What is lockWindow?

What it prevents

Without a lock, a re-engagement campaign reaching a recently acquired user could claim their subsequent events — so the source that acquired them loses credit to a source that merely advertised to someone already using the product.

lockWindow prevents that by holding attribution with the acquiring source for a defined period, regardless of what other campaigns reach the user in the meantime. It is a rule about precedence rather than about measurement accuracy.

Setting it

• Long enough to cover the acquisition's real influence — the period during which the user's activity is plausibly a result of being acquired rather than of anything since.
• Short enough not to block genuine re-engagement, which does eventually deserve credit.
• Consistent across sources, or the comparison between them is meaningless.
• Documented on every report, since a number produced under a different lock is not comparable.

Why the settings layer matters

lockWindow sits with attribution windows, reattribution windows and time of inactivity in a group of configuration values that quietly determine reported results.

None of them are measurements. They are policy choices about how to divide credit, and a campaign's apparent performance can be changed substantially by adjusting them without anything real changing at all.

This is why cross-vendor and cross-account performance comparisons need the settings checked first, and why a holdout — which does not depend on any of these settings — is the only comparison that survives them.