Consent Mode basic vs advanced: which conversions you lose and what gets modelled

Consent Mode is a Google mechanism: the consent banner passes the visitor's choices to Google tags, which then change how much data they send. Google's Tag Manager help describes two ways to implement it, basic and advanced¹.

The first row is the one to remember. Under basic, Google hears nothing from a visitor who has not consented. Under advanced, it hears a stripped-down signal. Which setup is allowed for your site and markets is a decision for your counsel and your consent management platform (CMP).

Under basic, you lose measurement for everyone who declines or ignores the banner. Their orders still happen. Your tags just do not see them. Google Ads and GA4 show fewer conversions than your shop does. Google's help says conversion modelling in Ads then relies on a general model, not one built on your account¹.

Under advanced, those visitors are not invisible. Cookieless pings give Google something to work with, and Google describes advanced as giving better, advertiser-specific modelling¹. Modelled conversions are estimates. They are not recorded orders.

Two things stay true under both modes:

• Your order system still has the real figure. Use it as the reference.
• Visitors who consent are tracked as before. If few consent, you see a thin slice of reality.

If you want to widen what Google can match, enhanced conversions send hashed first-party data, such as an email address, to Google. That is a separate feature, and it still needs the right consent basis. For the full picture of what a tracking pixel misses, and how server-side tracking fits in, see those pages.

Google publishes eligibility conditions for behavioural modelling in GA4²:

• Consent Mode runs on every page, and Google tags load before the banner, whether or not the visitor consents (the advanced setup)².
• The property collects at least 1,000 events a day with analytics storage denied, for at least 7 days².
• The property has at least 1,000 daily users with analytics storage granted, on at least 7 of the previous 28 days².

Google adds that meeting the thresholds does not guarantee eligibility, and that training can take longer than 7 days². The page does not describe a basic implementation.

What follows:

• A smaller shop can fail the thresholds, and then GA4 does not model the gap.
• If you are on basic, GA4 behavioural modelling does not apply, because tags do not load before consent.
• Check the property's own status in GA4 rather than assuming it qualifies.

Our view: if you are close to the thresholds, treat any modelled figure as a bonus, not a plan.

This is the question behind searches like "consent mode v2 conversions dropped" and "Google Ads conversions dropped after consent mode". There are two ordinary causes.

1. Honest loss. You switched to basic, or consent rates are low, so unconsented orders are no longer counted.
2. Broken signals. The banner and the tags are not talking to each other. A PPC Land article retold one agency founder's account of a client whose banner looked fine but did not pass consent signals to Google tags, and reported Google Ads conversions fell sharply overnight³. It is one account, not a norm.

The second one is a bug, not a loss. It is also the one you can fix. After every banner, CMP or tag change, check three things: the signals fire, they carry the right values, and the conversion count compares sensibly with your order system.

Google's EU user consent policy page names two consent signals: one for sending user data to Google for advertising, and one for ad personalisation⁴. Check that your banner passes both. What your consent obligations are is for your counsel to decide.

Bidding follows the data it can see. If conversions fall because of consent, automated bidding sees a weaker signal. That can change spend before anyone looks at the cause.

In reporting:

• Label modelled conversions as estimates.
• Compare platform conversions with order-system orders, week by week.
• Read a drop in platform conversions next to total revenue before cutting budget. A tracking change is not always a demand change.

Our view: do not compare periods before and after a consent change without a note on the change. You will otherwise credit or blame the ads for a banner.

CMP vendors explain the settings in their help pages, for example one CMP vendor's article on advanced and basic Consent Mode v2⁵. The effect on your bidding is yours to check. Ask:

• Which mode are we on, and who decided?
• Do the consent signals reach every tag, including the server side?
• What share of visitors consent, by market?
• Do we meet the modelling thresholds, and how do we know?
• After the last change, did anyone compare platform conversions with orders?