hi, i’m amr. when a client’s GA4 sessions drop after a cookie banner goes live, the first question is usually “where did the traffic go?” the honest answer depends on a threshold most teams have never looked up.
what google actually requires before it models the gap
with consent mode, visitors who decline analytics cookies don’t just vanish from GA4. google can fill part of the gap with behavioral modeling, estimating what those visitors did based on visitors who did consent. but the modeling is conditional. according to google’s own help documentation, the property needs:
- at least 1,000 events per day with analytics_storage set to “denied”, for at least 7 days
- at least 1,000 daily users sending events with analytics_storage “granted”, for at least 7 of the previous 28 days
google adds that more than 7 qualifying days may be needed, and that meeting the thresholds doesn’t guarantee eligibility. the model also weighs factors such as the ratio of new to returning users and the user-to-session ratio, which the page doesn’t quantify.
why this matters more for smaller sites
read the two thresholds as a volume gate on both sides: you need enough denied traffic for the model to learn from, and enough granted traffic to learn with. a B2B site that gets a few hundred visits a day can fail both without anyone noticing. my reading, not google’s wording: if the property is below the gate, there is nothing modeled to fill the gap, so the visitors who declined stay unmeasured in your reports.
that changes how you read a post-banner drop. on a low-volume site it’s not a reporting glitch you can wait out. it’s a permanent blind spot, and it lands unevenly: whichever segment declines most (often a specific country or device mix) is the segment your channel reports undercount.
a quick check before you rebuild any dashboard
- count your daily denied-consent events and granted-consent users against the two thresholds above
- compare GA4 sessions from paid channels with the clicks reported in Meta Ads and Google Ads for the same days. a widening gap after the banner went live is your undercount
- compare GA4 lead counts with form submissions in HubSpot or your CRM. leads are first-party records, so they don’t depend on the cookie decision
- report CRM leads and ad-platform conversions next to GA4, not instead of it
the point isn’t to avoid the banner. consent rules exist for a reason, and i’m not suggesting you work around them. the point is to know which part of your numbers is measured, which part is modeled, and which part is simply missing.
where this stops applying
if you have no consent banner, or your traffic is high enough to clear both thresholds comfortably, modeling may already be filling the gap and this is a smaller issue. modeled data is also an estimate, not a count, so even a qualifying property shouldn’t treat it as exact.
start scaling — if your attribution numbers moved after a consent banner and nobody can say why, a one-hour gap check usually finds it. let’s connect.
Sources: Google Analytics Help, “[GA4] Behavioral modeling for consent mode” (support.google.com/tagmanager/answer/11161109). Figures are published product documentation, not verified results from Amr’s own client accounts.