The Outliers feature helps you quickly identify statistically significant events in your data over time. It’s available in all timeline widgets across both the Workbench and Dashboard.
What is an Outlier?
An outlier is a data point that stands out significantly from the usual pattern. Specifically, we use z-scores to detect when a value deviates from the average trend within the time period currently shown on your chart.
For example, if the frequency of a theme increases dramatically compared to nearby months, it will be flagged as an outlier.
Where Does This Apply?
The Outliers feature is automatically enabled for the following timeline-based widgets:
Score (e.g., CSAT, custom metrics)
NPS
Sentiment
Timeline
You’ll find these widgets in both:
Workbench
Dashboard
How to Use It
Showing Outliers
Any chart that has outliers available will have an expandable section just above the chart area. Simply click "Show Outliers" to expand it and see outliers based on the chart's data.
Highlighting Outliers
Hover over an outlier to see it highlighted in context on the cart.
Interacting with Outliers
Click an outlier to open up interaction options, such as:
Applying a date filter to narrow down the view to that specific period
Applying a segment filter related to the spike or drop
Drilling into a theme to explore what might have caused the change
Pagination
Up to 3 outliers are shown at a time to avoid clutter.
Use the pagination controls to browse additional outliers in the chart.
💡 Tip
If you don’t immediately see outliers, try adjusting your date range. Outliers are calculated relative to the visible time window.