Consent mode and marketing measurement in Europe: what you lose, what you keep and what to fix

Why measurement changed

European privacy rules require consent for many non-essential cookies and similar technologies, including most advertising and many analytics uses. A proportion of visitors decline, ignore banners or choose limited consent. The result is fewer observed sessions, fewer attributed conversions and advertising platforms with less information for optimisation.

Marketing teams often notice this as a sudden drop in reported conversions after introducing a compliant banner. Sales may not have fallen at all. The measurement system has simply become more honest about what it can observe.

Google also requires advertisers using its measurement and advertising features for users in the European Economic Area to pass consent signals, which made consent mode a practical requirement for many businesses rather than an optional improvement.

What consent mode actually does

Consent mode is a mechanism that tells Google tags whether a visitor has granted consent for specific purposes, such as analytics storage, advertising storage, advertising user data and advertising personalisation. Tags then adjust their behaviour. When consent is granted, they work normally. When consent is denied, they do not set or read cookies for those purposes.

In basic implementations, tags are blocked until the visitor makes a choice, and no data is sent for visitors who deny consent. In advanced implementations, tags load with consent denied by default and can send cookieless signals without identifiers, which platforms can use for conversion modelling.

The choice between basic and advanced implementations is not purely technical. It should be assessed with privacy advisers against applicable law, the banner’s wording and the organisation’s risk appetite.

Consent granted

Tags use cookies and identifiers for the permitted purposes.

Consent denied, basic mode

Tags do not send data for the denied purposes.

Consent denied, advanced mode

Tags may send limited cookieless signals used for modelling.

Modelled conversions are estimates

Advertising and analytics platforms can estimate conversions from visitors who did not consent, using patterns from visitors who did. Modelling can make reports more complete, but it is still an estimate. Accuracy depends on traffic volume, the share of consented users and the similarity between consented and non-consented behaviour.

Teams should treat modelled numbers as directional rather than exact. Comparing platform-reported conversions with actual sales, leads or CRM outcomes remains essential.

Common implementation mistakes

The most common mistake is a banner that looks compliant but does not control tags. Scripts load before choices are made, the reject button does nothing technically, or consent mode signals are sent incorrectly. This creates legal risk and unreliable data at the same time.

Other mistakes include default consent states set to granted in regions where consent is required, missing signals for newer consent types, tags hardcoded outside the tag manager, and conversion pixels on pages where sensitive information is involved.

Testing should use browser developer tools and platform diagnostics to confirm what happens before consent, after acceptance and after rejection.

First-party data and server-side measurement

Measurement becomes more reliable when important outcomes are recorded in first-party systems. A CRM that records lead source, qualification and closed revenue gives a clearer picture of marketing performance than advertising platform dashboards alone.

Offline conversion imports can send qualified outcomes back to advertising platforms where consent and legal basis allow it. Server-side tagging can improve control over what data is sent to third parties, but it does not remove consent obligations and must be configured carefully.

Reporting that survives incomplete data

Marketing reporting should combine several views: platform data for optimisation, analytics for on-site behaviour, CRM and sales data for business results, and periodic tests such as geographic or time-based experiments for incrementality.

Teams should also explain data limitations to management. A report that says a share of traffic is unobserved because visitors declined tracking is more useful than a dashboard that silently undercounts and leads to wrong budget decisions.

This article describes measurement approaches. It is not legal advice, and organisations should confirm consent requirements with privacy advisers.

Where this sits in what we do

This article covers one decision inside a wider engagement. The solution page sets out how that engagement runs, what it includes and what it costs to find out.

• Digital Growth — The ongoing programme — strategy, content, paid, organic and reporting — run as one thing that answers to pipeline rather than to channel dashboards.
• Why more traffic rarely fixes a pipeline problem
• Lead scoring that sales will actually use
• Tender documentation for construction and engineering firms: stop rebuilding every bid from scratch
• Cookie banner design: why refusing must be as easy as accepting
• All insight articles