> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lunarmc.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Outliers Overview

## Overview

**Outlier Management** is SuperOrbital's anomaly detection and notification system. It allows you to automatically monitor important client KPIs and identify significant changes in performance without manually checking reports every day.

Outliers can monitor standard platform metrics as well as **Custom Filtered and Calculated Metrics**, making it possible to create alerts around the KPIs that matter most to each client.

When an Outlier is detected, SuperOrbital creates a Task for review and can optionally send your team a Slack notification.

***

## How Outliers Work

Each Outlier defines a metric to monitor, the period of performance to analyze, and what that performance should be compared against.

You can compare performance against:

* **Prior Performance** – Identify unusual changes by comparing current performance against a previous day, week, month, or other historical period.
* **Forecasts or Goals** – Identify when actual performance is significantly above or below a predefined client target.

For example, Outliers can help your team identify:

* Unexpected increases or decreases in Ad Spend
* Significant Revenue changes
* Budget over- or under-pacing
* Revenue target misses
* Unexpected changes in CPA, MER, ROAS, or other KPIs

You can also configure different severity thresholds to distinguish between smaller warnings and more significant performance anomalies.

***

## Outlier Notifications & Tasks

When performance exceeds your configured threshold, SuperOrbital automatically creates an **Outlier Task** for your team to review.

You can also configure **Slack notifications** based on severity, allowing the appropriate users or channels to receive alerts when important performance changes occur.

Slack notifications provide a snapshot of the anomaly, including the **Severity, Current Metric Value, and Baseline Value**.

From the notification, you can open the Outlier directly in SuperOrbital or **Ask Houston** for a quick analysis of what happened and recommendations for next steps.

***

## From Detection to Action

Detected Outliers are sent to the **PMT Queue** for review.

From there, your team can determine whether the anomaly requires action. Outliers that don't require follow-up can be rejected, while actionable Outliers can be approved and added to PMT.

This creates a workflow from **automated detection → notification → investigation → action**, helping teams identify important performance changes quickly and turn them into actionable next steps.