your campaign learning from your enquiries, or from your sales?