Consent Mode v2: why did my Google Ads conversions drop?

TL;DR

• Google now switches off conversion tracking, remarketing and demographic reporting for EEA and UK traffic on accounts without a working Consent Mode v2 setup. The big enforcement wave landed in July 2025.
• Your campaigns didn't change. Your measurement did. Real sales can stay flat while the reported conversion count falls off a cliff.
• Most broken setups fail in the same spot: the cookie banner never sends a consent signal, or sends it after the Google tag has already fired.
• Google's Marketing Platform blog says conversion modelling recovers more than 70% of the ad-click-to-conversion journeys lost to consent choices. Google Ads Help says modelling needs 700 ad clicks over 7 days per country and domain grouping before it starts.
• Start in the Diagnostics tab of a conversion action. It shows your consent mode status in about a minute.

If your Google Ads conversions fell off a cliff and your actual sales didn't, look at your cookie banner before you touch a single campaign.

Google spent two years phasing in Consent Mode v2 for the European Economic Area and the UK, then started enforcing it. Accounts that never sent proper consent signals lost conversion tracking and remarketing for that traffic. The campaigns kept spending. The reporting went quiet. Plenty of European business owners spent that quiet period rewriting ad copy that was never the problem.

Why did my Google Ads conversions drop in Europe?

Because Google stopped counting conversions from visitors whose consent status it can't read. Consent Mode v2 has been required for EEA and UK traffic since 6 March 2024, and from July 2025 Google began turning off conversion tracking, remarketing and personalisation on accounts that never implemented it properly. The ads are fine. The pipe carrying the data back is cut.

The phase-in ran in steps, which is why so many advertisers felt it as a slow leak rather than one bad Tuesday. March 2024 stopped new EEA and UK visitors going into audience lists. Later that year, existing remarketing audiences stopped topping up from that traffic. In 2025 the modelled conversions went too, and then the wider enforcement in July 2025 took personalisation and reporting features with it.

Here's the number worth knowing before you panic. Google's Marketing Platform blog puts conversion modelling recovery at more than 70% of ad-click-to-conversion journeys lost to consent choices. That figure was published in 2021 and Google hasn't refreshed it since, so treat it as a ceiling rather than a promise. It also only applies once modelling is switched on for your account, which brings its own conditions.

What is Consent Mode v2, in plain terms?

Consent Mode v2 is a set of signals your website sends to Google saying whether each visitor agreed to be tracked for ads and for analytics. Four signals do the work: ad_storage, analytics_storage, ad_user_data and ad_personalization. Google reads them, then decides what it's allowed to record about that person.

Two of those signals were added in v2. That's the whole upgrade. The reason it broke so many accounts is that adding signals meant editing the tag setup, and most small business sites had a cookie banner installed once, years ago, by whoever built the site.

Is my cookie banner the problem?

Usually, yes. Four failures cover most of the accounts we look at:

1. No default consent state is set before the Google tag loads. Google then treats every visitor as denied, including the ones who clicked Accept.

2. The banner loads after the Google tag. The signal arrives too late to matter.

3. The banner is decorative. It looks like a consent tool, blocks nothing, and sends nothing to Google.

4. The consent update on click never fires, so Accept and Reject produce identical signals.

None of these throw an error. Your site loads, your ads serve, your GA4 property still reports something. That's what makes it expensive. You can run this way for months and only notice when someone compares Shopify orders against Google Ads conversions and finds a gap they can't explain.

What's the difference between basic and advanced consent mode?

Basic consent mode blocks Google tags entirely until someone consents. Advanced consent mode loads the tags and sends anonymous, cookieless pings when someone refuses. Advanced gives Google more to model from, so it usually reports closer to reality.

 

Pick advanced unless your lawyer has told you in writing not to. If nobody has told you which one you're running, that's answer enough about the state of the setup.

How do I check if consent mode is working on my site?

Two checks, both free, both under ten minutes. Start inside Google Ads, then confirm on the live site with Tag Assistant.

