Consent Mode v2: User consent and data that doesn't disappear
The Consent Mode v2 setting is about whether the tags obey the human choice, not about having a window with a button. The site announces the consent state even before the counters are loaded, and a click in the banner updates this state on the fly, without reloading the page. Four signals — ad_storage, analytics_storage, ad_user_data, ad_personalization — must be denied before the tag manager snippet hits the page. Then you have to do the same with chat, maps and call tracking: they are not managed by Google.
four consent signals and the behavior of each system in both pre-choice and post-choice states
after a free audit
Price
30
Guarantee
days after project sign-off
four consent signals and the behavior of each system in both pre-choice and post-choice states
What are we setting up?
3-12
Term of works
working days after free inspection; counts the number of systems, not the size of the site
free verification of what is loaded before clicking: 2-3 working days
Before the estimate
at the plastic surgery clinic, the consent mode was set correctly, and the tag container behind it was not loaded at all - measurement 07/31/2026
A typical find
in the tactical store there are six of them: tag container, analytics, advertising, call tracking, mailing, Google login
How many systems usually fall under
legal assessments and rejection bypass: we close the technical part, the legal part is provided by your lawyer
What we don't do
Who it's for
Situations where this service delivers results
Scenario 1 of 5
You work in European markets or plan advertising there
From 2024, Google's advertising systems expect consent signals for traffic from the EEA. When the signal does not come, restrictions are turned on quietly: remarketing audiences stop being filled, part of the measurement in this area is lost. Warnings in the cabinet are easy to miss.
We'll review your situation in a free audit
Scenario 2 of 5
Banner put the plugin in and no one checked what it was doing
The most frequent configuration: the window appears, the button is pressed, and the scripts worked even before that. There is a formal banner. In fact, the behavior has not changed a single step, and you can see it in half an hour - in the network tab you can see which requests went to the first click.
Scenario 3 of 5
A lot of third-party widgets live on the site
Chat, road map, call tracking, video insert, social network buttons. Everyone puts their own on the device, and Google's consent mode doesn't control them — each requires a separate delayed launch. This is the biggest part of the work and the reason why the term is counted in systems and not in pages.
Scenario 4 of 5
Are you going to start analytics or advertising from scratch
Then it is cheaper to make the consent first. The "counters first, then banner" order costs a lot of work: tags have to be pushed to a state of agreement when they are already scattered across the theme, modules, and container, and then test each one again.
Scenario 5 of 5
The data dropped after installing the banner, and it's unclear why
We analyze what exactly happened: people opt out, tags do not update their status after selection, or the banner blocks the counter forever. Three different causes with the same symptom reported and treated differently.
What's included
Complete list of work and what you get as a result
A consent banner with categories and a link to change the choice in the footer - a person can change their mind without a letter of support
The default denied state declared above for the container snippet: tags are limited to selection, not "almost limited"
Signaling to the tag manager and on to analytics and ad counters — selection goes where the collection decision is made
Update signals at the moment of click without reloading the page - the first action of the user is not lost
Making third-party systems — chats, maps, call tracking, video embeds — dependent on consent: they are the ones that are most often forgotten
Limited collection mode where supported: on failure, the system receives an impersonal signal instead of nothing
Run each scenario in preview mode: agree, decline, partially agree, change decision
Comparison of the text of the privacy policy with the actual behavior of the site - the discrepancy between them is closed before launch
A document with a diagram: which system belongs to which category and what to do when a new one appears
Guarantee of 30 calendar days for completed works
When this service isn't right
What's not included — so there are no surprises at delivery
Legal expertise and drafting of a privacy policy is the work of a lawyer
Guarantee of compliance with specific legislation - we implement the technical part
Consent Management Platform subscription fee
Conducting advertising campaigns
Bypassing user opt-out in any form
Free tag compliance and behavior checker
The consent banner on the site does not mean anything yet: what matters is whether the tags are really waiting for the user to choose and what happens before the click. We are checking the behavior, not the presence of the window.
