Anomaly Alerts
Never miss a significant change in your analytics. Get alerted when metrics deviate from their expected patterns — via in-app notifications, email, Slack, or Discord.
What are Anomaly Alerts?
Anomaly Alerts use statistical analysis to continuously monitor your GA4 metrics during each daily sync. When a metric deviates significantly from its recent trend, Plask flags it as an anomaly, assigns a severity level, and notifies you through your configured channels. This means you catch traffic spikes, revenue drops, or engagement changes before they become problems — without manually checking Google Analytics every day.
How It Works
Modified Z-Score with Median Absolute Deviation
Each metric is analyzed over a 28-day rolling window using the modified Z-score method. This approach uses the median and Median Absolute Deviation (MAD) instead of mean and standard deviation, making it robust to existing outliers in your data. When today's value produces a Z-score that exceeds the configured threshold, an alert is triggered and classified by the magnitude of the deviation.
For example, if your daily active users typically range around 1,200 and suddenly drop to 640 — a 47% decline — the Z-score might hit -3.2, which exceeds the default threshold of 3.0. Plask flags this as a warning-level anomaly and notifies you immediately through your configured channels.
Info
Moderate deviation from trend
Warning
Significant deviation
Critical
Extreme deviation from trend
Configurable Sensitivity
Not every property has the same traffic patterns. A small blog with volatile daily traffic needs different thresholds than a high-volume e-commerce site. Plask lets you configure sensitivity per property so you get the right balance between catching real issues and avoiding noise.
Low
Z-score threshold: 4.0
Only extreme deviations trigger alerts
Default
Z-score threshold: 3.0
Balanced detection for most properties
High
Z-score threshold: 2.0
Catches smaller deviations early
Smart Alert Prioritization
When multiple alerts fire at once, Plask ranks them using an AI-powered priority score. The score combines three factors: severity (the magnitude of the Z-score deviation), recency (newer alerts rank higher), and metric importance (revenue and user-count metrics are weighted higher than engagement metrics). Alerts are displayed with unread items first, sorted by priority score, so you always see the most important issues at the top.
What Metrics Are Monitored?
Plask monitors 8 core GA4 metrics during each daily sync. Every metric is independently tracked with its own 28-day baseline, so an anomaly in one metric does not affect detection in others.
Notification Channels
Choose where you receive alerts. Free users get in-app notifications. Pro users can add email, Slack, and Discord delivery so the right people see anomalies without logging into Plask.
In-App Notifications
View all alerts from the dashboard bell icon. Alerts are sorted by priority with unread items surfaced first. Available on the Free plan.
FreeEmail Alerts
Receive instant email notifications for warning and critical alerts. Emails include the metric name, deviation magnitude, and AI-generated root cause analysis.
ProSlack Webhooks
Post anomaly alerts to any Slack channel via incoming webhooks. Your team sees alerts in real time without checking the dashboard. Webhook URLs are encrypted with AES-256-GCM.
ProDiscord Webhooks
Send alerts to your Discord server via webhook integration. Configure once in settings and receive formatted alert messages in the channel of your choice.
ProFrequently asked questions
- How does Plask detect anomalies?
- Plask uses modified Z-scores calculated with Median Absolute Deviation (MAD) over a 28-day rolling window. Unlike standard deviation, MAD is resistant to distortion from existing outliers, making it more reliable for real-world analytics data. When a metric's Z-score exceeds the configured threshold, an alert is generated and classified by severity.
- Can I adjust alert sensitivity?
- Yes. Each GA4 property has its own sensitivity setting with three levels: low (Z-score threshold of 4.0, fewer alerts), default (3.0, balanced), and high (2.0, more sensitive). You can change sensitivity at any time from the property settings page. Higher sensitivity catches smaller deviations but may produce more alerts for volatile properties.
- What notification channels are supported?
- Plask supports four notification channels: in-app notifications visible from the dashboard bell icon, email alerts delivered to your inbox, Slack webhooks that post to a channel of your choice, and Discord webhooks for server notifications. In-app alerts are available on the Free plan. Email, Slack, and Discord require the Pro plan.
- Are alerts available on the free plan?
- Yes. In-app anomaly alerts are included in the Free plan for up to 2 GA4 properties. You can view all alerts in your dashboard, see severity levels, and mark them as read. Email delivery, Slack webhooks, and Discord webhooks require the Pro plan at $9/mo or $79/yr.
- How are alerts prioritized?
- Plask uses AI-powered smart prioritization that scores each alert based on three factors: severity (the magnitude of the Z-score deviation), recency (how recently the anomaly was detected), and metric importance (revenue and user metrics are weighted higher than engagement metrics). Alerts are sorted by priority score with unread alerts surfaced first.
Start detecting anomalies today
Connect your GA4 properties and get anomaly alerts within 24 hours.