Spring naar content

This is a common issue with Consent Mode v2 in 2026. A visitor sees the banner, makes a choice and the tag fires while the dashboard stays green, but the consent decision often does not reach Google. Much of the setup looks healthy right up to the point where the data disappears. Because there is usually no warning email, many teams discover the problem only when a report falls apart.

What changed and why it matters now

Consent Mode v2 is Google’s mechanism for telling its tags whether a visitor granted consent to tracking. It has been mandatory for advertisers serving or measuring users in the EEA and the UK since March 2024, driven by the EU’s Digital Markets Act rather than by GDPR directly. For a while, “mandatory” carried little consequence, and plenty of teams implemented it loosely or skipped it.

That changed when Google began enforcing it. From July 2025 onward, accounts that were not passing correct consent signals for EEA and UK traffic started losing access to personalized advertising, remarketing and conversion tracking for the affected users.

Consent Mode v2 works through four signals. Two carry over from the earlier version: ad_storage, which governs advertising cookies, and analytics_storage, which governs analytics cookies. Two were added in v2: ad_user_data, which controls whether user data may be sent to Google for advertising, and ad_personalization, which handles remarketing. A consent platform that sets only the first two is running an incomplete integration. The banner appears, the tag fires, and the signal Google needs is missing from the request.

Why a healthy-looking setup can still be broken

The problem is that the visible parts of the chain look fine. A consent banner that collects users’ choices but does not share the corresponding updates with Google’s tags is basically decorative. A clean report in a tag testing tool confirms that a tag fired, not that the consent signal followed with it. So the failure tends to hide in the step that teams rarely inspect, which is the handoff between the banner collecting a choice and Google receiving it.

This is the opposite of what most people expect. Consent problems are usually thought of as a banner that does not appear or a cookie that fires too early. Here, the banner behaves correctly, and the wiring behind it drops the signal before it reaches its destination.

What server-side tracking changes

Most consent platforms run in the browser. The script loads from a third-party domain, collects the visitor’s choice and hands it to the tags, all of which happens client-side. It is quick to deploy and inherits the weaknesses of client-side tracking. Ad blockers can interfere with the third-party script, and browsers like Safari significantly cap cookie lifetimes, which repeatedly resets returning visitors and prompts them for consent again.

Server-side tracking changes the setup in a way that directly relates to consent collection. The consent choice is still collected in the browser, since it has to be gathered where the visitor is. Instead of the browser deciding what fires and what gets sent, that decision moves to a server you control, usually running on a subdomain of your own site. Your consent management platform (CMP) remains where consent is collected, but it is now enforced on the server.

In practice, the consent choice passes from the browser to your server container as a signal attached to the event. The tags inside that container read it and act on it. A tag fires when the visitor has agreed to its purpose and the vendor behind it, and data reaches an advertising platform when the visitor has consented to that specific use. Because this enforcement happens on your infrastructure, there are fewer places for the signal to be dropped, blocked or altered. You can inspect what was received rather than trusting that it arrived.

Relying on the server for enforcing consent decisions also changes who controls the rules and whether you can review them later. In a browser setup, the logic that decides what fires lives inside third-party scripts you neither own nor can inspect after the fact. For example, when a vendor changes its script, your enforcement can change with it. With server-side tracking, those rules are defined in one place you control. They are applied consistently to every tag, and each decision leaves a record you can go back to when a regulator, a client or your own team asks what was sent and why.

Conversion modeling can help fill part of the gap

Compliance is one reason to get consent signaling right. Another is recovering some of the conversions you lose when visitors decline tracking. Consent Mode offers two configurations, Basic and Advanced. In Advanced mode, tags load in a restricted state before the visitor decides and send anonymous, cookieless pings. Google uses those pings to estimate conversions that go unmeasured in those cases, recovering attribution that would otherwise go unrecorded.

However, conversion modeling doesn’t guarantee recovery and it does not apply to every account. Google only models conversions for accounts above a certain threshold of daily events, so smaller advertisers may see little or none of it. Correct consent signaling is the part you control; modeling is a variable addition on top.

What to check in your tracking setup

First, confirm that your consent platform sets all four signals, not just the two carried over from the previous version. An integration that omits ad_user_data and ad_personalization is incomplete by definition.

Verify that the consent choice reaches Google – don’t just assume it does because the banner appears and the tag fires. An error manifests itself in the gap between those two. It’s hard to confirm by eye, since a broken signal won’t be visible in the interface. TAGGRS offers a free website tracking checker that helps you verify whether consent is actually passed through. Scan your site and get a report on which tags fire, what cookies are set and how long they last, and whether your consent signals reach each platform.

Decide between Basic and Advanced mode on purpose rather than defaulting into one. The choice is important as it determines whether you are eligible for modeled recovery at all.

Lastly, consider where consent enforcement happens. Handling it on a server you control instead of a chain of browser scripts gives you a place to verify the signal and reduces the points where it can break without warning.

The bigger picture

Consent Mode v2 is a specific technical requirement, but it points to a broader shift. Measurement is moving out of the browser because it has become unreliable, whether that is due to a limited cookie lifetime, an ad blocker or a consent signal that has to reach its destination complete. Collecting consent is one part of the task. Making sure the decision is honored, and being able to prove it was, will help you strengthen your setup and its compliance.

A healthy-looking dashboard is worth as much as your confidence that the signal behind it is arriving. Checking whether it does takes little time and costs far less than discovering the gap when a report stops showing you the data.

Taggrs is sponsor of DDMA Digital Analytics Summit 2026. More information and tickets for Digital Analytics Summit taking place on 1 October at Hotel Casa in Amsterdam. 

Małgorzata Poddębniak

Content Marketeer | Taggrs

Ook interessant

Lees meer
Digital Analytics |

Your consent banner works. The signal behind it might not

Sometime in the past year, a number of advertisers watched their Google Ads conversions drop sharply. Campaigns were still running, clicks were arriving, budgets had not changed, yet the conversions…
Lees meer
Digital Analytics |

The consent debate is a distraction from the real measurement problem

The measurement conversation in our industry is currently dominated by regulation. Since the European Commission published its Digital Omnibus proposal in November 2025, I have seen countless posts and panels…
Lees meer
Digital Analytics |

What does it really cost to own an analytics platform? Software, migration, and maintenance costs explained

Analytics tools are often viewed as a cost rather than an investment. That framing alone tells you a lot. A “cost” suggests operational overhead. An investment is capital deployed to…