Which ads delivered units?

A shopper visits three listing sites, clicks a Meta ad, searches your dealership by name, then calls. Every one of those platforms reports the lead. Add up what your vendors claim and you have sold each car two or three times over.

The only arbiter is your DMS: which of those shoppers actually took delivery, and what the gross was. Matching that against the lead exports gives one number per channel that sums to the units you really sold, which is the report that makes a vendor conversation short.

A large share of dealership sales still start with a phone call or a walk-in, and those are exactly the buyers that click attribution cannot see. Treating them as untracked and ignoring them means ignoring most of the store.

Matching on the customer's phone number or email catches the phone-up who saw an ad three weeks earlier. What genuinely cannot be traced stays in Direct / Unknown rather than being spread across the paid channels, so the paid numbers keep meaning something.

A deal is written on Saturday, the customer drives away on a spot delivery, and the contract is funded — or it is not. Some of those cars come back. The sold log and the delivered log disagree for a week or more every month, and the controller closes on the second one. If the file you export is the sold log, you are crediting campaigns with revenue for units that never legally left the lot. Export what accounting recognizes, at the gross accounting recognizes, and the channel report will reconcile with the statement instead of arguing with it.

The date matters as much as the amount. A unit written on the thirty-first and delivered on the second belongs to different months depending on which column you send, and in a store that pushes hard at month end that is not a rounding difference. Pick one convention, keep it in every import, and say which one it is. The value of comparing this August against last August rests entirely on both files having been built the same way, and nothing in the report can detect a convention that changed halfway through the history.

Unwinds and buy-backs should come out too. A cancelled deal left in the file quietly inflates whichever channel produced that buyer, and it inflates it by a full unit of gross rather than by a click. Most dealer management systems can filter to funded, delivered contracts before anything leaves the building; if yours cannot, take them out before upload, because the report counts every row in the sales file as a sale and has no status filter. Either way the point is the same: the number has to survive the controller reading it over your shoulder.

No dealership lives on front-end gross alone, and everybody in the building knows it. The customer who took delivery in March comes back for the first service, then for tires, then for brakes, and eventually trades back in. Marketing is judged on the delivery and on nothing after it, which means a source that produces owners who stay with the store looks identical on the report to one that produces buyers who are never seen again. Those two sources are worth very different amounts, and the gap compounds every year the customer stays.

Because the match is on the person rather than a session, repair orders raised years later still credit the channel that produced the original buyer. Export closed repair orders alongside delivered units, carrying the same customer phone or email, and the ranking usually changes shape. A source that looked mid-table on front-end gross moves up once the fixed operations revenue behind those customers sits next to it. That is not an arithmetic trick; it is the store's real economics finally appearing in a marketing report rather than only in the financial statement.

It also gives the fixed operations side its own answer. Service marketing — reminders, declined-work follow-up, the local search spend aimed at people who bought elsewhere — can be ranked on repair-order revenue in the same report as vehicle campaigns, rather than being defended separately with appointment counts nobody trusts. A department column mapped as Product / service is what makes that cut possible, and it is the single breakdown most worth adding to a dealership export. Without it the two economies are averaged into one figure that describes neither of them.

Acquisition has become a budget line of its own. Stores buy search terms, run appraisal offers and pay for inquiries from people who want to sell a vehicle rather than buy one, and that spend is defended with appraisal counts. An appraisal is not an acquisition, and an acquisition is not a retailed unit. The chain from advertisement to money runs through three separate steps, and most stores measure only the first because it is the only one an ad platform can see without help from the dealership.

There is a second reason to add the appraisal lists to the lead file, each row tagged with its source. A meaningful share of the people who sell you a vehicle also buy one, either the same day or the next time they need something, and they appear in the sales file under the same phone number. Ranked on delivered gross like any other source, an appraisal campaign stops being a cost center defended by volume and becomes a channel with revenue attached. For many stores that would be the first evidence anybody has seen in either direction.

The wider point is that any list of people with contact details can be uploaded and ranked beside paid media. Appraisal inquiries, service customers sitting in a strong equity position, a lapsed-owner list pulled out of the dealer management system, the campaign a vendor ran on your behalf and reported on itself. None of them are advertising channels in the platform's sense of the word, and all of them produce delivered units. Putting them on one axis is the only way to find out which of the store's several demand engines is carrying the month.