Across the public TagEasy audit corpus, the most common breakage isn't "no GTM at all" — most sites have one. It's subtler.
The four ways a GTM container quietly breaks
1. Debug mode leaks to production
When a developer is previewing a container, GTM appends `?gtm_preview=env-3>m_auth=...` query params to the script src. If those params survive the deploy, every visitor loads the debug container instead of the published one. We've seen this on real production sites in the corpus.
How to spot it: search your homepage HTML for `gtm_preview` or `gtm_auth`. If present, you're live in debug mode.
2. Multiple GA4 properties on the same page
A site adds GA4 (G-AAAAA) directly via gtag.js for the marketing site, then someone adds a SECOND GA4 (G-BBBBB) inside GTM for the app. Now every pageview fires twice — once to each property. The reports look "fine" individually but rows don't reconcile.
3. Pixels configured in TagEasy / settings, not loading on the page
The setup wizard collects pixel IDs (Meta, TikTok, Google Ads). They get saved. But if the GTM tag template was deleted, paused, or had its trigger broken, the pixel ID lives in your settings and never fires in the user's browser.
This is the exact gap TagEasy's AI debugger compares against — your configured pixels vs. what the live homepage actually loads.
4. Consent Mode v2 absent, Meta + Google Ads still firing
See the separate post on Consent Mode v2 — short version: in our corpus, fewer than 20% of sites that run Meta or Google Ads have a working consent gate.
The 4-step quarterly audit any team can run
- Search your homepage HTML for `gtm_preview` and `gtm_auth`. Should return zero matches.
- Grep for `G-` and `AW-` patterns; confirm each ID is one you recognise + intended. Multiple GA4 IDs deserve a comment in the container documenting why.
- Cross-check every pixel ID in your tag manager against your ad platform UI — does each one map to a live, billing campaign?
- Run the TagEasy free audit on your homepage. Each check has an "Explain in detail" expander that walks the specific finding.
TagEasy's public auditor runs all four of these checks (plus 9 more) automatically. The /benchmarks page shows the failure rate across our corpus so you can see how your site stacks up.
Why most teams discover this too late
GTM publishes; no one watches what it actually does. Drift creeps in over weeks. The first signal is usually "Why are our Meta CAPI numbers down 30% this month?" — which is the wrong question. The right one was: "Did the purchase tag stop firing in version 47 of the container?"
The TagEasy daily drift cron answers that automatically; you can also do it by hand quarterly with the checklist above.