Why Your EU Google Ads Data Just Vanished

Google Consent Mode v2 implementation checklist

I treat this section like fast triage. The goal is straightforward: get Consent Mode v2 working well enough that Google Ads can keep optimizing, GA4 stays usable, and consent choices are respected consistently.

Fast actions to fix first

1. Confirm your site fires tags via Google tag or Google Tag Manager
   If parts of the site still use hard-coded tags that ignore consent, you’ll end up with inconsistent behavior. I prefer having core GA4 and Google Ads tags run through Google tag or GTM so they can reliably listen to consent signals.

2. Ensure consent defaults are set correctly
   Where your legal obligations require it, consent should default to denied until the user makes a choice. In practice, that means analytics_storage, ad_storage, ad_user_data, and ad_personalization start as denied for new sessions in affected regions.

3. Connect your CMP to consent updates
   Your consent management platform should push updates into Consent Mode v2 when a visitor accepts, rejects, or adjusts settings. If there’s no update, Google never receives a usable signal. If you’re evaluating vendors, start with Google’s Consent Management Platform (CMP) Partner Program list.

4. Verify Google Ads and GA4 tags receive consent states
   Use GTM Preview or Tag Assistant to confirm each relevant tag can “see” the consent status for the four signals before it fires.

5. Test in Tag Assistant and GTM Preview
   In a private browser window, go through the banner multiple ways (accept all, reject all, partial) and inspect what fires. I’m looking for consistent behavior across those paths, not just one “happy” path. (If you need a refresher on how Preview works, Google’s GTM Preview documentation is a useful reference: Preview and debug.)

6. Recheck conversions and reporting
   Within a day or two, review GA4 and Google Ads for conversion stability, region splits, and audience growth. At this stage, stability matters more than dramatic jumps.

Pass or fail: simple Consent Mode v2 table

Use this as a quick scan with your team. You don’t need to read code to ask these questions and verify the screens.

If you can walk through that table and give clear answers, you’re already in better shape than most teams I see. A small set of screenshots from Tag Assistant or GTM Preview is often enough to keep conversations grounded when numbers get noisy.

If you want a clean baseline for what “good” tracking looks like in a B2B funnel, pair this with B2B Conversion Tracking Checklist: GA4 Events That Matter.

Consent Mode v2 in 2026

By 2026, Google Consent Mode v2 isn’t a “nice to have.” It acts as a gatekeeper for several GA4 and Google Ads features many teams assume will “just work.”

If you have EEA or UK traffic and you use GA4 or Google Ads, Google expects you to collect consent, send the four consent signals, and honor user choices. When that doesn’t happen, the practical impact is that remarketing lists can stall, demographic and interest reporting can disappear, and parts of conversion tracking can degrade for users in regulated regions.

For a B2B service firm, the consequence isn’t just fewer lines in a dashboard. It tends to show up as fewer trackable form fills and demo requests, shakier CAC calculations for paid channels, and messier attribution when you plan budgets. You may still close deals - you just lose visibility into which touchpoints helped, which bids to cut, and which campaigns deserve more spend.

Privacy enforcement pressure also plays into this. Frameworks like the GDPR and the Digital Markets Act push toward provable controls: personal data should only be used when a visitor actually agreed. That pressure flows into how platforms handle measurement, audiences, and personalization.

From a revenue lens, I think about three board-level risks: forecasts can swing because measurement changes (not demand changes), CAC can look worse when modeled conversions drop out, and channel ROI slides become less trustworthy unless someone checks consent behavior and modeling status first. Fixing consent won’t create demand by itself, but it can reveal more of the demand you already earned.

Consent Mode v2 vs v1

A common trap I still see is: “I already set up Consent Mode years ago.” Often that means Consent Mode v1 logic is still in place, while Google’s current expectations align with v2 behavior.

Consent Mode v1 mostly focused on storage. V2 adds expectations about how data may be used, introduces two additional signals, and tightens how audience building and ad personalization behave based on those signals.

Common migration pitfalls

Even when teams aim for Consent Mode v2, a few issues show up repeatedly. Legacy tag clutter is a big one: old GA and Ads tags sit next to newer ones, so reports become inconsistent. Duplicate tagging is another: the same event fires from GTM and from hard-coded scripts, so you get double counting in some places and missing data in others.

The CMP layer can also be the weak link. I often find category mapping gaps, where “marketing” toggles don’t translate cleanly to all four Consent Mode signals (for example, mapping to ad_storage but leaving ad_user_data denied). Finally, there’s timing: tags fire on the first page view while consent is still “default,” and only later does the CMP initialize and send an update - meaning the first hit of many sessions is misclassified.

For a simpler explainer you can share internally, see Consent Mode v2 in plain English for ecommerce (the mechanics apply to B2B sites too).

Consent signals

Consent Mode v2 revolves around four yes/no signals that apply per visitor and session. Each one controls how tags behave and what Google tools can do with the data that’s allowed.

analytics_storage controls whether storage can be used for analytics (think GA4 page views and events).
ad_storage controls storage used for advertising measurement (including conversion-related measurement).
ad_user_data controls whether certain user data can be sent to Google for ads purposes.
ad_personalization controls whether data can be used for remarketing and personalization features.

