What is Attribution Window?
The period after seeing or clicking an ad within which a platform will claim credit for a resulting conversion.
Common settings: 7-day click, 1-day view
A worked example
Someone sees your Instagram ad on Monday, searches your brand on Wednesday, and buys on Thursday. Under a 7-day click and 1-day view window, Meta may claim that sale. Google Ads may also claim it through the branded search. Your backend records one order, while the two dashboards together report two.
Why it matters
Attribution windows explain the single most common question in paid media: why platform-reported sales exceed the orders in your own system. Each platform measures independently, each claims what falls inside its own window, and nobody reconciles them for you.
This matters beyond reporting tidiness, because the window also shapes optimisation. A campaign optimised on a 7-day click window is being told that slower, more considered conversions count. Narrow it to one day and the algorithm shifts toward buyers who act immediately, which changes who it targets and what creative it favours.
For businesses with long consideration cycles, real estate, education, B2B and high-value healthcare among them, a short window systematically undervalues the channels doing the early work. Awareness activity looks useless because the conversion arrives weeks later and gets credited to whatever the customer touched last.
The nuance most people miss
View-through conversions are where the biggest overstatement usually sits. A view-through means someone saw the ad, did not click it, and later converted. Some of that influence is real and some of it would have happened regardless, and no platform separates the two for you. When reconciling, look at your click-based numbers first and treat view-through as directional. The honest reference point is always total delivered revenue in your own system compared against total ad spend, not the sum of what each platform claims.
Common mistakes
- Adding platform-reported conversions together across Google and Meta, which double-counts the same orders
- Using a short window for a product with a long consideration cycle
- Changing the window mid-quarter and then comparing performance across the change
- Reading view-through conversions as though they were clicks