When to run this checklist
Run it before a new banner or consent platform is switched on, and before any change to an existing one goes live: new wording, a category added or merged, a vendor added to the list, a new tag in the container, or a switch from one consent platform to another. Small changes count. Moving a tag from one category to another changes which visitors it fires for, and a vendor added to the list without a matching tag setting is exactly the gap that lets a pixel fire on a reject.
The list is usually run by whoever owns the tag manager container, with the person responsible for privacy reading the banner and records groups and the analyst reading the measurement group. It is a go-live gate for the banner itself. Proving that every conversion reaches analytics and the ad platforms is a wider job, covered by the tracking verification checklist, which treats consent as one group among many. When a tag is already firing before consent on the live site, the work is diagnosis rather than QA, and the tags firing before consent guide is the place to start.
Before you start
- The banner wording and category descriptions as approved, so the live banner is checked against an agreed text rather than against itself.
- The vendor and tag list, showing which tool belongs to which category and what each one receives.
- Access to the GTMTag management system. container with preview rights, and to the consent platform's settings.
- A clean browser profile or private window with no earlier cookies, plus a phone for the mobile checks.
- A keyboard and a screen reader you know how to use, or a colleague who does.
- The current analytics and ad platform numbers for a recent period, so the drop after launch can be compared against something.
The checklist
The groups follow the order a visitor and then a reporting team meet the banner: what it says, what it controls, whether it behaves, what it does to the numbers and what is kept on file. The tags group is the hard gate: if any line in it fails, the banner does not go live, however polished the wording is.
Banner
Tags
Testing
Measurement impact
Records
Where each check is explained
The widget keeps each line short enough to tick; the reasoning sits in the guides. Each check links to the page that explains it.
| Check | Explained in |
|---|---|
| Offer accept and reject with equal prominence where the law requires it | Privacy, consent and measurement |
| Describe each cookie category in plain language | Privacy, consent and measurement |
| Make the banner usable with a keyboard and a screen reader | Privacy, consent and measurement |
| Provide a link to change choices later from every page | Privacy, consent and measurement |
| Confirm no marketing tag fires before consent is given | GTM tags firing before consent |
| Set consent mode defaults before any tag loads | Consent mode |
| Update consent mode when the visitor makes or changes a choice | Consent mode |
| Map every tag in the container to a consent category | Container governance |
| Test accept, reject and partial choices in a clean browser | Debugging with preview mode |
| Check the choice persists across pages and on return visits | Debugging with preview mode |
| Check the banner on phones does not hide the page's main action | Speed and mobile for landing pages |
| Annotate the banner launch date in analytics and reports | GA4 data drop after a cookie banner |
| Compare traffic and conversions before and after launch and explain the gap | GA4 data drop after a cookie banner |
| Brief ad platform owners on modelled conversions and reporting gaps | Why Google Ads and GA4 disagree |
| Keep a record of the banner wording and vendor list at launch | Privacy, consent and measurement |
| Update the privacy notice to list every tool that receives data | Tag governance |
If an item fails
A failure in the tags group stops the launch. A marketing tag that fires before consent, or a consent mode default that arrives after the first tag, is a privacy problem before it is a measurement one, and the banner should not go live until preview shows the tag waiting. Fix the line, then rerun the whole tags and testing groups in a fresh clean browser, because the fix usually changes trigger order and the order is what failed.
Banner failures depend on where the site operates. Whether reject must be as prominent as accept, and what counts as a valid choice, varies by jurisdiction, so the equal-prominence line says "where the law requires it" and the decision belongs to whoever is accountable for privacy, not to the person building the banner. An accessibility failure, such as a banner that traps keyboard focus, should be fixed before launch: a visitor who cannot dismiss the banner cannot use the site at all.
The measurement and records groups rarely block a launch, but they should not be skipped. A banner can ship with the comparison still to do, provided the launch date is annotated on the day; without the annotation, the drop is discovered weeks later and blamed on a campaign. Ad platform owners should also know that their totals may include modelled conversionsConversions a platform estimates statistically for activity it could not observe directly, such as unconsented or cross-device journeys. for visitors who declined, so platform and analytics figures will not move together. The data drop after a cookie banner guide explains how to size and explain the gap once a few weeks of data exist. The banner governs cookie and tag choices only; marketing permissions and objections held in the CRM and the email platform are covered, purpose by purpose, in the tutorial on mapping consent across the marketing stack. The records group also keeps the published privacy noticeThe published notice that tells people what personal data the business collects, why, and which tools and partners receive it; marketing keeps it current. in step with the vendor list, so every tool the banner controls is named there too.
A ceramics shop's banner switch
Ospreth Ceramics, an online pottery shop, replaces a notice-only cookie bar with a consent platform offering accept, reject and settings. Preview in a clean browser shows the analytics tag waiting correctly, but the social pixel fires on the first page view whatever the visitor chooses: it had been added as a custom HTML tag two years earlier and never assigned a consent category. The launch waits a day while the pixel is moved to the vendor's own template and mapped to the marketing category. After launch, recorded sessions fall by around a third and the shop's paid social account reports fewer purchases. Because the launch date was annotated and the ads owner had been briefed on modelled conversions, the monthly report explains the drop in one paragraph instead of a week of investigation.
Common mistakes
- Testing in a browser that already said yes. The consent cookie from an earlier visit makes every tag look compliant; only a clean profile shows the default state.
- Testing accept and forgetting reject. The accept path works because everything fires; the reject and partial paths are where tags leak.
- Consent defaults that load too late. The default is set by a tag that fires after the analytics tag, so the first hit goes out without any consent state.
- Categories nobody can explain. Labels copied from a template describe tools the site does not use, and miss the ones it does.
- A settings link that only lives in the banner. Once dismissed, a visitor has no way back to change their mind.
- Surprise in the monthly report. Nobody told the ad platform owners, and the drop in reported conversions is treated as a campaign failure.
The tracking verification checklist is the companion list for proving tracking end to end once the banner is live, and the checklists library holds the rest.
