What is a reattribution window?

The two periods involved

Re-engagement measurement uses two distinct time settings, and conflating them is the most common configuration error.

The inactivity period is how long a user must have been absent before they count as lapsed at all. It defines who is eligible to be reattributed.

The reattribution window is how long after the re-engagement ad a return can still be credited to it. It defines how long the credit lasts.

Setting them defensibly

• Base the inactivity period on your usage rhythm, not on a platform default. It should sit clearly beyond a normal gap between sessions.
• Keep the reattribution window short. Re-engagement is meant to prompt an immediate return; a long window collects spontaneous ones.
• Separate click and view windows, as with any attribution window — view-through re-engagement is especially prone to over-claiming.
• Record both settings on every report, so a comparison across periods or platforms is possible.

Why the settings decide the result

A short inactivity period and a long reattribution window will report re-engagement as extraordinarily effective, because you are crediting ordinary returning usage to advertising. Lengthen the first and shorten the second and the same campaign looks mediocre.

Nothing about the campaign changed. This is why re-engagement performance should never be compared across accounts or vendors without verifying both settings, and why the only durable read on the channel is a holdout against the lapsed population — see incrementality.