Marketing tag audit

A marketing tag audit answers one question: does the tracking you believe is running actually run, on every page, for every visitor? It is a different question from "is the tag installed", and the gap between them is where reporting quietly stops matching reality.

Docket runs 6 checks in this area on every audit. They are the least glamorous part of the report and the most likely to be costing money right now, because a missing tag does not break anything a visitor can see. Nothing errors. The site looks fine. The numbers in your dashboard are simply smaller than the truth, and they have been for as long as nobody checked.

Why "installed" is the wrong question

Tag managers are installed once, on a template, and then the site grows. A landing page built for one campaign gets published from a different template. A checkout flow gets rebuilt by a contractor. A blog moves to a subdomain. Each of those is a page where the tag is absent, and none of them announce it — the tag is still installed, on the pages it was installed on.

So the check that matters is coverage: of the pages Docket read, how many carried the tag. One page missing a pixel in a funnel of four is a quarter of your conversions attributed to nobody.

What Docket looks for

• Advertising pixels mar.ad_pixels
• Analytics installed mar.analytics
• Consent and privacy compliance mar.consent
• Email list building mar.email_capture
• Marketing stack inventory mar.stack
• UTM tagging hygiene mar.utm_hygiene

Two of those need explaining, because they are not what people expect from a tag audit.

Consent and privacy compliance is checked because a tag firing before consent is worse than a tag missing. A missing tag costs you data. A tag that fires before the visitor agrees is a legal exposure in the EU and the UK, and it is invisible in every dashboard — the data arrives and looks perfect.

UTM tagging hygiene is about your own links. Inconsistent casing splits one campaign into two rows in your reports, and internal links carrying UTM parameters overwrite the original source of the session — so a visitor who arrived from search and then clicked your own tagged banner is recorded as arriving from the banner. Your best channel gets credited to your worst.

What we found looking at real sites

On 2026-08-14 we audited four live business sites — an Italian food retailer, a French bakery chain, a US coffee retailer and a US museum — and the tracking lane produced a finding on the coffee retailer that is worth quoting exactly as it was written:

No analytics tag on any page Docket could see

The important words are could see. That site renders its pages with JavaScript and ships almost no HTML, so a tag injected at runtime would not appear in what Docket read. The finding says what was observed and stops there, rather than concluding that a large retailer has no analytics — which would be a claim about somebody's business made from an absence of evidence.

That is deliberate, and it is the same rule everywhere in the product: a check that could not see must not report nothing as though it were zero. If you want the JavaScript-rendered answer, --render runs the page in WebKit and audits what a browser actually gets. How that works, and when it is worth the time.

Four sites is not a sample and there is no rate here to quote. It is one measurement, reported because it shows what the lane does and where it stops.

Doing this yourself, without Docket

You do not need a tool for the first pass. Open your site, view source on five pages chosen from different templates — homepage, a product or service page, a blog post, a landing page, the thank-you page after a form — and search each for your measurement ID. The thank-you page is the one to check first: it is the page that proves a conversion happened, it is usually built last, and it is the page most often published from a template nobody added the tag to.

If all five carry it, you have ruled out the common failure. If one does not, you have found the thing this lane exists for, and it took four minutes.

What this lane deliberately does not do

It does not tell you whether your tracking is correct — whether the events fire on the right actions, whether revenue is passed accurately, whether your goals mean what you think. Those need to be tested in your analytics account against real sessions, and a crawler cannot see them. Docket reports what is present in the pages it read and does not guess at the rest.

Download Docket

Common questions

What is a marketing tag audit?

A check that the analytics and advertising tags you believe are running actually appear on every page, for every visitor, and fire at the right time. It is different from checking a tag is installed: tags are installed on templates, and sites grow pages that use other templates.

Why would a page be missing my analytics tag?

Because tags are added to a template, not to a site. Landing pages built for a campaign, checkout flows rebuilt by a contractor, and blogs moved to a subdomain are all pages published outside the template the tag was added to. Nothing errors when it happens.

Can Docket see tags that JavaScript injects at runtime?

Not by default — it reports what is present in the HTML it read, and says so rather than concluding a site has no analytics. Running an audit with --render executes the page's JavaScript in WebKit first and audits what a browser actually receives.

Does Docket check whether my tracking is correct?

No. It checks what is present in the pages it read. Whether events fire on the right actions and whether revenue is passed accurately can only be tested against real sessions in your analytics account, and a crawler cannot see them.