On double attribution, attribution windows, and the right way to measure each channel's contribution.
Add the revenue Meta reports to the revenue Google reports, and you get a sum higher than the store's total revenue. This is not a technical error. It is how the platforms are built.
Each platform counts for itself
A customer saw an ad on Instagram on Sunday, searched for the brand on Google on Tuesday, clicked a search ad and bought. Meta records the purchase because its ad was seen within the attribution window. Google records it because the last click was its own. One purchase was counted twice.
Attribution windows
Each platform defines how long after a click or a view it still takes credit. A long window increases reported revenue, a short one reduces it. Comparing platforms with different windows is comparing two different rulers.
What to measure instead
- The ratio between total revenue and total ad spend. One metric that does not depend on attribution.
- Revenue from new customers separately from revenue from returning customers.
- An external attribution tool, such as Triple Whale, that sees all channels on the same ruler.
- Holdout tests: what happens to total revenue when one channel is paused for a defined period.
The reports of the ad platforms are useful for comparing campaigns within the same platform. To decide how much to invest in each channel, you need to look at the till.