Google Signals sunset: what changed on June 15, 2026 (and what to do)

On June 15, 2026, Google Signals stopped acting as the ad-data co-controller in GA4. If you built remarketing audiences or cross-device reporting on Signals, the ground shifted under you — and, as usual, nothing in your dashboard lit up red when it happened.

What actually changed

Google Signals used to enrich GA4 with signed-in Google users' cross-device and demographic data, and it doubled as an ad-data control surface. As of the sunset, that ad-data authority consolidates under Consent Mode v2 signals managed inside Google Ads — ad_user_data and ad_personalization. Signals no longer gates ad data.

• Cross-device reporting that leaned on Signals degrades to modeled / first-party data.
• Remarketing audiences built on Signals need to be rebuilt on Consent Mode v2 signals.
• The allow_google_signals gtag flag is now effectively a no-op.

The gotcha nobody warns you about

Plenty of GA4 snippets in the wild still emit allow_google_signals: true. It used to mean something; now it is dead weight. Worse, if your setup was quietly leaning on Signals as the ad-data control point, ad data may now flow based purely on Consent Mode v2 state — which makes a missing or misconfigured CMP more consequential than before, not less.

💡 Note: If your generated GA4 snippet still sets allow_google_signals, drop it. TagEasy stopped emitting the flag in generated snippets at the sunset — it does nothing now, and set to true it is exactly the kind of stale ad-data signal our own consent audit flags.

What to do now (a 30-minute checklist)

1. Confirm Consent Mode v2 is live and correct — ad_user_data + ad_personalization must reflect real user choice. (See our Consent Mode v2 on Shopify post for a config that actually works.)
2. Rebuild any Signals-based remarketing audiences in Google Ads on Consent Mode v2 + first-party data.
3. Remove allow_google_signals from your gtag config — it is a no-op.
4. Re-audit: the consent-v2 check in the free TagEasy auditor tells you whether the gtag consent defaults are present in your initial HTML.

Why this is easy to miss

Same reason most tracking breakage is invisible: nothing errors. Reports keep rendering, pixels keep firing, and the audience just quietly stops growing. The feedback loop is broken — which is the whole reason we publish tracking-quality scores at /benchmarks. A visible number beats a silent regression.