Cost per sale
What is cost per sale: cost per sale explained
Cost per sale is the marketing spend it takes to deliver one vehicle, calculated by dividing total spend in a period by the units sold from that spend. It is the number that decides whether a campaign paid for itself, and it is the one most dealership marketing reports quietly leave out.
You will also see it as cost per unit sold, cost per acquisition, CPA, marketing cost per vehicle retailed, or advertising per unit. NADA reports the same idea as advertising expenditure per new unit sold.
How is cost per sale calculated?
Cost per sale equals total marketing spend in a period, divided by the number of vehicles delivered from that spend in the same period.
There is a second route to the same answer, and it is the more useful one for diagnosis: cost per sale equals cost per lead divided by close rate. The two formulas always agree, but the second one shows you which half of the problem you are looking at.
Cost per sale explained: a worked example
A single rooftop spends $40,000 on marketing in March and logs 500 leads, so cost per lead is $80. Forty of those leads take delivery, an 8 percent close rate.
Direct division gives $40,000 divided by 40, or $1,000 per delivered vehicle. The second route gives $80 divided by 0.08, which is also $1,000. Same money, same month, and now you can see that the $1,000 is made of two separate things: what the leads cost, and what share of them the store converted.
Halving the media bill takes cost per sale to $500. So does lifting the close rate from 8 percent to 16 percent, without spending a dollar less. Most stores have far more room in the second lever than the first.
What does a normal cost per sale look like?
The closest published reference point is NADA Data 2025, which puts average dealership advertising at $739 per new unit sold across US franchised dealerships in 2025. Total dealership advertising expenditure reached $9.96 billion, an average of $586,246 per dealership for the year.
Read that $739 carefully before you hold anyone to it. It divides all advertising spend by all new units, including the walk-ins, the repeat customers, and the buyers who never touched a tracked campaign. A cost per sale measured against lead-sourced units only, the way most marketing reports do it, will land well above $739 at the same store. The two numbers are not interchangeable, and comparing them is how a perfectly healthy campaign gets killed in a Monday meeting.
NADA also reports where the money went in 2025: search engine marketing took 21.1 percent of dealer advertising, third-party listing sites 20.0 percent, SEO 19.5 percent, and social media advertising 14.2 percent. Roughly three quarters of the budget is now digital, which is exactly the portion where cost per sale is hardest to measure honestly.
How cost per sale differs from cost per lead
Give two sources $9,000 each for the month.
- Source A returns 300 leads at a $30 cost per lead. Six deliver, a 2 percent close rate, so cost per sale is $1,500.
- Source B returns 60 leads at a $150 cost per lead. Twelve deliver, a 20 percent close rate, so cost per sale is $750.
Source A looks five times cheaper and costs twice as much per car. Identical spend, and Source B put twice the units on the ground. A cheap lead that never closes is not cheap. It is spend with a nicer label on it.
This is why the two metrics belong on the same line of the same report. Cost per lead tells you what you bought. Cost per sale tells you what it was worth.
Where cost per sale gets misread
- The delivery lag. March spend delivers into April and May. Dividing this month's cost by this month's units mixes two different cohorts and makes a growing budget look inefficient for as long as it keeps growing.
- Attribution decides the numerator and the denominator. Which channel gets credited for a sale is a modelling choice, so the same store can publish two very different costs per sale for the same channel. See first touch attribution and last touch attribution for how far apart those answers can sit.
- Blending new and used. Different spend, different close rates, different gross. One combined figure hides which side is carrying the other.
- Treating it as a cost, not a ratio. A $1,400 cost per sale on units averaging $4,200 in total gross is a good month. The same $1,400 against $2,000 gross is not. The figure only means something next to what the car made.
- Small denominators. A source that delivered three cars has a cost per sale that swings by hundreds of dollars on one deal. That is noise, and stores cut good channels over it every quarter.
Using it to make a decision
The practical test is not whether cost per sale went down. It is whether the store could scale the source that produced it. A channel at $600 per sale that caps out at four units a month matters less than a channel at $1,100 per sale that will take another $10,000 and keep converting.
So run the number monthly, keep it next to average gross, and check it against the close rate before you touch the budget. Half the time the fix is not in the media at all: the leads were fine and the follow-up was slow.
