First touch attribution
What is first touch attribution: first touch attribution explained
First touch attribution gives 100 percent of the credit for a sale to the earliest recorded marketing touch in the customer's path, ignoring every touch that comes after it. It answers one question well, which channel started the conversation, and nothing else, because it assigns all the credit by convention rather than by evidence.
You will also see it called first click attribution, first interaction attribution, or first-touch. It is one of the two single-touch models, the other being last touch attribution.
How is first touch attribution calculated?
For each delivered vehicle, find the earliest recorded touch in the customer's path and assign the whole unit to that channel. Then:
- Channel share equals units whose first recorded touch was that channel, divided by total units, times 100.
- First-touch cost per sale equals that channel's spend, divided by the units it was credited with.
No fractions and no splitting. Every unit lands whole in exactly one column, which is why the model is easy to build and easy to over-read.
First touch attribution explained: a worked example
One buyer path: day 1 an organic search reaches the used inventory page, day 5 a click on a third-party marketplace listing, day 12 an opened BDC email, day 19 a click on the branded paid search ad, day 21 delivery. First touch credits the whole car to organic search. The other three channels book zero.
Now the month. A store spends $40,000 and delivers 40 units. First touch credits SEO 16 units, marketplace 14, social 8, paid search 2, and email 0.
SEO is therefore 16 divided by 40, or 40 percent of March. At $10,000 of SEO spend, first-touch cost per sale is $625. Paid search, on $12,000 and two credited units, comes out at $6,000 a car. Social looks strong at $6,000 divided by 8, or $750.
Under last touch the same 40 cars credit paid search with 12 units, which prices it at $1,000. Same month, same spend, and a six-fold difference on one channel driven entirely by which end of the path you decided to reward.
The case for first touch attribution
- It shows what creates demand. Discovery channels rarely close anyone directly. Under last-click reporting they look like they do nothing, and first touch is the only common model that keeps them visible.
- It corrects the branded search illusion. A customer who already decided on your store and searched your name is not a conversion your paid search created. First touch declines to hand that credit over.
- It is stable. The first touch is set on day one and never changes. Later touches cannot rewrite history, so the report does not shift under you mid-month.
- It suits long purchase cycles. When a deal takes three weeks and eight touches, knowing which channel opened the file is a genuinely useful planning input for next quarter's budget.
- It is cheap to run. One stored field per customer record. No modelling, no volume threshold, no vendor.
The case against first touch attribution
- It is not the first touch. It is the first one you logged. A shopper who saw your used stock on a friend's phone, watched an OEM ad, and read a forum thread before ever hitting your tracking gets filed under whatever your system happened to catch first.
- Devices and cookies reset. The same person on a work laptop in week one and a phone in week three often arrives as two people with two first touches. Longer purchase cycles make this worse, and a car purchase is a long cycle.
- It rewards volume, not quality. A cheap channel that pulls in thousands of early browsers will collect a large share of first touches by arithmetic. That is a statement about reach, not about selling cars.
- It makes closing work invisible. The BDC follow-up, the appointment confirmation, the retargeting that pulled a stalled deal back. All zero, forever, in every report.
- It cannot be checked. There is no experiment that shows the first touch deserved all the credit. The 100 percent is a convention chosen before the data arrived, and the model would produce a confident answer even if the first touch had no effect at all.
How first touch differs from last touch
They are mirror images with the same flaw. One says the channel that opened the conversation earned the sale, the other says the channel that was present at the end earned it, and neither claim survives contact with a 21-day path involving four channels.
Used together they are more honest than either alone. Run both, put them side by side, and read the disagreement rather than the numbers. A channel that scores high on first touch and low on last touch is doing discovery work. The reverse pattern is a closing channel. A channel that is low on both is the one worth questioning.
Just do not average the two into a single figure and present it as the truth. Two conventions averaged together is still a convention, and mixing models is a reliable way to lose track of what the report is even claiming. See marketing attribution for the multi-touch and time-decay alternatives, and for why privacy changes have made all of this less reliable than it was five years ago.
