Jump to a popular page, or start typing.

    BLOG / how-to

    Consent Mode v2 without breaking Google Ads optimization

    Consent Mode v2 has been mandatory for EEA and UK ad personalization since March 2024, and it is still the most common thing we find broken in Google Ads accounts.

    Consent Mode v2 has been mandatory for EEA and UK ad personalization since March 2024, and it is still the most common thing we find broken in Google Ads accounts. Usually nobody noticed the break, because campaigns kept spending. They just learned less from every euro.

    What it actually is

    Consent Mode passes the user’s consent state with every Google tag and API call, across four signals: ad_storage, ad_user_data, ad_personalization, analytics_storage. Your consent banner records the choice; Consent Mode transmits it. Two of these matter enormously for performance. ad_user_data gates whether Google may use your hashed identifiers at all, which means Enhanced Conversions and Customer Match silently stop working for users where it is missing. ad_personalization gates remarketing.

    The failure modes we see weekly

    The banner fires after the tags, so the first pageview reports no consent regardless of the user’s choice. The CMP updates the browser state but the server-side pipeline never hears about it, so your CAPI events carry no consent signal. Advanced mode was enabled without understanding that unconsented pings feed modeling only. Or the v1 implementation was never upgraded, so ad_user_data does not exist and Google treats it as denied.

    Why doing it right recovers performance

    With Consent Mode implemented, Google models conversions for the consent gaps instead of counting them as zero, and consented users’ hashed identifiers keep Enhanced Conversions alive. Advertisers frequently see reported conversions rise after a correct implementation. The data was always there; the permission slip was not attached.

    The implementation checklist

    CMP configured for the four v2 signals. Consent state readable before any tag fires. Server-side propagation, so every event leaving your infrastructure carries the flags the browser collected. Verification in Tag Diagnostics, then in the Ads interface consent diagnostics. And a regression test whenever the banner vendor ships an update, because they do.

    Consent is also simply the deal you made with your users. Enforcing it end to end, browser through server, is what makes the rest of your first-party data program defensible. How Signals handles consent covers the mechanics.

    Get your Consent Mode implementation checked

    PUT IT TO WORK

    Reading is good. A live setup is better.