Quebec SMBs: what consent really takes away from your measurement
Google Consent Mode promises to recover, through modelling, the conversions lost when a visitor refuses cookies. What the documentation states and no one mentions: modelling requires 700 clicks in seven days per domain country. Below that threshold, you recover nothing.
- Google sets the activation threshold for conversion modelling at 700 ad clicks in seven days per domain country. Many Quebec SMBs never get there.
- In basic mode, a refusal blocks everything: no data is sent to Google, not even the consent state. The tags do not fire.
- Below the threshold, you steer a budget on a fraction of your real sales, without knowing which fraction. That is a decision risk, not a technical problem.
- The answer is not to get around consent. It is to measure the gap, report it, and tie your real sales to your management system rather than to the platform.
- A vendor who promises to recover everything with Consent Mode either does not know the threshold or is not telling you.
On this page
What is the consent gap
The consent gap is the difference between the conversions a company actually produces and the ones its advertising platforms can observe, once you remove the visitors who refused cookies. It does not close by switching tools. It is measured, reported and offset by tying real sales to the management system rather than by trying to recover data the visitor refused to give.
What Consent Mode actually does
This article is for owners and marketing managers who buy advertising and have watched their measured conversions drop without sales following. It is about measurement, not compliance itself, which we cover in our article on compliance with Law 25, Quebec's private-sector privacy law.
Consent Mode is not a banner. It is a mechanism that passes your visitors' choice to Google and adjusts tag behaviour accordingly. It comes in two versions, and the gap between them is considerable.
In the basic version, when the visitor refuses, Google states that no data is sent, not even the consent state, and that tag firing is fully blocked. Conversion modelling, the statistical estimate of unobserved sales from the observed ones, then relies on a general model rather than on your own data.
In the advanced version, the tags load anyway and send cookieless signals as long as consent is not granted. Google can then model from your real situation, provided you reach a threshold.
The threshold most SMBs never reach
This is the point almost no sales proposal mentions.
Google's documentation states that the click threshold for conversion modelling is 700 ad clicks in seven days per domain country. Below it, the diagnostic tool shows that Consent Mode is implemented but that the required thresholds are not met.
Do the math for your own account. Seven hundred clicks in seven days means about a hundred clicks a day, steadily, in a single country. A Quebec company with a monthly budget of a few thousand dollars and a cost per click of $3 or $4 does not get there, except at peak.
The consequence is rarely spelled out: you paid for an implementation that does not deliver what you were promised. Compliance is genuinely covered, and that alone is reason enough to have put it in place. The conversion recovery, though, will not happen.
A vendor who justifies a budget by promising to recover lost conversions through modelling, without checking your click volume, is selling a feature your account will never turn on. The question to ask is direct: do we reach 700 clicks in seven days, and does the diagnostic tool confirm that modelling is active?
Measuring your consent gap
You cannot recover the refused data. You can, however, know how much of it is missing, and that one number changes how you read your reports.
Three numbers are enough, and you can pull them in half a day. The acceptance rate of your banner, which your consent management platform provides. The number of real sales recorded in your management system over the period. And the number of conversions reported by your advertising platforms over the same period.
The gap between the last two numbers, set against the first, is your consent gap. If it is consistent with your refusal rate, the situation is understood. If it is much wider, you have another tracking problem, independent of consent, and it is better to know that before blaming the regulation.
This number then has to appear in your monthly reports. A dashboard that shows 180 conversions without noting that roughly 40% are missing leaves management deciding on an incomplete picture.
Steering a budget on partial measurement
The good news is that partial measurement stays usable, provided it is stable and reported.
What stays reliable: comparisons over time, at a constant consent rate. If your measured conversions rise 30% from one month to the next and the acceptance rate has not moved, the increase is real even if the absolute level is understated.
What becomes false: comparisons between channels whose consent rates differ, and any return calculation in absolute value. A return computed on 60% of the sales directly understates performance, and it can get a profitable campaign cut.
What stays true in every case: the sales recorded in your management system. It is the only source that depends on no consent, and that is why it should become your reference measure, as explained in our article on cost per signed quote.
The three costly mistakes
Trying to get around the refusal. Techniques that restore tracking despite a refusal expose the company to a penalty and settle nothing in the long run. The legal risk far outweighs the measurement gain.
Making the banner deliberately confusing to inflate acceptance. Consent obtained through an ambiguous flow is not valid consent. You then stack the regulatory risk on top of data you have no right to use.
Comparing periods before and after the banner. This is the most common mistake and the most costly in decisions. The drop in measured conversions when a banner goes live is not a drop in sales, yet it gets budgets cut every year.
