How to Get an Email When GA4 Traffic Drops (or a Slack Message)

Why GA4 does not just tell you

GA4 collects the data, but it is built around you opening the interface. Its automated insights appear on the Home page and the Reports snapshot, and they are never emailed. If nobody logs in, nobody learns about the drop. That is fine for one property you check every morning. It falls apart when you run five sites, a couple of apps, and a client property or two, and a search update or a broken tag silently halves traffic on one of them.

There are three ways to fix this. Two are built into Google products. One is a dedicated monitoring tool.

Option 1: GA4 custom insights (free, email only)

Custom insights are GA4's alerting feature. You define a condition on a metric, GA4 evaluates it on a schedule, and it emails you when the condition is met. Here are the exact steps in the current GA4 interface.

1. Open your GA4 property. On the Home page, scroll to the Insights section and click View all insights. You can also get there from Reports → Reports snapshot → Insights card.
2. Click Create. GA4 shows a list of suggested insights (for example, "Anomaly detection for Active users"). You can tick one of those or choose Create new to define your own.
3. Set the evaluation frequency: Hourly, Daily, Weekly, or Monthly. Hourly is only available for web data streams, not app streams.
4. Choose a segment (All Users by default; you can restrict to a country, device, or channel) and a metric, such as Active users, Sessions, Key events, or Total revenue.
5. Choose a condition. "Has anomaly" lets Google's model decide what counts as unusual. A percentage-change condition such as "% decrease more than 20" is explicit and predictable.
6. Name the insight, enter the email addresses to notify under Manage notifications, and click Create.

You need the Editor or Administrator role on the property to create custom insights, and each property is limited to 50 of them. Triggered insights also show up in the Insights dashboard for everyone with access to the property, and are retained for a year.

What custom insights cannot do

• No Slack, Discord, or webhooks. Email is the only delivery channel.
• No sensitivity control for anomalies. "Has anomaly" is a black box: Google's Bayesian model decides, and you cannot make it stricter or looser. Your only lever is to switch to a fixed percentage condition, which then fires on normal weekend dips unless you tune it carefully.
• One property at a time. Insights are configured per property. With ten properties you set up and maintain ten sets of insights, and the emails arrive with no portfolio context.
• No explanation. The email says the metric changed. It does not tell you which channel, device, or country drove the change.

We cover the model itself in more detail in GA4 anomaly detection: how Insights works.

Option 2: Looker Studio scheduled reports (a report, not an alert)

Looker Studio (Google's free reporting tool) can email a PDF of a report on a schedule: daily, weekly, monthly, or a custom cadence, to up to 50 recipients. Many people use this as a poor man's alert by scheduling a daily email of a traffic chart and eyeballing it.

It works, but it is a report, not an alert. The email arrives whether or not anything changed, so after a week most people stop opening it. The free tier allows one schedule per report, delivered at most once a day. Looker Studio Pro (a paid per-user license) adds Slack and Google Chat delivery and conditional "chart alerts" that email you when a value crosses a threshold. Those chart alerts are threshold-based, email only, and have no anomaly detection, so they inherit the same weekend-dip problem as a fixed percentage rule in GA4.

If you want a weekly summary rather than an alert, Looker Studio is a reasonable choice. See three ways to automate a GA4 weekly report email.

Option 3: A GA4 monitoring tool (Plask)

The third option is a tool whose only job is to watch your GA4 properties and tell you when something is off. Plask is one of these. It connects to GA4 with read-only OAuth, syncs each property once a day, and runs a statistical check on eight core metrics (active users, sessions, new users, screen views, engagement rate, revenue, transactions, and event count).

Here is what that looks like in practice:

• Sign in with Google and pick the properties to monitor. Plask backfills history so detection starts working within the first day.
• Every day at 06:00 UTC, Plask compares yesterday's value of each metric against the previous 28 days using a modified Z-score based on the median and Median Absolute Deviation (MAD). When enough history exists it compares against the same weekday, so a quiet Sunday is not flagged as a drop.
• Deviations are classified as info, warning, or critical. On the Pro plan, warning and critical alerts are emailed and every alert is posted to your Slack or Discord webhook. On the Free plan, all alerts appear in the dashboard bell.
• Each warning or critical alert includes an AI root cause analysis: Plask pulls the channel, device, and country breakdowns for the anomaly day versus the baseline and explains which segment moved. That is the "organic mobile traffic from the US fell 40%" line that GA4's email never gives you.
• Sensitivity is a per-property setting with three levels (Z-score thresholds of 4.0, 3.0, and 2.0), so a volatile hobby site and a steady SaaS marketing site can be tuned separately.

Comparison: GA4 custom insights vs Looker Studio vs Plask

Which one should you use?

One property, you already log into GA4 weekly: create two custom insights, one "Has anomaly" on Active users and one "% decrease more than 30" on Sessions evaluated daily. It costs nothing and covers the obvious failures.

You need a weekly PDF for a client or a manager: Looker Studio's scheduled delivery is the right tool, and it is free.

Several properties, or you live in Slack: use a monitoring tool. The cost of a missed drop on one of eight properties is usually more than $9 a month, and the root cause line saves the twenty minutes you would otherwise spend clicking through GA4 reports to find out what happened. Plask's free tier covers two properties with in-app alerts, so you can test the detection quality before paying for delivery channels. See how the alerts work on the anomaly alerts feature page or compare in detail on Plask vs GA4 custom alerts.