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.
A click can only get credit for a sale that happened after it and within the attribution window. Anything older is left out.
Pick your window
In the Dashboard and in the Ads Manager, next to the attribution model, there is a button with a sliders icon. Click it and choose how long a click can still get credit:
| Window | Clicks that count |
|---|---|
| 1 day | From the 24 hours before the sale |
| 7 days | From the week before the sale |
| 14 days | From the two weeks before the sale |
| 28 days (default) | From the four weeks before the sale |
| Lifetime | Every click in the customer's journey, however old |
Lifetime goes back as far as your project keeps data.
The window applies to every attribution model, Last Click included. Visits without an ad follow the same window under Linear, the only model that counts them.
Your choice only changes what you see. Each teammate picks their own window, and the Assistant answers with yours. Workflows and requests made with an API key use 28 days.
Changing the window recalculates the numbers right away, past periods included.
How the window works
For each order, Clakta looks at the time between each touchpoint and the order. A touchpoint after the order, or older than your window, doesn't count. The model then shares the credit between the touchpoints left. If none are left, the order is unattributed.
With a 7-day window, a customer who clicked 10 days ago and again 2 days ago keeps only the second click. Under Last Click that changes nothing. Under First Click it changes a lot: the "first" click becomes the one from 2 days ago, because the older one no longer counts. With a 14-day window, both clicks count.
Which window to pick
Pick a window about as long as your customers usually take to buy.
A short window (1 or 7 days) only credits clicks close to the sale. It fits products people buy on the spot. Ads that brought the customer in weeks earlier get nothing.
A long window (28 days or Lifetime) also credits the click that first brought the customer in. It fits expensive products and decisions that take weeks, and it gives prospecting campaigns their share of the sales.
When customers take longer than your window
If your customers usually think for two weeks and your window is 7 days, the click that brought them in has expired by the time they buy. The sale then goes to whatever retargeting closed it, or to nothing.
Signs this is happening to you:
- First Click and Last Click show almost the same numbers
- Prospecting campaigns show almost no revenue while retargeting looks amazing
- Many orders are unattributed even though your ads are tagged
What to do:
- Pick a longer window and see how the numbers change
- Compare Clakta's attributed revenue with your total revenue to see the size of the gap
- Read retargeting numbers as "closed the sale", not "caused the sale"
Privacy-friendly projects
In privacy-friendly projects, Clakta only remembers a visitor for the day. So in practice a click only gets credit for a purchase on the same day: a click on Monday and a purchase on Wednesday belong to two different anonymous visitors. A longer window doesn't change that.
In full attribution projects the visitor is remembered, so the whole window counts. If your customers usually take several days, privacy-friendly mode will show less paid performance whatever model or window you use. See Tracking modes.
Check this before deciding a channel doesn't pay off. A privacy-friendly project whose customers take 10 days to decide can make good prospecting look like wasted money.
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 models
The five models Clakta offers, how each one shares the credit for a sale, and which one to use for each decision.
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.