Attribution models
The five models Clakta offers, how each one shares the credit for a sale, and which one to use for each decision.
The attribution model decides how the credit for a sale is shared between the ads that came before it. Clakta has five. Change yours from the dashboard toolbar or any Ads Manager page, and your existing data is shown again under the new model right away. To see all five applied to the same purchase, read How attribution works.
Last Click
All of the sale goes to the last paid click before the order.
The default. Only paid clicks count; visits without an ad are ignored. If there is no paid click inside the window, the order is unattributed.
Good for: deciding about campaigns that close sales, and whole conversions that are easy to explain.
Watch out: whatever created the interest gets no credit, so prospecting looks worthless.
First Click
All of the sale goes to the first paid click in the journey.
The mirror of Last Click: same rules, opposite bias.
Good for: seeing which campaigns bring in new customers.
Watch out: retargeting and campaigns that close sales look like they do nothing.
Linear
The sale is split equally between every touchpoint, including visits without an ad.
Each paid click and visit inside the window gets 1 / n of the sale.
Good for: seeing the whole journey, including how much revenue comes through channels you don't pay for.
Watch out: a visit without an ad counts as much as a paid click, which lowers the paid channels' results. Don't judge ROAS with it.
Linear Paid
The sale is split equally, but only between paid clicks.
Good for: comparing paid channels fairly when customers usually see several ads before buying. It is usually the best choice for advertisers on several channels who find Last Click too harsh.
Watch out: every paid click counts the same, whether the person was just looking or ready to buy.
Triple Attribution
Every paid channel that touched the order gets full credit.
If Meta and Google both touched a sale, both get a full conversion, so the totals across channels go above your real number of orders on purpose. A channel that touched the order twice still counts once.
Good for: comparing Clakta with what each ad platform shows in its own reports, because that is how each of them counts the same sale.
Never add up Triple Attribution conversions across channels and treat the result as revenue: it counts the same order more than once. Use it to compare with the platforms, not to get a total.
Which one to use
| Your question | Model |
|---|---|
| What closed the sale? | Last Click |
| What brought the customer in? | First Click |
| How do my paid channels compare? | Linear Paid |
| How much comes through channels I don't pay for? | Linear |
| Why does Meta show more sales than Clakta? | Triple Attribution |
Pick one model as the number your team uses for budget decisions, and use the others to investigate. Switching models in the middle of a discussion is a quick way to lose everyone's trust in the data.
What the model doesn't change
- Total revenue and number of orders: they come from your orders.
- Ad spend: it comes from the platforms.
- The conversions the platforms report: Clakta never changes them.
Only Clakta's attributed conversions, attributed revenue and what is calculated from them (ROAS, CPA, profit) change with the model.
Related articles
How attribution works
How Clakta decides which ads led to a sale, what each of the five models does with the same purchase, and why Clakta's numbers differ from the ad platforms'.
Attribution windows
How long an ad click can still get credit for a sale, how to pick your window, and what happens when customers take longer than that to buy.
Unattributed orders
Why some orders get no ad credit, how to tell a real organic sale from a tracking problem, and what to fix in each case.