What we measure
Is there a banner and what does it actually doA common picture: the window is there, and all scripts are loaded at once, regardless of the selection. This is worse than not having a banner because it creates a false sense of security.
Default state of consent signalsOn one of the sites we measured, the consent mode was set to default denied — that is, tags are limited until the user chooses. This is correct behavior, and it does not happen often.
Which systems must obey consentAnalytics, advertising pixels, chats, call tracking, maps, video inserts. We list everything that puts files on the user's device.
Are signals updated after selectionConsent must be transferred to tags not only at the start, but also at the moment of clicking - otherwise the user's first action will remain unrecorded.
What happens to the data upon failureSome systems support a limited collection mode without identifiers. Let's see if it is used or if everything just turns off.
Privacy policy on the siteDoes it describe what is actually going on? The discrepancy between text and behavior is a separate risk.
What you get
A description of the actual behavior of the site before and after the user's selection.
List of systems that currently ignore consent.
Assessment of the expected impact on the volume of data after correct implementation.
Conversation for 30 minutes on the document.
Timeline: 2-3 working days
Why is it free
Because the check takes two or three days, and the situation "there is a banner, but the behavior has not changed" occurs so often that it would be strange to check it for money.
What's next
We provide an estimate for the implementation: banner, transmission of signals to tags, transfer of systems to consent, verification.
Short form: your contact and site URL
Contract, act and 30-day warranty
Every project gets a written contract: scope, deadlines, amount, acceptance procedure. After delivery — act and invoice, then 30 calendar days of warranty.
Sole proprietor & bank transfer
The contractor is a registered sole proprietor. Payment by invoice with closing documents.
Rights & access — yours
Code, design and materials transfer to you after full payment. Domain, hosting, repository and analytics are registered to you.
Client portal instead of email chains
During the project you get access to a portal: contracts, invoices, acts and project status in one place.
European clients
Among our work — projects for Norway, Bulgaria, Moldova and Spain.
Verifiable numbers
Every case in the portfolio comes with a link to a live site and a technical measurement.
Audit first, then pricing
There is no price list on the site intentionally: the scope of the same work differs multiples between clients.
We say "no" when unsure
If the task isn't ours or the deadline is unrealistic — we tell you upfront.
What affects the price
Why two seemingly identical tasks are priced differently
How many systems obey the choiceOne counter and banner — several days of work. Analytics, two ad pixels, chat, call tracking, and map—six separate translations, each with its own method of delayed launch. This is where the difference between the lower and upper bounds of the estimate lies.
Where does the banner itself come from?The consent management platform takes care of texts, categories, translations and the consent log — and the subscription fee. Your own banner is cheaper to own and easier to upload, but you are responsible for the magazine and updating the texts for changes in the regulations.
Collection mode to chooseAdvanced requires more testing: tags are loaded, but behave differently, and both states need to be tested separately. The basic one is set up faster, but then the data is noticeably less, and you need to be prepared for this in advance.
Number of markets and language versionsRequirements vary by region, so the banner often needs to behave differently depending on the visitor's country. Add text translations and a separate script check on each language version.
Who edits the codeWhat is done in the tag container, we do ourselves and quickly. Widgets sewn into a theme or a frontend assembly are contributed by whoever has access to the repository: then the deadline begins to depend on his queue, and not on ours.
Process steps
Transparent stages with approval at every step
Total duration:3–12 days
1
Free tag behavior checker
2-3 working days, to estimate
Let's see what loads before the first click, what state the default signals are in, and which systems ignore the selection altogether. The output is a description of actual behavior and a list of systems to be translated.
2
Default categories, banner texts, and status
1-3 working days
We record which categories a person sees and what he falls into. We put the denied declaration above the container snippet — the order of the lines in the markup here is more important than any setting in the interface.
3
Transferring systems to consent signals
1-5 working days
We hang Google tags on the state of consent, third-party widgets on delayed launch in the same category. Chat, map and call tracking appear after selection, not with the page.
