Where can I see if GA4 receives consent signals?

In Admin, under Data collection and modification, open Consent settings. It shows the EEA's share of your traffic and conversions, and flags problems with the advertising and behaviour signals per data stream. Google says the alerts can take 48 to 72 hours to update.

Is it safe to switch GA4's reporting identity?

Yes. Google says the option you choose does not affect data collection or processing, and you can switch back at any time. You need the Editor role. Switching from Blended to Observed simply hides the modeled users, which makes it a quick way to size them.

How long does GA4 take to start modeling after I fix consent mode?

At least a week, often longer. Google wants 7 days of enough declined and accepted traffic within a 28-day window, and says training can take longer still. Estimates then run only from the date the property qualified, not backwards.

How to check GA4 data after cookie consent, step by step

Find where your cookie banner shows, confirm consent mode is on, and see whether GA4's behavioral modeling is active and how much it adds. Then compare GA4's purchases with Shopify's orders in banner and non-banner countries. The gap between the two regions is roughly what consent costs your data.

By Joris van Huët, Founder & CEOUpdated 2026-10-018 min read

Usually nine checks over one closed month. Find where your banner shows, confirm consent mode is on, and see whether GA4's modeling is active and how much it adds. Then compare GA4's purchases with Shopify's orders in banner and non-banner countries. The difference between the two regions is roughly what consent costs your data.

You need a closed month, the Editor role in GA4 and access to Shopify's reports. Steps 1 to 4 check the setup, 5 to 7 check the modeling, and 8 and 9 measure the cost. The menu paths come from Google's and Shopify's help pages as they read on 1 October 2026.

Step by step

List the countries that see a banner. Shopify's automated settings show its banner in the UK and EEA when those markets are active. In those regions, non-essential data waits for consent. Menu path: Shopify admin > Settings > Customer privacy > Cookie banner > Regions and content.
Confirm consent mode is on. Google says consent mode switches on automatically once Shopify's cookie banner is active. A third-party banner must pass consent to Shopify through the Customer Privacy API. Google says the status Consent mode is implemented can take 48 hours, or up to 2 weeks, to appear. Menu path: Google Ads > Goals > Conversions > Summary > (website conversion action) > Diagnostics.
Check what GA4 receives. The Consent settings page opens with the share of your traffic and conversions that comes from the EEA. Select your web stream to see alerts on the advertising and behaviour consent signals. Menu path: GA4 > Admin > Data collection and modification > Consent settings > (web stream).
Find out whether tags wait for consent. In basic mode, tags stay blocked until a visitor accepts, so a decliner sends nothing. In advanced mode, tags load with consent denied and send the cookieless pings that modeling needs. On Shopify, pixels in banner regions run only with the permissions they require, and Pixel Helper says Pixel is awaiting consent while one waits. Menu path: Shopify admin > Settings > Customer events > (Google & YouTube pixel) > View customer privacy, then Test.
See whether modeling is switched on. GA4 turns behavioral modeling on by itself once a property qualifies, and the Blended identity then lists it. Google asks for at least 1,000 events a day from decliners for 7 days. It also asks for 1,000 daily users who accepted, on 7 of the previous 28 days. Menu path: GA4 > Admin > Data display > Reporting identity.
Read the data-quality icon. Including estimated user data means estimates are in. Excluding estimated user data means your identity setting hides them. Estimated user data unavailable can mean the property does not qualify, or the dates fall before it did. Menu path: GA4 > Reports > Acquisition > Traffic acquisition > data-quality icon.
Measure how much is estimated. Note the month's users under Blended, switch to Observed, note them again, then switch back. Google says the choice changes nothing in collection or processing. The gap is the share GA4 estimated. Menu path: GA4 > Admin > Data display > Reporting identity > Observed > Save.
Compare purchases with orders by country. Build a free-form exploration with Country as rows and Ecommerce purchases as values for the month. In Shopify, open Total sales by billing location for the same dates. Divide purchases by orders for banner and non-banner countries. Menu path: GA4 > Explore > Free form; Shopify admin > Analytics > Reports > Category filter > Sales.
Compare Direct by country. Nest Session default channel group under Country in the same exploration, and add Purchase revenue as a value. Google says modeled key events go to Direct when traffic is too thin to model them. Menu path: GA4 > Explore > (your exploration) > Rows.

A worked example

For illustration, here is one invented month in round numbers.

