# How to avoid alert fatigue

> Too many ad alerts and your team stops reading them. Five fixes for alert fatigue: fewer rules, severity, separate channels, snooze vs off, and a weekly prune.
- **Author**: Marxx
- **Published**: 2026-11-06
- **Category**: Alerts
- **URL**: https://marxx.ai/library-guides/how-to-avoid-alert-fatigue

---

You avoid alert fatigue by sending fewer, sharper alerts: only rules that need someone to act, each with a severity, in a channel where the right person watches. Then prune the list every week, so an alert that never leads to action does not survive.

## Signs you have it

- People react with an emoji and do nothing.
- The alerts channel is muted by half the team.
- The same alert fires every day and everyone knows why.
- A real problem sat in the channel for a day before anyone saw it.

If two of these sound familiar, the problem is the rules, not the team.

## Five fixes

1. **Cut to the alerts that need action.** For each rule, ask: "When this fires, what will someone do?" If the answer is "look at it", move it to a weekly report instead. Start with four or five rules.
2. **Give every alert a severity.** High means act today, Medium means this week, Low means review on Monday. People learn to scan for High.
3. **Use separate channels.** Put alerts in their own channel per client, apart from leads and digests. A lead message and a High CPA alert compete for the same eyes otherwise.
4. **Snooze known causes, turn off wrong rules.** Snooze when you already acted and are waiting. Turn off when the rule should never have fired.
5. **Add a minimum spend.** An ad set that spent INR 400 cannot tell you much. Skip it until it spends enough to judge.

## A worked example

Say an agency runs alerts for three clients. In one week it got 63 alert messages and acted on 7. That is 1 useful alert in 9.

| Change | Alerts per week |
|---|---|
| Starting point | 63 |
| Removed 4 "nice to know" rules | 38 |
| Added INR 2,000 minimum spend to CPA and CTR rules | 21 |
| Snoozed 2 ad sets already being fixed | 14 |

The team still acted on 7 alerts. Now 1 in 2 messages leads to action, and people read the channel again.

## Run a weekly prune

Every Monday, spend 10 minutes on this list:

- Which alerts fired most? Is each one still worth a message?
- Which alerts were snoozed? Was the cause fixed?
- Did we find any problem that no alert caught? Add a rule for it.
- Is any rule firing on tiny spend? Raise its minimum.

Fold the answers into your [weekly ad report](/guides/how-to-build-a-weekly-ad-report), so the client sees what changed and why.

## Where Marxx fits

We make Marxx. Its alerts carry a High, Medium or Low severity, go to the client's own Slack channel, and have Snooze 7 days and Turn off buttons in the message. Setup is in [Meta ads alerts in Slack](/guides/how-to-set-up-meta-ads-alerts-in-slack), and a [channel naming convention](/templates/slack-channel-naming-conventions-for-agencies) keeps alert channels easy to find.

## FAQ

### How many ad alerts is too many?

There is no fixed number. If fewer than half your alerts lead to action, you have too many.

### Should alerts go to email or Slack?

Send High alerts where the responder works all day, usually Slack. Low alerts can go to email or a weekly report.

### Is snoozing an alert a bad habit?

No, as long as the cause is known. Reply in the thread with why you snoozed, so the next person knows it is handled.

---
- [More Alerts articles](https://marxx.ai/library-guides/category/alerts)
- [All articles](https://marxx.ai/library-guides)