Your Google Ads campaigns look healthy. Tag Assistant shows consent as “granted.” No errors in GTM. The conversion action exists and is enabled. Yet your reported conversions are half of what they should be, and your cost per acquisition keeps climbing for no visible reason.
This is the #1 complaint from EU-market advertisers right now. And in most cases, the cause is not your tracking setup - it is your cookie banner.
What Changed in March 2024
Google began requiring advertisers to send consent signals with all Google Ads requests starting March 2024. This was not optional.
If you serve ads to users in the European Economic Area (EEA) or the UK without implementing Consent Mode v2, Google can:
- Limit your ad serving
- Restrict conversion modeling
- Reduce the accuracy of Smart Bidding signals
The requirement covers two specific consent parameters that advertisers must pass:
ad_storage - Controls whether advertising cookies can be set (including the GCLID cookie that ties clicks to conversions).
ad_user_data - Controls whether user data can be sent to Google for advertising purposes (required for Enhanced Conversions to work).
Without both parameters being properly communicated to Google’s tags, your conversion tracking degrades significantly. Not immediately. Not with a clear error. Just quietly, over time, in ways that are hard to trace.
The “Everything Looks Correct” Problem
The cruelest version of this issue is when your setup passes every diagnostic check - and your conversions are still under-reported by 40-80%.
Here is why this happens.
Tag Assistant and the Google Tag debugger show consent status at the moment you test. When you test your own site, you have likely already accepted cookies (or you are testing in a mode where consent defaults to granted). So the tool reports: consent granted, tag fired, conversion recorded.
What you are not seeing is what happens to the other 90% of real visitors who never interact with your cookie banner at all.
The Default Consent State
Consent Mode works with two possible default states:
| Default State | Behavior |
|---|---|
granted | Tags fire normally before consent is given |
denied | Tags do not fire (or fire in limited mode) until consent is given |
The correct GDPR-compliant setup is to default to denied and only upgrade to granted after a user explicitly accepts.
The problem is what happens when a user never interacts with your cookie banner.
The Cookie Banner Design Problem
This is where most conversion data disappears.
If your cookie banner is a small, unobtrusive bar at the bottom of the screen - easy to scroll past, not blocking content, no bright button - a large percentage of visitors will never click anything. They will just scroll past it and use your site.
Those visitors never grant consent. Their consent state stays at the default (which should be denied). Your Google Ads tags do not fire for them. Their visit is not tracked. Their conversions are invisible.
The Numbers Are Worse Than You Think
When a cookie banner is genuinely non-intrusive - a subtle bottom bar rather than a blocking popup - interaction rates can drop below 10%. That means 90%+ of visitors never grant or deny consent. They just ignore it.
In markets with high organic traffic or where users are privacy-savvy (Germany, Netherlands, Scandinavia), this is the default behavior. People have learned to ignore cookie banners.
The practical result:
| Banner Type | Approximate Interaction Rate | Data Captured |
|---|---|---|
| Full-screen blocking popup | 85-95% | 85-95% of potential |
| Medium overlay with clear buttons | 50-70% | 50-70% of potential |
| Bottom bar, easy to ignore | 5-20% | 5-20% of potential |
| Bottom bar, never clicked | <5% | <5% of potential |
If only 10% of your visitors ever click “Accept” and you are defaulting to denied, you are capturing roughly 10% of your conversion data. The other 90% is gone.
Why This Is Specifically Painful for Google Ads
Google Analytics data loss from Consent Mode is significant. But the impact on Google Ads is more acute because of how Smart Bidding works.
Smart Bidding Needs Conversion Signals
Smart Bidding strategies - Target CPA, Target ROAS, Maximize Conversions - rely on conversion data to optimize bids. The algorithm learns which users, times, devices, and queries are likely to convert, and adjusts bids accordingly.
When you lose 80% of your conversion signals, Smart Bidding is operating with a fraction of the data it needs. The algorithm makes worse decisions. Bids are miscalibrated. Performance declines. And you do not know why.
The Bidding Loop Breaks Down
Here is the compounding problem:
- Consent rates are low → fewer conversions recorded
- Smart Bidding has less data → bids become less accurate
- Inaccurate bids → worse auction performance
- Worse performance → fewer clicks from high-intent users
- Fewer high-intent clicks → even fewer conversions recorded
The decline is self-reinforcing. You may try to fix performance by raising budgets or adjusting targets, but the underlying data problem means the algorithm cannot learn effectively regardless.
GCLID Storage Requires ad_storage Granted
Even if a user converts in the same session as their click, the GCLID (Google Click ID) that connects the click to the conversion must be stored in a cookie. That requires ad_storage to be granted.
If ad_storage is denied at the point of click, the GCLID is not stored. When the user completes a purchase 20 minutes later, Google Ads has no way to attribute that conversion to your campaign.