Line	Banner countries	Other countries
Shopify orders, Total sales by billing location	400	600
GA4 Ecommerce purchases, free-form exploration	260	570
Share of orders GA4 saw	65%	95%
Direct's share of GA4 purchase revenue	50%	35%

Say the shop shows 400 orders from banner countries and 600 from the rest. If the exploration shows 260 and 570 purchases, GA4 saw 65% of the first and 95% of the second. Ad blockers and definitions hit both regions, so most of that thirty-point gap is likely the banner's doing.

Suppose Direct also takes 50% of GA4's purchase revenue in banner countries against 35% elsewhere. That is the pattern Google's help predicts: key events it could not model, filed under Direct. If step 7 also adds users under Blended, those users are estimates, so trends hold and counts go soft.

One store's export shows what those two symptoms look like when everything sits in one file, with no country split. Direct holds 57.7% of revenue in last click, first click and touched views alike (Channels sheet). Journeys with 4 to 9 touches took 16.9 days to buy (Journeys sheet), the kind of path one declined visit can break. Steps 8 and 9 tell you how much of a pattern like that is consent.

What should I check when the numbers look wrong?

Users jumped on one date. That is often the day modeling started, because reports with estimates show higher user counts. Compare periods on the same side of that date.
first_visit and session_start look inflated. GA4 cannot tie a decliner's events together, so it counts those events on every page they load. In reports, modeling leaves event counts alone.
A path or funnel exploration disagrees with a report. Those explorations model first_visit and session_start, so their counts come out lower.
Audiences look small. Audiences never include modeled users, so in banner countries they only hold people who accepted.
Countries do not line up. GA4's Country is where the activity came from, while Shopify uses the billing address. Gifts and travellers blur both, so read the gap between regions, not single orders. Definition gaps are covered in reconciling GA4 and Shopify revenue.
Google Ads says consent mode is implemented, but nothing is modeled. Google's bar is 700 ad clicks over 7 days for each domain and country. Smaller countries may never clear it.
Tag Assistant cannot connect until you accept. Google says that can mean your tag is blocked before consent, the basic setup. Decliners then send GA4 nothing to model.

What to do this week

Decline the banner yourself. Open your store in Tag Assistant with your location set to a banner country, decline, and open the Consent tab. Pass: the On-page Default column reads Denied and changes only when you choose. Fail: tags run as granted before you choose, which is a compliance problem before it is a data one.
Read Google Ads' modeling status by country. Open the Diagnostics tab of the purchase conversion action from step 2. Pass: it says modeling is active and shows uplift for your main countries. Fail: it only says Consent mode is implemented, so Google Ads is not modeling your decliners yet.
Repeat steps 8 and 9 next month. Keep both regions' shares in one sheet, month after month. Pass: the gap between regions holds steady. Fail: it widens after a theme, banner or app change, so check that change first.

Check the homework. Your GA4 Attribution paths export already holds the evidence. Causality Engine reads that one file and shows what each channel caused next to what last-click gave it, in 1 to 2 minutes, for €99 once (excluding VAT), refundable within 30 days. Check the homework

Sources, 1 October 2026: Configuring customer privacy settings (Shopify Help Center); Set up Shopify to obtain user consent (Tag Manager Help); Verify consent mode implementation (Tag Manager Help); [[GA4] Verify and update consent settings in Google Analytics](https://support.google.com/analytics/answer/14275483?hl=en) (Google Analytics Help); About consent mode (Google Ads Help); Pixels overview (Shopify Help Center); App pixels (Shopify Help Center); Troubleshoot consent mode with Tag Assistant (Google for Developers); [GA4] Behavioral modeling for consent mode (Google Analytics Help); Traffic acquisition report (Google Analytics Help); Reporting identity (Google Analytics Help); [GA4] Free-form exploration (Google Analytics Help); Analytics dimensions and metrics (Google Analytics Help); Sales reports (Shopify Help Center); [GA4] About modeled key events (Google Analytics Help); About consent mode impact results (Google Ads Help).

Frequently asked questions

Where can I see if GA4 receives consent signals?
In Admin, under Data collection and modification, open Consent settings. It shows the EEA's share of your traffic and conversions, and flags problems with the advertising and behaviour signals per data stream. Google says the alerts can take 48 to 72 hours to update.
Is it safe to switch GA4's reporting identity?
Yes. Google says the option you choose does not affect data collection or processing, and you can switch back at any time. You need the Editor role. Switching from Blended to Observed simply hides the modeled users, which makes it a quick way to size them.
How long does GA4 take to start modeling after I fix consent mode?
At least a week, often longer. Google wants 7 days of enough declined and accepted traffic within a 28-day window, and says training can take longer still. Estimates then run only from the date the property qualified, not backwards.

Go deeper: Causal attribution, explained.

Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.

Keep reading

Terms in this article

Browse the full glossary