4
Running scenarios and reconciliation with politics
1-4 working days
Preview mode, four scenarios: consent, refusal, partial consent, change of decision through footer. In parallel, we compare the text of the privacy policy with what the site actually does.
5
A week of observation of the share of consent
7 calendar days, does not take working days
Let's see what share of sessions gives consent and how the amount of data in the reports has changed. It is with this figure that all subsequent periods are then compared, so it should be removed immediately, and not after a quarter.
Did not find your case?
Describe how it works on your side — we will tell you whether “Consent Mode v2” fits and what it means in your situation. No brief and no call: one question, one answer.
Technologies & integrations
What we build on and what it connects to
Stack
Consent Mode v2 — four signals: for analytics, advertising, user data and personalization. Manages Google tags and only them.
Advanced mode - tags are loaded before selection and send impersonal signals without identifiers. There is material for modeling, but not every lawyer accepts it.
Basic mode - nothing is loaded until consent. Legally the simplest, but there is nothing to model.
Google Tag Manager — changes signals without site release. One condition: the declaration of the default state is above the container snippet itself in the markup.
Consent management platforms — texts, categories, consent log. They pay a subscription fee and add their script to the critical path of loading the page.
The preview mode is the only place where you can see the state of the signals at the moment when a particular tag is triggered. This is not visible on the live site.
Integrations
GA4
Google Ads
Meta Pixel
call tracking
chat widgets
maps and video inserts
consent management platform
Display-only banner vs consent that manages tags
How this option differs from the alternative
Banner without signalingOur approach
What happens before you clickall scripts worked before the window appearedtags are either loaded in the denied state or not loaded at all — depending on the selected mode
What gives refusaldoes not change anything - data is collected furthercollection is stopped; where supported, leaves an impersonal signal without identifiers
Third party widgetslive a separate life, the banner does not see themchat, map and call tracking start after selection
Data volume in reportsremains the same, so it looks betteris decreasing, and we show how much more there is to work
What to show during the inspectionwindow screenshotcategory schema, selection history, and reproducible behavior
Cases
Tasks and results in numbers — all metrics measured by us
plastic surgery clinic, a site on the JS framework — our technical dimension
Task
Understand why a configured consent did not produce any data in the reports.
Solution
Read page markup, inline tag container snippet, and default consent state.
Result
The consent mode v2 is correct: default denied, the status is declared before the tags. The other thing didn't work - in the inline snippet, the exit condition compared the identifier with itself, so the return always worked and the container was never loaded; the ID field contained the GA4 code, not the container. Consent governed what is not. The rest of the site is done neatly: the bot is given 126 KB of pre-rendered HTML in 145 ms, ten types of structured data. Measured on 07/31/2026.
Calculate how many systems on the site will have to obey the user's choice.
Solution
Analyzed page markup and listed everything that puts identifiers on the device or transmits data to the outside.
Result
Six systems at the same time: tag container, analytics, advertising counter, call tracking, mailing service and login through Google account. Plus custom dataLayer events in checkout. Each is transferred to consent separately - there is no common switch for all six. The size of the site does not play a role: 54,302 products, the first response is 413.8 ms. Measured on 07/31/2026.
a brand of natural spices with its own online store is our technical dimension
Task
Check that the website collects data about the visitor.
Solution
Reconciled counter IDs in the markup with what those counters actually do.
Result
Three collection systems at the same time: live GA4, Universal Analytics, which Google turned off on July 1, 2023, and Yandex.Metrika. A dead meter still loads and writes its message to the device, so it is subject to consent on an equal basis with a live one. There is another defect nearby: the site map has 3,971 entries for 590 addresses. Measured on 07/31/2026.
the online shop for the delivery of sweets and snacks is our technical dimension
Task
Find out whether there is something to conquer consent on the site.
Solution
Search for all counters and external scripts in the markup, check the sitemap and PHP version.
