What Changed
GA4's Advertising section was reorganised: the Attribution paths and Model comparison reports were merged into a single tabbed Attribution report, and Journey analysis now defaults to showing only the top ten conversion paths. A separate June change altered how traffic sources are grouped and how consent controls govern what GA4 can share with Google Ads.
The practical effect is that period-over-period comparisons straddling these dates show movement that is a measurement artifact rather than a real performance change, and remarketing lists built from GA4 shrink to true consent capture rate.
Source pending: replace with the Google Analytics Help release notes before publishing.
Why It Matters
The failure mode here is not lost data, it is a false performance read. A channel that appears to have declined may simply be receiving less attributed credit under the reorganised reporting, and budget cut in response to that would be a self-inflicted wound.
Annotate the change dates in every reporting view and stop making period-over-period comparisons that straddle them. Where a client needs a clean trend line across the boundary, rebuild it from BigQuery event-level data with a consistent model applied outside GA4, rather than arguing about which interface number is correct.
Two secondary effects worth checking. Journey analysis now surfaces fewer conversion paths by default, so multi-touch analysis needs the report adjusted manually rather than read as-is. And GA4-built remarketing lists now reflect true consent capture rate, so list shrinkage is expected and is not a tagging fault.
[Replace source with Google Analytics Help release notes before publishing.]