Consent Mode v2: how much data you really lose and what it does to your campaigns

What Consent Mode v2 is — in one sentence
Consent Mode v2 is the mechanism by which your site tells Google's tags what a user has consented to — and Google adjusts its data collection accordingly. Since March 2024, it's effectively mandatory in the European Economic Area for anyone who wants to use remarketing, audience lists, and full conversion measurement in Google Ads and Google Analytics 4. Without it, Google stops accepting EEA traffic data for those features — and that's not a contractual penalty, but a technical cutoff of the signal.
In practice, Consent Mode v2 adds two new consent parameters to the standard setup: ad_user_data (whether user data may be sent to Google for ads) and ad_personalization (whether it may be used for personalization and remarketing). These decide whether a given user feeds your campaigns at all, or disappears from measurement. The rest of this article is about how much you really lose and how to limit it.
How much data do you lose without Consent Mode v2?
Without a correctly implemented Consent Mode v2, you lose exactly as much measurement as the share of users who reject marketing cookies — and in the EEA that's usually 40 to 60% of traffic, depending on the industry and how the consent banner is built. Those users still buy, fill in forms, and call, but their conversions have no way to reach Google, so in the dashboard they look like they never happened. The result: understated conversion volume, overstated cost per acquisition, and Smart Bidding optimizing against an incomplete picture.
That „other half” isn't a rounding error — it's half your budget being judged blind. If you're fighting to make the dashboard numbers match real profit — which we covered in our piece on the marketing dashboard — the consent layer is the foundation the rest of measurement stands on.
Basic or Advanced — the difference and what to choose
There's one difference, but it's decisive: in Basic mode, Google's tags don't fire at all until the user clicks „accept,” while in Advanced mode the tags load immediately and, when there's no consent, send cookieless, anonymous pings. Those pings are exactly what later lets Google model the lost conversions. For most results-driven companies, that settles it in favor of Advanced — Basic is simpler and feels „safer” legally, but it gives up more data that you'll never get back.
| Trait | Basic | Advanced |
|---|---|---|
| When Google tags load | only after consent | immediately, before the decision |
| What it sends without consent | nothing | cookieless pings (no identifiers) |
| Eligibility for conversion modeling | heavily limited | yes, above the volume threshold |
| Impact on remarketing and audiences | larger loss | smaller loss |
| Implementation complexity | simpler | requires proper CMP integration |
If you operate in the EEA and care about Smart Bidding and remarketing, default to Advanced. Basic makes sense where volume is so small that modeling won't trigger anyway, or where the legal team demands a hard „no signal before consent.” But that's a decision about how much measurement you knowingly give up — not a technical detail.
How many conversions does Google's modeling really recover?
Modeling recovers part of the lost conversions, not all of them — and it's worth saying plainly before you believe the marketing slides. Based on cookieless pings and historical data, Google estimates how many conversions „probably” came from unconsented traffic and adds them to reports as modeled conversions. That's a statistical estimate, not counted transactions, so treat it as a good approximation, not gospel.
The key catch: modeling only kicks in above a volume threshold. Google requires a certain number of events and clicks over a few-day window (on the order of hundreds of clicks per week per domain and country) before it will model anything at all. A small account with a dozen-odd conversions a month won't get modeling — it'll just get a hole in the data. That's why the smaller your traffic, the more it matters not to lose a single real conversion to implementation errors.
What Consent Mode v2 does to your Smart Bidding and remarketing
Consent Mode v2 affects campaigns in two places at once: it feeds (or starves) Smart Bidding with conversion data, and it decides how many users reach your remarketing lists at all. With no consent and no Advanced mode, a user won't feed any audience list — so your remarketing audience shrinks by exactly the share of people rejecting cookies. Smaller lists mean more expensive, narrower remarketing, and the bidding algorithm gets a weaker signal to learn from.
It's the same mechanic we described with Performance Max and inflated ROAS: the automation is only as good as the data you feed it. If Smart Bidding sees only half your conversions, it will underbid where it „can't see” the result and reallocate budget toward campaigns that happen to have better-measured traffic — not necessarily the ones that actually earn. Correct consent isn't compliance for peace of mind; it's fuel for the bidding.
Consent Mode v2, server-side, and Enhanced Conversions — how they fit together
Consent Mode v2, server-side tagging, and Enhanced Conversions are three different layers that work best together, not as substitutes. Consent Mode sets what's allowed to be sent; Enhanced Conversions improves the match on the conversions you do have consent for (via hashed data such as email); and server-side tracking moves data collection to your server, where you have more control over what reaches Google and when. Order matters: consent first, then match quality, then infrastructure.
In practice, there's no point investing in a server-side GTM setup or solid GA4 measurement if the consent layer leaks and half your conversions can't count anyway. Seal Consent Mode v2 first, then build the rest on top of it. As a marketing agency, this is usually the layer we start a measurement audit from — because an error here contaminates every number above it.
How to check whether Consent Mode v2 works correctly
You can check Consent Mode v2 in a few minutes by looking at which consent parameters leave the site before and after the banner is accepted. Don't trust that „the plugin is installed” — as many as 6 in 10 implementations we audit look enabled but actually send the wrong signals, or none at all. The list below is the minimum worth running after every implementation and every change to the consent banner.
- In Google Tag Assistant, check that before clicking the banner the tags send a default consent state (denied), and that after acceptance it updates to granted.
- Confirm both new parameters are passed: ad_user_data and ad_personalization — a missing one is the most common Consent Mode v2 error.
- In GA4 (DebugView and reports), check whether behavioral and modeled data appear — their presence signals that Advanced mode and modeling are actually working.
- Verify the consent banner (CMP) is integrated with Consent Mode, not just „nicely displaying” a cookie notice without passing a signal to the tags.
- Verify the default state before consent: it should be denied for all marketing categories, or you risk collecting data with no basis.
FAQ: Consent Mode v2
Q.Is Consent Mode v2 mandatory?
In the EEA it's effectively mandatory since March 2024 for anyone using remarketing, audience lists, and full conversion measurement in Google Ads and GA4. Without it, Google stops accepting EEA traffic data for those features. Formally it's not a penalty, just a technical cutoff of the signal — but the effect is the same: you lose data and reach.
Q.What's the difference between Basic and Advanced mode?
In Basic mode, Google's tags don't fire until the user consents. In Advanced mode, the tags load immediately and, without consent, send cookieless, anonymous pings that let Google model the lost conversions. Advanced loses less data but requires proper integration with the consent banner.
Q.How many conversions does modeling recover?
A portion, not all — and the share depends on account volume and market. Modeling is a statistical estimate based on cookieless pings and historical data, so treat modeled conversions as a good approximation, not counted transactions. Below the volume threshold, modeling doesn't run at all.
Q.Does Consent Mode v2 affect Smart Bidding?
Yes, directly. Smart Bidding learns from conversion data — if a lack of consent means it sees only part of the results, it bids on a truncated picture and reallocates budget toward better-measured traffic, not necessarily the most profitable. Correct consent is fuel for the algorithm, not a mere formality.
Q.How do I check whether Consent Mode v2 works?
Use Google Tag Assistant to see the consent state before accepting the banner (should be denied) and after (granted), and confirm the ad_user_data and ad_personalization parameters are sent. In GA4, check whether modeled data appears. If the banner only displays a notice but doesn't pass a signal to the tags, Consent Mode isn't working, even though it looks enabled.


Builds conversion measurement so the numbers in the dashboard match the money in the bank. Focused on consent, modeling, and measurement that survives cookie loss.
Related articles
Let's talk growth
Send a brief or drop your contact — we'll reply within 24h.
Get a quote