How each signal maps to impact

In practice, I sanity-check each signal without reading code: I confirm that “accept all” produces granted states where expected; that “reject all” produces denied states; and that partial choices map consistently to the signals you intend. When consent updates arrive late (after tags fire), you can do everything “right” on paper and still get unreliable results in reporting and attribution.

If remarketing is a meaningful part of your acquisition mix, this pairs well with Smart remarketing architecture for B2B and ecommerce because audiences and personalization are where consent mistakes show up first.

Consent Mode modes

Consent Mode v2 generally runs in two modes: basic and advanced. Both rely on the same four signals; the difference is what happens when a user says “no.”

I don’t treat this as a purely technical decision. It’s a measurement strategy decision: what level of visibility do you need, and what constraints are you operating under?

Basic mode

In basic mode, if a user denies consent for a category, tags for that category don’t fire. That means no cookies and no additional pings for that category. When consent is granted, tags behave normally.

Basic mode can be acceptable if you run very limited paid media, if your organization explicitly prioritizes legal risk reduction over measurement depth, or if you rely heavily on other internal systems for performance validation. The trade-off is sharp data loss, and in B2B - where conversions are already low-volume - losing even a small number of conversions can materially distort channel and campaign decisions.

When basic mode is harming measurement, I usually see the same patterns: EU and UK conversions drop close to zero while sales activity in those regions doesn’t mirror that decline; Google Ads shows far fewer conversions than your internal lead counts; and bidding systems become conservative because the platform is starved of signals. At minimum, I want consent choices to map cleanly to the signals, tags to respect those signals on every template (including older pages), and a parallel regional view in your CRM or lead system so you can spot measurement drift.

Advanced mode

Advanced mode is more flexible. When a user denies consent, tags can still send anonymous, cookieless pings. The intent is not to identify or target the user who said no, but to preserve enough aggregated signal for modeling and trend analysis.

Advanced mode only works as well as the setup behind it. I focus on a few prerequisites: the four signals must be correct and arrive on time; tags must behave differently under denied vs granted states; CMP integration must be reliable with clear category-to-signal mapping; and you need enough volume to meet platform thresholds for modeling eligibility.

Operationally, I see better outcomes when a single owner is accountable for consent and tracking decisions, changes are tested after releases and CMP updates, documentation is kept simple (a basic flow diagram is often enough), and GTM changes are versioned clearly so rollbacks are possible.

If bidding performance is part of the conversation, link your consent work to how optimization actually happens in-platform: Smart bidding in simple words and when to use manual bids.

Consent management platform

Many teams conflate Consent Mode with the consent banner itself. I separate them like this: the CMP is the interface and system that collects and stores the visitor’s choice, while Consent Mode is the framework that translates that choice into behavior for Google tags and tools.

In practice, you typically need both: the CMP asks, and Consent Mode listens and adjusts GA4 and Google Ads behavior. When I evaluate whether a CMP is suitable for a B2B service site, I prioritize fundamentals over design polish: whether it integrates cleanly with GTM or Google tag, whether it supports geo-based rules (for example, stricter defaults in EEA and UK), whether it maintains usable consent logs, what its performance impact looks like, and whether it’s maintained well enough to keep up with policy and platform changes.

When configured correctly, a CMP plus Consent Mode v2 helps ensure cookies and personal ad data don’t flow before consent where consent is required. If you run advanced mode, the “no consent” path can still produce anonymous, aggregated measurement signals used for modeling - without turning “no” into “yes.”

Conversion modeling

Conversion modeling often gets described like magic, but I treat it as statistics built on top of reliable signals. If Consent Mode v2 is misconfigured, modeling won’t fix bad inputs.

Modeling exists because not every conversion can be tied back to a click or user - especially as more people decline tracking. Platforms also apply eligibility thresholds before modeling kicks in, and smaller B2B sites can sit close to those thresholds. That means small changes in traffic, consent rates, or implementation timing can flip modeling quality on and off.

Integrity checks I rely on

• Compare GA4 conversion trends with CRM and lead-system outcomes by market - I look for directionally similar movement, not perfect one-to-one matching.

• Watch for sudden breaks by region, device, or browser. Sharp EU-only or Safari-only drops often point to consent or storage behavior rather than real demand changes.

• In Google Ads, monitor the relationship between observed and modeled conversions. Abrupt swings often indicate an implementation change, consent mapping shift, or timing issue.

• Validate timing: consent signals should arrive before key events (especially conversions) fire, otherwise results drift into “unknown” or get misattributed.

When numbers look wrong, the root cause is often simple: the CMP never sends the update event, tags fire before consent initialization, duplicate implementations double-count or conflict, defaults are wrong for a region, or web vs server-side setups interpret consent differently.

If I have limited time to diagnose, I use a short workflow: test one or two key pages while walking through multiple consent paths in Tag Assistant and GTM Preview; capture evidence of the four signal states under each path; compare recent vs prior conversions by region in GA4; and check whether Google Ads reporting shows continuity in observed vs modeled behavior for top campaigns. After fixes go live, I monitor daily for sharp changes (especially in EEA and UK), keep a simple change log for site and tag updates, and then re-compare CRM vs GA4 trends after a week to confirm the system stabilized.

The aim is simple: Consent Mode v2 should give you numbers that are defensible, explainable, and consistent enough to base budget and forecasting decisions on - even when consent rates vary by market. If you want to push reliability further, tie web measurement to outcomes using Offline Conversion Imports: The Only Signals Google Ads Should Optimize For.