The third mistake is worth pausing on, because it happens in a meeting and it sounds reasonable. A management team sees conversions fall 40% in the month of the compliance rollout, concludes that marketing no longer works, and cuts. Sales, meanwhile, had not moved.
The durable way out
It comes down to three decisions, in this order.
First, move the reference measure to your management system. Your real sales depend on no consent, and it is the only source that will stay accurate whatever the regulatory or technical changes.
Next, keep the platforms as optimization tools rather than as sources of truth. They receive a signal, they train bids, and their number serves to compare campaigns against each other, not to measure your revenue.
Finally, report the gap in every report. It is a matter of honesty, and it is also what keeps an owner from deciding on a cut based on a number they believe is complete.
What to verify before concluding anything about your numbers
These five checks take half a day and prevent budget decisions made on incomplete measurement.
- What is your banner acceptance rate, month by month, over the past year?
- Do you reach 700 clicks in seven days, and does the diagnostic tool confirm that modelling is active?
- What gap separates your recorded sales from the conversions the platforms report?
- Do your monthly reports state this gap, or do they present the conversions as a total?
- If the gap far exceeds your refusal rate, who checks the tracking, and within what timeframe?
A solid answer gives an acceptance rate and a diagnostic state. A vague answer talks about compliance and modelled data. A vendor who promises to recover your conversions without having checked your click volume is selling a feature your account will not turn on.
Measuring your real gap and deciding which source is authoritative is exactly what we establish in a paid audit.
What is yours to own: the real sales from your management system, the decision on which source is authoritative, and the refusal of any workaround technique. What does get delegated: checking the modelling threshold, computing the gap, configuring Consent Mode and reporting the gap in the reports. A company that compares its recorded sales with its reported conversions understands its numbers in half a day. A company that does not will one day cut a profitable campaign, believing it is reading a drop in performance.
The overall view of conversion is in what conversion rate optimization really fixes.
Is the problem after the click and you already have a team? See our work on conversion rate optimization.
Frequently asked questions about Consent Mode
Does Consent Mode recover my lost conversions?
Only if your account reaches the modelling threshold, which Google sets at 700 ad clicks in seven days per domain country. Below it, the feature is implemented but inactive. Check the state in the diagnostics tab of your conversion actions before counting on this recovery.
Do you need basic mode or advanced mode?
Advanced mode lets Google model from your real situation, whereas basic mode fully blocks the tags on a refusal and relies on a general model. The choice between the two carries compliance implications that go beyond measurement and deserve legal validation, not just technical.
My conversions dropped after the banner, did I lose sales?
Almost never. Compare the same period in your management system: if sales are stable, you did not lose customers, you lost visibility into them. This is the most costly interpretation error in this file, because it gets working budgets cut.
Can you improve the banner's acceptance rate?
Yes, through clarity and readability, never through ambiguity. Consent obtained through a confusing flow is not valid, which leaves you with the regulatory risk and unusable data. Legitimate gains come from understandable wording and fast loading, not from a hidden refuse button.
Does server-side tagging solve this problem?
No, and it is important to say so. A visitor who refuses stays refused, wherever your tags run. Server-side tagging, meaning running your tags on a server you control rather than in the browser, solves other problems: the reliability of transmission and control over what is sent. It does not restore data the person refused to provide.
What should you report in a monthly report?
The acceptance rate for the period, the estimated gap between reported conversions and recorded sales, and the activation state of modelling. Three lines are enough. They keep a partial number from being read as a total, which is the starting point of most bad decisions on this subject.
- Google Tag Manager Help Centre, Verify your Consent Mode implementation, official documentation, accessed July 2026. Source for the threshold of 700 clicks in seven days per domain country.
- Google Tag Manager Help Centre, About Consent Mode, official documentation, accessed July 2026. Source on the full blocking of tags in basic mode and on the cookieless signals of advanced mode.
- Commission d'accès à l'information du Québec, Lignes directrices 2023-1, Consentement : critères de validité (in French), official guidelines, consulted in July 2026. Source for the principle that an interface that highlights acceptance over refusal can render consent invalid.
- Falia framework, arithmetic of the measurement gap. The rates are explicit scenarios, to redo with your own acceptance rate.
Gabriel almost always takes your first call and carries out your audit. He builds the strategy starting from your growth goal: where to put your budget, which market to test and how to connect each lead to a real sale in your CRM. He mainly leads engagements for three goals: Optimize the profitability of your digital campaigns, Develop a new market, and Generate demand and growth. With Geneviève, he also works on organic search (SEO), AI visibility (GEO) and conversion rate optimization (CRO). The sales a Google Ads or Meta Ads campaign brings in depend on the page that receives the click. He writes mainly about marketing strategy, paid advertising and measurement.
About Falia →
SEO
Paid advertising
Conversion
AI visibility
Web design