Tag Debug Assistant

Debugging your tags made simple and efficient.

Overview

On the Tag Monitor main page, a Tag Diagnostics column is shown at the end of the Error Monitoring table. Each of your broken tags is assigned a severity level: Low, Medium, or High.

📌 Please note: an insight is one failing tag, grouped by the pages it failed on. It is not a broken page. Consent is never the reason a tag appears here.

Severity

Severity ranks on recent activity: how many times did the tag fail in the last 24 hours?

Severity	When it is assigned
High	More than 50 failures in 24 hours, or more than 5,000 failures in total
Medium	More than 20 failures in 24 hours, or more than 500 failures in total
Low	More than 5 failures

Trend

Next to the diagnosis, a trend marker (▲) may appear.

Marker	Meaning	Most likely origin
▲	Most failures occurred in the last 48 hours, or the whole run of failures is less than 24 hours old	A recent GTM publish or a site deployment
No marker	Failures are spread evenly across the window	-

📌 Please note: both severity and the trend marker are calculated by fixed rules, not by the AI, so they are consistent from week to week.

💡 Suggestion for triage of diagnoses: combine High severity diagnoses with the ones marked with the trend marker (▲). This combination means a lot of traffic is failing and something changed very recently. The fix is usually a rollback of a deployment or a one-line correction in the last publish.

Individual diagnosis

A diagnosis is given per row. Click the diagnosis to open the complete diagnosis.

🧠 The diagnosis column only appears when there are diagnoses.

In the diagnostic panel, the Tag Debug Assistant gives you the following information:

Card	What it tells you
Plausible root cause	Names the specific parameter, flag, event, or timing at fault
Possible business impact	The consequence for reporting, marketing, or compliance
Analysis	The evidence trail behind the diagnosis
Potential fixes	Guidance shown open, with escalate / hand-off steps collapsed

The evidence behind the diagnosis

In the evidence, expect to see references to:

The event context: which GTM event the tag fired on, for example gtm.js or add_to_cart
Parameter presence: which parameters the tag actually received, with exact ratios such as "price was null in 35 of 47 failures"
Consent state: only when a flag was denied in more than 70% of sampled failures, and only ever as a gating gap
Failure timing: how the failures are distributed across the window

Potential fixes

Section	Owner	Typical action
Analysts' Actionable (shown open)	GTM / analytics owner	Change a variable, trigger, or consent check inside the container
Escalate / hand off (collapsed by default)	Front-end / web team	Change what the site pushes to the dataLayer, or how a vendor script is loaded
Escalate / hand off (collapsed by default)	Marketing / CRM team	Adjust expectations on reporting, or brief stakeholders on the data gap

💡 Suggestion for building a ticket: use the copy icon on the Potential fixes card. It copies every fix as plain text, including the collapsed hand-off ones, already grouped and labelled by team. Add the root cause sentence, and add the sample URLs from the footer so the assignee has a page to reproduce on.