Consent mode conversion drop: how to prove the banner caused it

Why do conversions drop after a consent mode or cookie banner change?

Conversions drop after a consent change because fewer visitors can be measured with cookies, not because fewer people buy. When a visitor declines, Google tags stop reading and writing cookies, so a conversion that used to be tied to an ad click is no longer observed directly.

How big the gap is depends on how the banner is set up and on how many people decline. Google notes that consented users are typically 2–5x more likely to convert than unconsented users, so even the conversions you can still observe are not a fair sample of everyone who visits.

The fingerprint to look for: the drop starts on the day the banner or consent configuration went live, it hits every campaign that sends traffic to that site, and it is larger in countries where the banner is shown.

What is the difference between basic and advanced consent mode?

Basic consent mode blocks Google tags until the visitor answers the banner; advanced consent mode loads the tags straight away and sends cookieless pings when consent is denied. That difference decides how much of the lost data Google can model back.

Consent mode works with four signals that matter for measurement and ads: ad_storage, analytics_storage, ad_user_data and ad_personalization. Your banner or consent management platform (CMP) sends a default state when the page loads and an update when the visitor makes a choice. If someone on the web team moves from one mode to the other, or swaps the CMP, log it as a tracking change: it can move reported conversions without any change in real sales.

How do I check if consent mode caused my conversion drop?

Check three things in order: the date of the consent change, the gap between clicks and conversions, and the consent status Google reports for your conversion actions. If all three line up, the banner is your prime suspect.

1. Date the consent change. Ask the web team or the CMP admin for the exact day and time the banner, its wording, its buttons or its consent mode settings went live. Check Google Tag Manager version history if the CMP is deployed there.
2. Compare clicks with conversions. If clicks and sessions are stable and only conversions fell, a measurement change is more likely than a demand change. If clicks also fell, look at bids, budgets and the auction first.
3. Open the consent status in Google Ads. Click the Goals icon, open the Conversions dropdown, click Summary, select a website conversion action and open the Diagnostics tab. You will see either "Consent mode is implemented" or "Consent mode is implemented and modeling is active".
4. Test the banner yourself. Use Tag Assistant: its Consent tab is empty when consent mode is not implemented or the Google tag is blocked. Check that a default is set before tags fire and that it updates when you accept.
5. Compare with your CRM or back end. Orders and qualified leads in your own system do not depend on cookies. If they are flat while Google Ads conversions fell, you have a measurement gap, not a sales problem.

The same logic applies to GA4: see how to annotate GA4 when something changes so the date of the banner change sits next to the chart, and how to tell a GA4 tracking problem from a real loss of visitors.

Why are modeled conversions not showing up yet?

Modeled conversions only appear after your account crosses Google's data thresholds, and the status can take days to update. A drop in the first week after a consent change is normal even when the setup is correct.

Google Ads

• Conversion modeling needs a correct consent mode (or IAB TCF v2.0) setup and a threshold of 700 ad clicks over a 7-day period, per country and domain grouping.
• The active status can take 48 hours to appear in Google Ads, and up to 2 weeks in some cases.
• Modeled conversions appear in the normal Conversions column, at the same granularity as observed conversions.
• Impact results (modeled conversions divided by observed conversions) can appear 7 days after you implement consent mode, are shown for 4 weeks from the start of modeling, refresh weekly and are not reported when uplift is below 1%.

GA4

• Behavioral modeling needs at least 1,000 events per day with analytics_storage denied for at least 7 days, and at least 1,000 daily users with it granted for at least 7 of the previous 28 days.
• Training the model can take more than 7 days after you meet those thresholds.
• Modeled data only shows when Admin > Reporting identity is set to Blended. Reports then carry a data-quality message such as "Including estimated user data".

Small sites often never reach these thresholds. In that case the drop in reported conversions is permanent, and your targets, bid strategies and client reports have to be reset to the new baseline.

How do I isolate the cause when several things changed at once?

Put every change from the days before the drop on one timeline, rule out tracking first, then test one hypothesis at a time. Consent changes rarely ship alone: they often go out with a site release, a new tag or a CMP upgrade.

1. Build the timeline. List your team's edits (bids, budgets, ads, audiences), platform automations (auto-applied recommendations, bid strategy learning), site, tracking and CRM changes (banner, CMP, tags, forms) and external events (holidays, competitor promotions). Pull the ad account side from Google Ads change history.
2. Rule out tracking first. If the banner went live the same day as a new form or a GTM container version, check that the conversion tag still fires on an accepted session before you blame consent.
3. Use regions as a control. Many sites only show the banner, or only deny by default, in the EEA and the UK. If conversions fell in those countries but not elsewhere, consent is the likely cause. If they fell everywhere, look for a change that affected every visitor.
4. Check the size of the gap. Compare the drop with your accept rate if your CMP reports it. A conversion drop at or below your decline rate is consistent with consent, since people who decline tend to convert less. A drop far larger than the decline rate points to a broken tag.
5. Change one thing and wait. If you fix the default or switch to advanced mode, do it alone and give it at least 7 full days before judging.

Example: an agency sees Google Ads conversions for a client fall 35% from 12 March, with clicks flat. The timeline shows a new CMP went live on 11 March, EEA traffic only, while non-EEA conversions are stable and CRM orders are unchanged. The agency logs the banner change, resets the reporting baseline and checks the Diagnostics tab weekly until modeling becomes active.

For the client conversation, the guide on how to explain a conversion drop to a client has a structure you can reuse. For a wider checklist of causes, see why conversions drop.

What should I log when the consent setup changes?

Log every consent change with the same detail you would give a bid strategy change, because it moves the numbers in the same way. Six fields are enough to answer "what changed" months later.

• Date and time live, not the day the ticket was opened.
• What changed: banner design, button wording, CMP vendor, consent mode type (basic or advanced), default states per region.
• Where: domains, subdomains, countries and apps affected.
• Owner: who made the change and who approved it (often legal).
• Expected effect: "observed conversions down in EEA; modeling may fill part of the gap after the thresholds are met".
• Follow-up date: when you will check the Diagnostics tab and GA4 reporting identity again.

In Changeline you can log this as a tracking change on the client's project timeline, next to the Google Ads and Meta Ads edits from the same week, so the next person who sees the drop finds the banner change before blaming the campaigns.