Tag anomaly notifications

Tag anomaly alerts tell you when one of your tags suddenly fires much more, or much less, than expected. This allows you to fix broken tracking before it affects your reports, your campaigns and your ad spend.

What will you do?

• Learn what a tag anomaly is and why it matters
• Choose which tags you want to be alerted about
• Understand an anomaly alert and know what to do next

⏰ Estimated time: 5–10 minutes

🔧 Requirements: Tag Monitor running on your website, an Advanced or Premium plan, and access to the Code Cube portal

What is a tag anomaly?

Every tag on your website has its own normal rhythm. Your purchase tag fires a certain number of times per hour, your page view tag a lot more. Code Cube learns that rhythm for each of your tags and lets you know when a tag suddenly breaks from it.

There are two kinds of anomalies:

• Spike: the tag fires much more than usual. For example, after a website update your purchase tag fires twice for every order, and your reports show double the revenue.
• Drop: the tag fires much less than usual, or stops completely. For example, after a change to your cookie banner your conversion pixel stops working, and your ad platforms no longer see conversions.

🧠 Why this matters: An anomaly doesn't show up as a tag error. The tag technically still works, so without anomaly alerts these problems can go unnoticed for weeks.

When will you hear from us?

Every morning at 09:00 (Central European Time (CET, UTC+1)) you receive one summary per day of the anomalies we found in the past 24 hours.

• A tag that behaved unusually for several hours appears only once, with its most unusual hour highlighted.
• The most urgent tags are always at the top of the alert.
• When there are no anomalies you won't receive a message. No news is good news.

How urgent is it?

Every anomaly gets a label, so you can see at a glance what needs your attention first.

An example: your purchase tag normally fires 100 to 150 times an hour.

• One hour it fires 700 times, almost 5 times the normal maximum. That's a Warning.
• Another hour it fires only 10 times, 90% below the normal minimum. That's Critical.

🧠 Good to know: We always compare with the edges of a tag's normal range, never its average. Everyday ups and downs therefore never trigger an alert.

Choose which tags you want to be alerted about

You decide which tags we watch for anomalies. That way you only hear about the tags that matter most to you, like purchases, leads or sign-ups.

You set this up in the portal, under Tag Volume alerting. For each monitoring rule:

1. Choose the tags, domains and GTM containers the rule applies to.
2. Under What to monitor, choose Anomaly, or Both to also get alerts when a tag stops firing.
3. Under Alert me on, choose how much you'd like to hear:
• Major only (recommended): get alerted only about Critical and Warning anomalies
• All anomalies: the above as well as alerts about smaller changes (Info)

🔗 Step-by-step instructions: Tag Monitor setup → Tag Volume alerting

⚠️ Important: We only send alerts for tags that are part of a monitoring rule with anomaly monitoring switched on. Is a tag not in any rule? Then you won't hear about anomalies, regardless how big the change.

Where will you receive alerts?

Anomaly alerts arrive in the same places as your other Tag Monitor notifications, so there's nothing extra to set up. Every channel you use, whether it’s email, Slack, Teams or another channel, receives the same alerts.

Want to add a channel or change who gets the alerts? Follow Setup notification channel.

💡 Tip: Send the alerts to a shared channel your marketing and analytics team already use. A sudden spike or drop is often easiest to explain together, for example when a campaign has just gone live.

What's in an alert?

Whether it arrives by email, Slack or Teams, every alert looks the same. At the top you'll see the period it covers and how many anomalies were found, for example 1 critical, 2 warning.

Below that you find per tag one card (most urgent first):

🧠 Good to know: On a day with many anomalies, Slack and Teams may split the summary over a few messages. Email always arrives as one message.

What to do when you get an alert

1. Click View this tag → to see the tag's activity in your dashboard.
2. Think about what changed around that time: a website release, a new GTM version, a cookie banner update or a new campaign.
3. Was it expected, like a big sale or newsletter?
4. Not expected? Check the tag's triggers and your recent GTM changes, or forward the alert to the person who manages your tags.

We won't repeat the alert

Once we've informed you about an anomaly, we won't send it again every day while it continues. In that way new issues can’t be missed.

We'll only contact you again about the same tag when:

• It gets worse, for example when a Warning becomes Critical
• It's still going on after a week. We'll send a friendly reminder, and repeat that every 7 days until it's resolved
• It comes back. The tag returned to normal, and then behaves unusually again

🧠 Good to know: A tag counts as back to normal once it has had a full day without any unusual hours.

Check that everything is set up

You can't trigger an anomaly on purpose, but a quick check makes sure you're ready for the next morning summary.

1. In the portal, open Tag Monitor > Configuration > Tag Volume alerting.
2. Make sure at least one active rule has What to monitor set to Anomaly or Both.
3. Make sure the rule's container, domain and tag cover the tags you care about.
4. Choose Alert me on: Major only or All anomalies.
5. Make sure at least one notification channel is switched on and reaches the right people.
6. Take a look at the anomaly view in your Tag Monitor dashboard. Anomalies you see there for tags in your rule will show up in the next morning's summary.

❓ Didn't get a message? That usually just means there were no new anomalies. Monitoring is still running, and you'll hear from us the morning after something unusual happens.

Troubleshooting

Not getting the alerts you expected? Most of the time a small change to your rule settings fixes it.