If you run Google Ads to visitors in the EU or UK and your cookie banner is not wired into Google Consent Mode v2, your conversion numbers are probably lower than reality, your remarketing lists may be shrinking, and Smart Bidding is working with a thinner signal than it should. The fix is a three-part job: a consent banner that actually passes consent signals to Google, correctly configured tags, and Enhanced Conversions to win back some of the data that consent refusals remove. Done properly it takes a competent person a day, not a month.
We are writing this in October because Q4 is when tracking mistakes get expensive. Budgets go up, bidding algorithms learn faster, and any gap in measurement gets multiplied by spend. Fix the plumbing now and Black Friday runs on clean data.
What is Consent Mode v2, in plain language?
Consent Mode is a way for your website to tell Google’s tags what the visitor agreed to in the cookie banner. Version 2 added two signals on top of the original two:
- ad_storage and analytics_storage (the original pair): may cookies be stored for advertising and for analytics?
- ad_user_data: may user data be sent to Google for advertising purposes?
- ad_personalization: may the data be used for personalised advertising such as remarketing?
Since March 2024 Google has required these signals for advertisers who want to use audience and remarketing features with traffic from the European Economic Area and the UK. Without them, those features stop collecting from that traffic. So this is not an optional nicety: for a European advertiser it is the entry ticket to remarketing and to a good part of conversion measurement.
Basic mode vs. advanced mode: which one should a small business use?
This is the choice that matters most, and it is mostly a legal and risk decision rather than a technical one.
| Basic mode | Advanced mode | |
|---|---|---|
| Google tags before consent | Blocked entirely | Load, but send cookieless pings only |
| Data for visitors who decline | None | Aggregated signals used for modelling |
| Conversion modelling quality | Weaker, relies on consenting users | Better, has more signal to model from |
| Privacy posture | More conservative | Needs your legal basis to be comfortable with pings |
Our default recommendation for small advertisers is advanced mode, provided you are comfortable with it from a data-protection standpoint and your privacy policy describes it accurately. If your legal adviser or your sector (health, finance, anything sensitive) pushes you towards the cautious reading, basic mode is perfectly defensible. You will simply see fewer modelled conversions and should expect a bigger gap between reported and real results. We are a marketing agency, not a law firm, so treat this as a decision to make with someone qualified to give legal advice.
Why do conversions drop when you switch it on?
Two different things happen at once, and people confuse them.
First, a real measurement loss. Every visitor who declines cookies is a visitor Google can no longer tie to a click directly. That loss existed before Consent Mode; you just could not see it. Second, Google fills part of the gap with modelled conversions: statistical estimates of conversions from unconsented traffic, based on the patterns of consenting users. Modelling needs a minimum amount of volume to switch on, and Google documents those thresholds in its help centre (they are measured in hundreds of clicks over a week, so check the current figures). A tiny account may never reach them.
The practical consequence: after implementation, expect reported conversions to wobble for two to four weeks. If you compare against a period when the banner was not passing signals correctly, you may even see an apparent increase from modelling. Do not make bidding decisions on the first fortnight of data.
The implementation checklist we use
1. Choose a consent banner that supports Consent Mode v2
Most mainstream consent management platforms (CMPs) now support it, and Google maintains a list of partner CMPs. What matters is that the banner sets the default state to “denied” for EEA/UK visitors before any Google tag fires, then updates the state when the visitor chooses. A banner that merely hides itself while tags fire regardless is not compliant and does not give Google what it needs.
2. Set defaults and updates in the right order
The default consent command has to run before the Google tag loads. If you use Google Tag Manager, use the consent initialisation trigger for that command, not the ordinary page-view trigger. This single ordering mistake is the most common reason we find “Consent Mode installed” accounts that are not really working.
3. Check the four signals separately
Open Tag Assistant or the GTM preview, accept all cookies and confirm all four signals flip to granted; then reject all and confirm they stay denied. Then test the middle case: analytics yes, advertising no. Many banners map their categories to only two of the four signals and silently leave the new ones in the wrong state.
4. Add Enhanced Conversions
Enhanced Conversions for web lets your site pass hashed first-party data (typically a hashed email address or phone number from a form or checkout) with the conversion event. Google matches the hash against signed-in Google accounts to recover conversions that cookies alone would miss. Key points:
- The data is hashed with SHA-256 in the browser before it is sent; you should still describe it in your privacy policy.
- It only works if your form or thank-you page actually exposes the email or phone field to the tag, which sounds obvious and is the most common failure.
- It respects consent: with ad_user_data denied, the enhancement data is not used.
5. Verify in Google Ads diagnostics
In Google Ads, under the conversion action, look at the diagnostics tab. It will flag whether consent signals are being received and whether Enhanced Conversions are being matched. Do not rely on your own testing alone. Give it a few days of traffic and look again.
How do you know if it is working?
Three checks, in order of how much we trust them:
- Tag diagnostics show consent status and match rates. This is a technical confirmation.
- Consent rate from your banner reporting. Many European sites see consent rates that vary hugely by banner design, sector and country; if yours is dramatically low, the banner design, not Google, is the problem. A banner that makes “reject” easier than “accept” will lose data, but that is a legal-compliance choice you should not undo for tracking’s sake.
- Reconciliation with real business data. Compare Google Ads conversions with orders, leads in your CRM, or booked calls for the same weeks. You are not looking for equality. You are looking for a stable ratio. If Google reports 70 percent of your actual leads one month and 72 percent the next, your setup is consistent. If it swings between 40 and 90, something is broken.
What does this do to bidding?
Smart Bidding strategies such as Target CPA and Target ROAS learn from the conversions they can see. If a third of your real conversions are invisible, the algorithm underestimates what each click is worth and may bid too cautiously, particularly in markets with low consent rates. Restoring signal through Consent Mode and Enhanced Conversions tends to help most on small accounts where every conversion counts. It will not rescue a weak offer, but it removes a handicap. If you are deciding whether to take tracking further with server-side setups, our guide to server-side tracking for small businesses covers when that extra step is worth the cost.
Common mistakes we see in audits
- Banner on the page, signals not wired. The banner records a choice, but the tags never hear about it.
- Defaults set to granted. Convenient for numbers, and a compliance problem in the EEA and UK.
- Two tracking installs. A hard-coded Google tag and a GTM tag both firing, with only one of them consent-aware.
- Conversion action counted twice. A thank-you page URL trigger plus an event trigger, which inflates numbers and misleads the bidding.
- Testing only on desktop Chrome. Safari and mobile browsers cut data in other ways, so check there too.
A sensible rollout plan before Q4 spending ramps up
If you are reading this in October with Black Friday on 27 November, work backwards. Week one: choose and install the banner, set defaults. Week two: verify all four signals, add Enhanced Conversions, and make sure only one conversion tag fires per action. Weeks three and four: let data accumulate and reconcile against real orders or leads. Only then raise budgets. Our note on how much to raise your Q4 ad budget assumes your measurement is already trustworthy, which is exactly what this work buys you.
One more thing worth saying plainly: tracking is never perfect any more, and it does not need to be. The goal is a stable, honest approximation that bidding algorithms can learn from and that you can reconcile to money in the bank. If you get there, you are ahead of most advertisers we audit, including some that spend ten times as much. For a broader view of how to read the numbers once they are flowing, see how to vet a paid media agency, where measurement questions come near the top of our list.