The short answer
A conversion window determines how long after an eligible ad interaction a conversion can be counted under a platform’s rules. Changing the window changes which conversions are eligible for reporting. It does not change the underlying order revenue in the store.
Record the actual measurement settings
Separate click-based, view-based, and other eligible interactions. Also record the attribution model, which allocates credit among eligible interactions. The window and the model answer different questions. Google documents conversion windows in its official guide; settings and supported options should be checked in the relevant account.
Illustrate the timing effect
Illustrative scenario: a campaign spends $1,000. Under a shorter reporting window, it receives $2,000 of revenue credit, giving $2,000 ÷ $1,000 = 2× ROAS. A longer-window view of the same activity includes an additional $500 of eligible credited revenue, giving ($2,000 + $500) ÷ $1,000 = 2.5×. This scenario assumes only the eligible revenue changes and does not predict the size of a real window effect.
A larger number is not automatically better evidence
A longer window can include purchases occurring farther from an interaction. That can be useful for a longer buying cycle, but the resulting credit is not proof that the advertising caused each purchase. A shorter window can miss later purchases. Choose and disclose a rule suited to the measurement question rather than choosing the one with the most attractive ratio.
Avoid comparing incompatible periods
If settings change, record the date and whether reports are comparable. Do not assume old data is retroactively recalculated. Keep an export or a record of the earlier definition where appropriate. A trend crossing a settings change should be annotated, especially when a budget decision depends on it.
Separate eligibility from reporting delay
A conversion may be inside the chosen window but not yet have happened or been reported. That is a maturity issue, not necessarily a window issue. Before comparing recent and older periods, assess how much time each has had to collect conversions.
For a review, write one sentence that states the source, interaction type, window, model, date basis, and extraction date. If you cannot obtain these details, treat precise cross-platform ROAS rankings cautiously.
Record your reporting rules
- Gather: Note the platform, model, eligible interactions, window and date basis.
- Produce: Annotate settings changes and compare reports using compatible rules.
Decision boundary: A longer window changes eligibility; it does not prove additional causal sales.
Next reading
Examples are illustrative, not client results or benchmarks. The calculations use the inputs shown; real-world decisions depend on your records and assumptions. Report a correction.