Attribution model
An attribution model is the rule for which ad gets the credit when a person saw several.
The model decides who gets the credit; the attribution window decides how long after a click or a view a sale still counts. The semantic layer names both in the description of each version of revenue, so a reader can tell why two systems report two different numbers for the same sales.
In the data model, each order carries the channel the brand's attribution rule credits, so new customers by channel and the customer acquisition cost built on them rest on that rule, and the rule can be wrong.
Any return on ad spend (ROAS) built on attribution can count sales the ads did not cause. An incrementality test, which holds some people back from the ads, counts only the sales the ads caused.
Example
Pinterest's Conversion insights report credits the last Pin a person engaged with before checkout, organic or ad. Google's Ads Help expects discrepancies, often up to 20%, from different attribution models, with Google Analytics against Google Ads as its example.