In Google Ads, open Goals, then Conversions, then click one conversion action and open the Diagnostics tab. It reports a consent mode status. Google Ads Help notes the active status can take 48 hours to appear and occasionally up to two weeks, so a blank reading on a fresh setup isn't proof of failure.

In Tag Assistant, enter your site URL, then accept everything on your banner when the tab opens. Open the earliest Consent event in the summary and check that ad_storage, ad_personalization and ad_user_data are all set. Run it again and press Reject. If both runs look identical, your banner isn't talking to Google.

Why do GA4 and Google Ads disagree now?

They model separately, they count separately, and they hit their thresholds at different times. GA4 fills gaps with its own behavioural modelling. Google Ads fills gaps with conversion modelling that needs 700 ad clicks over 7 days per country and domain grouping before it will run at all, per Google Ads Help.

That threshold is the part small advertisers keep tripping over. A local business in Munich or Manchester spending a few hundred a month across two countries may never clear 700 clicks in a week in either one. Consent mode still matters for compliance and for remarketing, and the modelled top-up simply won't arrive. In that case, server-side tagging and enhanced conversions do more for your reported numbers than another week of bid tinkering.

Will the missing conversions come back?

The historical ones, no. Google doesn't backfill the period when your signals were absent. Going forward, some of the gap closes once your setup is fixed and modelling starts, and there's a training period after that before reporting settles.

Expect roughly this order: fix the banner, wait 48 hours for the status to show as active, then give it a couple of weeks before you judge the numbers. Comparing week one after the fix against your best month from 2024 will only make you miserable.

What should I fix first?

Work in this order and stop guessing at campaign level until the measurement is honest.

5. Confirm the consent mode status in the Diagnostics tab of your main conversion action.

6. Set the default consent state to denied for EEA and UK visitors, before any Google tag loads.

7. Move to advanced consent mode unless there's a written legal reason not to.

8. Turn on enhanced conversions so consented traffic sends stronger signals.

9. Note your consent rate. Published cookie consent benchmark reports from 2024 and 2025 put average European accept rates somewhere in the forties, and banner design swings that number more than anything else. Treat those figures as an estimate, not a target.

10. Only then look at campaign performance, using a clean date range that starts after the fix.

FAQ

Q) 1 Does Consent Mode v2 apply to me if my business is outside Europe?

Yes, if you advertise to people in the EEA or the UK. Google applies it based on where the visitor is, not where you are. A US company selling into Germany needs it.

Q) 2 Do I need a cookie banner just to run Google Ads?

You need one to track and remarket to European visitors lawfully. You can run ads without one, but you'll lose conversion tracking and remarketing for that traffic, which makes the spend very hard to judge.

Q) 3 Will fixing consent mode make my numbers accurate again?

Closer, not exact. You'll get consented conversions plus whatever modelling adds. Accept that reported conversions and real orders will differ, and pick one source of truth for decisions.

Q) 4 Does this affect Meta ads too?

Meta has its own consent requirements in Europe and its own signal loss, so a broken banner hurts there as well. The fix is different, though. Consent Mode v2 is a Google system and doesn't feed Meta anything.

Q) 5 Should I hire someone to fix this?

If nobody in your business can open Tag Assistant and read a consent event, yes. It's usually a few hours of work, not a retainer. If you're weighing up an agency for the wider marketing, our vetting checklist is written around the US market but the questions travel: https://mavixmarketing.com/blog/mavix-digital-marketing-agency-the-leading-seo-agency-in-texas-for-affordable-results-driven-growth 

Conclusion

A conversion drop that arrives without a matching drop in revenue is a measurement story, and in Europe it's usually this one. The fix is cheap compared to the cost of running blind, or worse, cutting a campaign that was actually working. Go and check the Diagnostics tab before you change anything else.

When did you last check what your cookie banner is telling Google?

Get a free 15-minute tracking check

Want someone else to look at it? Get a free 15-minute tracking check at  https://cal.id/mavixmarketing/strategy-call  and we'll tell you whether your consent signals are reaching Google, and what it's costing you if they aren't.