What is time of inactivity?

Two different thresholds

The same phrase covers two settings that operate on very different timescales, and confusing them causes real measurement errors.

Session inactivity is minutes. It decides when one visit ends and the next begins — see session.

Reattribution inactivity is days or weeks. It decides when a user counts as lapsed, and therefore when a returning visit can be credited to a re-engagement campaign — see reattribution window.

Setting them

• Base both on your product's actual rhythm, not on a platform default.
• Session timeouts should reflect how long a natural pause is — reading, thinking, switching apps briefly.
• Reattribution inactivity must sit clearly beyond a normal gap between uses, or ordinary behaviour gets credited to campaigns.
• Document them alongside every report, because a number produced under different settings is not comparable.

Why it decides results

These settings are unusually consequential relative to how little attention they get. A short reattribution inactivity period makes re-engagement look extraordinarily effective, because users who were between normal sessions are counted as recovered lapsed users.

Nothing about the campaign changed. This is why re-engagement performance cannot be compared across accounts or vendors without checking both settings first, and why a holdout is the only comparison that survives a settings difference.