Result
There are no collection systems: no tag container, no Google counter, no Meta pixel. The banner of agreement here would rule the void — first you have to start measuring at all. Nearby: Sitemap gives HTTP 200 with zero byte body and PHP 7.3.33 not supported as of December 2021. Measured on 07/31/2026.
What we need from you
We can't start without this — best to prepare in advance
1Access to the site or to the tag manager - without it, only external behavior is visible.
2A list of everything that puts files on the device: analytics, advertising, chat, call tracking, maps, video inserts, social media buttons.
3Deciding on the categories of consent and banner text is your choice with your lawyer, not ours.
4Existing privacy policy or agreement to prepare it with a lawyer before launch.
5List of markets served by the site: Requirements vary by region and the banner often needs to behave differently.
6A person on your side who makes decisions on texts and categories.
If something is missing — let us know, we'll help you gather it or do it as a separate task.
FAQ
Most frequently asked questions — with concrete answers
We already have a cookie banner. Is that enough?
Most often not. A typical situation: a window is displayed, a button is clicked, and all scripts are loaded immediately and regardless of the selection. A banner without tagging is a decoration that gives the false impression that the issue is closed. A behavior check takes several hours and shows this clearly.
What happens to the data after implementation?
There will be fewer of them - this is the expected effect, not a breakdown. Some users will opt out, and their actions will not be included in the reports in their entirety. The limited collection mode partially compensates for this with impersonal signals, but there will be a drop in performance in the first weeks. We warn about this in advance, because it is regularly confused with a drop in sales.
Is it possible not to ask consent at all?
This is a legal question, not a technical one, and the answer depends on the markets you serve and the data you collect. We do not give legal assessments and do not undertake to guarantee compliance with the law - this is the work of a lawyer. Our part is different: to make sure that the site technically behaves exactly as you decided together with the lawyer.
How does consent affect advertising?
Due to the size of the audiences and completeness of conversions. Remarketing lists are shrinking, some conversions are coming in an impersonal form. Ad systems use simulation to partially fill in the gaps, but accuracy drops. This is the price for correct work with data, and it is worth making plans with this in mind, not against it.
What systems must be subject to consent?
Everything that puts files on the device or transmits data to the outside: analytics, advertising pixels, chats, call tracking, maps, video inserts, social network widgets. Most often, third-party widgets are forgotten - no one thinks of them as collecting data, but they work the same way. We list them in the check by name.
Is it possible to let the user change the selection later?
It's not just possible, it's necessary: the link to the consent settings should be available at all times, usually in the footer of the site. This is both a requirement of most regulations and a practical convenience. We put it together with the banner, not as a separate add-on.
How to verify that consent is really managing tags?
Two ways, and both are available to you without us. First: open the site in a clean window, turn on the network tab, and see what requests went before you clicked on anything. If there is already an appeal to analytics and advertising counters, the banner is decorative. The second, more precise one: the preview mode in the tag manager shows the state of each signal at the moment when a particular tag is triggered. We run through four scenarios — consent, refusal, partial consent, and change of mind through the link in the footer. It is usually the last one that breaks.
Basic or advanced mode - what to choose?
Our default recommendation is extended with default denied: tags are loaded, but no identifiers are written to the selection and impersonal signals are sent, so advertising systems have something to model. We take the basic one when the lawyer insists that nothing be added to the agreement; at the same time, we honestly say that the reports will be poorer. And a tip that costs zero: take your baseline a week before your banner launch. Otherwise, in a month there will be nothing to compare the decline with, and instead of an analysis, a dispute will begin about who caused it.
Send the site address - we will see what is loaded on it before the user's choice.
In response, you will receive a description of the actual behavior before and after the click, a list of systems that currently ignore consent, and an estimate of how much data volume will change. If you have everything compiled correctly, you will hear it in the first email and without an estimate.
From measured casesThe consent mode v2 is correct: default denied, the status is declared before the tags
There is no price list on the site on purpose: the same work differs several times over between two clients, and a “from” figure explains nothing in that case. First a free audit — we count your pages, duplicates and speed — then we name the sum and the deadline and fix both in the contract.