Consent Mode modeling can recover some of this - but modeling requires a sufficient base of consented users to train against. If your overall consent rate is very low, the modeling is inaccurate or does not run at all.
How to Diagnose the Real Problem
Before you change anything in GTM or Google Ads, confirm what your actual consent rate is.
Check Your CMP Reporting
Most Consent Management Platforms (CMPs) - Cookiebot, OneTrust, Usercentrics, CookieYes - provide consent analytics showing:
- How many users saw the banner
- How many accepted all
- How many rejected
- How many never interacted
If your CMP shows that 70%+ of users “never interacted,” that is your problem.
Check GA4 for Consent State
In GA4, you can see consent state data if you have configured it. Look for events with consent_status set to not_applicable - these are visits where no consent signal was ever received.
Compare Conversion Volume to Server-Side Data
If you have server-side data (e-commerce orders in Shopify, WooCommerce, or a CRM), compare:
- Total orders from your system
- Conversions reported in Google Ads
A gap of more than 20-30% in EU traffic is a strong signal that consent-related data loss is occurring.
Use the Consent Mode Report in Google Ads
Google Ads has a diagnostic for Consent Mode under Tools → Diagnostics → Consent Mode. This shows how many conversion events are arriving with consent granted vs. denied vs. unspecified. If you see a large portion as “denied” or “unspecified,” your banner is the issue.
The Real Fix: Banner Design, Not Tag Configuration
Most guides on fixing Consent Mode focus on the technical implementation: GTM variables, tag sequencing, CMP integration. Those matter. But if your banner interaction rate is 5%, fixing your GTM setup will not recover your data.
The fix is the banner design.
What Makes Users Interact With a Banner
Blocking content until a choice is made. Users who cannot scroll past the banner must make a choice. This is the single biggest factor in interaction rate.
Visual prominence. A banner that occupies a large portion of the screen and uses contrast to stand out gets noticed. A bottom bar in grey blends into the page.
Clear, equally prominent choices. Both “Accept” and “Decline” buttons should be immediately visible. Users who see only an “Accept” button and have to dig for a way to decline often just close the tab.
Plain language. “We use cookies to improve your experience and show you relevant ads. You can choose to accept or decline.” This converts better than legal jargon.
The Legality of Intrusive Banners
GDPR requires that consent be freely given and not coerced. You cannot legally:
- Use a consent wall that blocks access to content entirely
- Make accepting harder to do than declining
- Use dark patterns (pre-ticked boxes, misleading language)
- Make the “Accept” button green and the “Decline” button barely visible
But you can legally:
- Show a prominent banner that occupies significant screen space
- Require the user to make a choice before continuing
- Make both options clearly available with equal visual weight
- Explain why you use tracking and what it is used for
There is a meaningful design space between “invisible bottom bar” and “GDPR dark pattern.” Most sites with low consent rates are sitting in the first category when they could be in the middle.
Targeting a Realistic Consent Rate
By region, realistic achievable consent rates with a well-designed, compliant banner:
| Region | Low (Poor Banner) | Target (Good Banner) |
|---|---|---|
| Germany | 15-30% | 45-60% |
| Netherlands | 15-30% | 40-55% |
| France | 20-35% | 50-65% |
| UK | 30-45% | 55-70% |
| Scandinavia | 15-30% | 40-55% |
| Southern Europe | 25-40% | 55-70% |
Moving from 20% to 50% consent rate in Germany recovers 37.5% more conversion data - the equivalent of more than doubling your observable conversions without changing a single ad.
Technical Checklist: After Fixing the Banner
Once your banner is designed to drive meaningful interaction, verify the technical implementation is correct.
1. Default State Is Set Before Any Tags Fire
In GTM, your CMP integration must set the default consent state before the Google tag initializes. This is often called “consent initialization” and must happen in the first few milliseconds of page load.
If you are using a CMP plugin, check the documentation to confirm it uses gtag('consent', 'default', {...}) before any gtag('config', ...) call.
2. All Four v2 Parameters Are Set
Consent Mode v2 requires four parameters:
gtag('consent', 'default', {
'ad_storage': 'denied',
'analytics_storage': 'denied',
'ad_user_data': 'denied',
'ad_personalization': 'denied'
});
If you are only setting ad_storage and analytics_storage, you are missing ad_user_data and ad_personalization, which affect Enhanced Conversions and remarketing respectively.
3. Update Fires Correctly After Acceptance
When a user accepts, the CMP must call:
gtag('consent', 'update', {
'ad_storage': 'granted',
'analytics_storage': 'granted',
'ad_user_data': 'granted',
'ad_personalization': 'granted'
});
Test this in GTM Preview mode. After clicking “Accept” on the banner, you should see a consent_update event in the preview, and the consent state should show all four parameters as granted.
4. Consent Status Persists Across Sessions
The CMP must store the user’s consent choice (typically in a first-party cookie) and re-apply it on subsequent visits. If a user accepts today but their choice is not remembered next week, consent defaults back to denied on the return visit.
