Attribution Windows Are a Policy Choice
A 30-day attribution window isn't a fact about how customers behave. It's a default nobody chose on purpose - and changing it changes which channel gets credit for the same sale.
Most attribution windows are whatever the platform shipped with. Thirty days is common. Some default to seven, some to ninety. Almost nobody sat down, looked at how long their own customers actually take to buy, and picked a number that matched. The window came pre-installed, and it stayed.
That would be fine if the number were neutral. It isn't. A window is a claim about how far back in time credit for a sale should reach - and that claim shapes the report before anyone reads a single row of it.
Shrink the window and only the last click or two before a purchase gets credit - usually whatever channel someone touched right before checking out. Stretch it and channels further back in the journey start getting counted too, even ones that only distantly nudged the outcome.
Neither version is lying, exactly. They're both counting real interactions. They're just drawing the line at a different point in the same story, and a different line produces a different winner.
Ad platforms don't set their default window arbitrarily. A longer window means more of a customer's journey falls inside it, which means more sales end up credited to that platform - including ones a shopper would have made anyway. That's not a conspiracy. It's just what happens when the entity setting the ruler also benefits from a longer ruler.
None of this means the platform's number is fabricated. It means the number reflects the platform's default assumption, not a business's actual sales cycle.
A more useful starting point is the business's own sales cycle - how long does a typical customer actually take between first contact and purchase? A window shorter than that cycle systematically undercounts the channels that start the journey. A window much longer than it overcredits channels for coincidence, not influence.
Whatever window gets picked, the choice is worth writing down somewhere everyone reading the report can see it - not because the number needs defending, but because a different window next quarter will otherwise look like a change in performance instead of a change in the ruler.
Not inherently - it's wrong only if it doesn't match how long this specific business's customers actually take to decide. For a fast, low-consideration purchase it might even be too long.
Not necessarily. A channel that tends to start a journey and one that tends to close it can reasonably be measured on different timeframes, as long as that difference is deliberate and documented, not accidental.
It can retroactively change how past conversions are credited, depending on the platform, which is exactly why switching windows without noting it is where most "unexplained" performance shifts actually come from.
Look at time-to-purchase in the actual data - the gap between a customer's first recorded touchpoint and their eventual conversion - rather than assuming a platform default already matches it.
If nobody can say why the attribution window is set where it is, that's usually a sign it was never actually set - just inherited.