Check that your CMP cookie has an appropriate expiry (typically 6-12 months) and that consent is applied immediately on page load for returning users, before any tags fire.
5. URL Passthrough Is Enabled
For users who deny consent but convert in the same session, URL passthrough preserves the GCLID in the URL rather than in a cookie.
In GTM, set the Google tag configuration variable url_passthrough to true. This extends attribution to within-session conversions even for non-consented users.
6. Enhanced Conversions Are Implemented
Enhanced Conversions use hashed first-party data (email, phone, name) to improve conversion matching without relying on cookies. This works for consented users and helps recover conversions even when attribution via GCLID fails.
Implement Enhanced Conversions in GTM using the conversion linker tag with the provide_signals parameter set to true, and ensure the transaction data layer includes customer information.
What Consent Mode Modeling Can and Cannot Do
After your banner is fixed and your technical implementation is correct, Consent Mode conversion modeling fills in some of the remaining gap.
What Modeling Does
Google observes patterns in your consented user data:
- Which ads, queries, and audiences tend to convert
- How long users typically take between click and conversion
- What device and geographic patterns correlate with conversion
It then applies statistical estimates to non-consented traffic based on those patterns. The modeled conversions appear in your Google Ads reports alongside observed conversions.
What Modeling Cannot Do
Modeling is not tracking. It cannot:
- Recover individual conversion records
- Provide accurate conversion values for non-consented transactions
- Work well if your consent rate is very low (not enough data to model from)
- Provide accurate attribution for remarketing or audience building
A consent rate of 20% does not mean modeling gives you the other 80% back. It gives you an estimate that may be directionally correct but can be off by a significant margin.
Modeling Quality Threshold
Google’s modeling requires a minimum volume of consented conversions to operate accurately. Below a certain threshold, modeling either does not run or produces unreliable estimates.
The exact threshold is not published, but in practice, accounts with fewer than 30-50 observed conversions per month per consent state often see poor modeling quality. For accounts with very low traffic, improving your consent rate is especially critical because modeling may not compensate at all.
A Common Misdiagnosis: Blaming the Tracking Setup
When conversions drop, the natural instinct is to investigate the tracking implementation. The GTM container. The conversion tag triggers. The data layer. The Conversion Linker.
For Consent Mode problems, this investigation will usually come up empty. The tags are correct. The triggers are correct. The Conversion Linker is firing. Everything looks fine - because it is fine, for the users you are able to observe.
The users who never interact with your cookie banner are invisible to the diagnostic tools. You cannot debug a tag that never fired because consent was never given.
The diagnostic question is not “why isn’t my tag firing?” It is “why aren’t my users granting consent?”
That shifts the investigation from GTM to your consent banner, your CMP configuration, and the user experience around the consent decision. Different tools, different team, different fix.
Prioritizing the Fix
Not every site will see the same impact. Prioritize fixing your consent setup if:
You have significant EU or UK traffic. Consent Mode requirements are legally enforced in the EEA and UK. If 30%+ of your traffic comes from these regions, the problem compounds quickly.
Your Google Ads account uses Smart Bidding. Target ROAS, Target CPA, and Maximize Conversions all degrade with reduced conversion signals. Manual CPC campaigns are less affected (though attribution is still broken).
You sell products with longer consideration cycles. If users typically click your ad and convert within 1-3 days (rather than immediately), multi-session attribution is critical. Losing GCLID storage on the first visit means losing most conversions.
Your conversion volume feels lower than business results. If your actual orders (from Shopify, your CRM, etc.) significantly exceed what Google Ads is reporting, data loss is likely the cause.
Key Takeaway
Since March 2024, Consent Mode compliance is mandatory for EU and UK ad serving. But compliance is not just about having a cookie banner. It is about having a banner that users actually interact with.
A barely visible bottom bar that 90% of visitors scroll past is technically compliant but practically equivalent to no consent collection at all. The result is the same: tags do not fire, conversions are not recorded, and Smart Bidding operates with a tiny fraction of the data it needs.
The fix starts with the banner. Make it prominent enough that users notice it. Make the choices clear enough that they make one. Then verify that the technical implementation correctly upgrades consent state when they accept.
Closing that gap - from 10% interaction to 50% interaction - can recover more conversion data than any GTM fix or tracking architecture change.
For EU-market advertisers, the consent banner is not a legal checkbox. It is the gatekeeper between your ad spend and your attribution data.
Related Posts
How to Implement Consent Mode v2 with Google Tag Manager
14 min read
How Consent Mode Impacts Your Tracking: What Changes When Users Decline
12 min read
GA4 and Consent Mode v2 - How to Keep Tracking Without Killing Data Quality
9 min read
Need Help With Your Google Ads?
I help e-commerce brands scale profitably with data-driven PPC strategies.
Get In Touch