# Root Cause Analysis: Track down anomalies in your data

> For AI agents: a documentation index is available at [/docs/llms.txt](/docs/llms.txt). Append `.md` to any page URL for markdown, or send `Accept: text/markdown`.

Root Cause Analysis (RCA) automatically finds the properties and user segments that explain an anomalous data point in a chart, so you don't have to manually pivot through breakdowns to track down what changed. RCA analyzes the properties of anomalous events, adds external context like country-specific holidays and product releases, and then explains the anomaly or rules out obvious causes.

Without RCA, finding that explanation isn't always easy. For example, you might need to navigate from _Platform > OS > Device family_ to discover that users on a specific device type drive an observed change. RCA surfaces that segment for you, so you can tell whether an anomalous data point is random noise or the beginning of a shift in user behavior.

Use RCA to answer questions like "Which user groups best explain this change?" or "How are other correlated metrics affected?"

- This feature supports Event Segmentation charts only. It doesn't support formulas.

## Analyze an anomalous data point

To use RCA, you must have an analysis in Amplitude that displays anomalous data.

For example, this Event Segmentation chart has an unusual-looking peak for December 30th:

![root_cause_analysis_1.png](https://amplitude.com/docs/images/analytics/root_cause_analysis_1.png)

To examine this anomalous data point using Root Cause Analysis, follow these steps:

1. Click [_Anomaly + Forecast_](https://amplitude.com/docs/analytics/anomaly-forecast) to confirm that the result you're interested in is actually a statistical anomaly. Amplitude enhances the chart to display the statistically-expected values for each day, as well as a range of values that would not be considered anomalous (in other words, values that could be attributable to random chance).

   ![root_cause_analysis_2.png](https://amplitude.com/docs/images/analytics/root_cause_analysis_2.png)

   In this example, the December 30 data point is inside the range of statistically normal values, but close to the edge.
2. Click on the data point to bring up the Microscope. Then click _Run Root Cause Analysis_. Scroll down to view the _Root Cause Analysis_ tab.

Amplitude scans the properties of the anomalous event and compares them to a baseline. It then generates a time-series graph for each property, so you can understand how each value tracks with the anomalous data point.

You can **expand** any of these graphs into a standalone Event Segmentation chart by clicking _Open Chart_ for the graph you're interested in.

You can also give **feedback** on the chart by clicking the thumbs-up or thumbs-down icon. This tells Amplitude whether you found the chart useful or not, and helps improve the ranking algorithm over time.

RCA scans event properties in batches of 30 to present the most relevant properties first. RCA automatically pauses scanning to avoid overwhelming you with graphs. To scan more properties, click _Continue Scanning_.

> **Tip:**
>
> Holidays are frequently the cause of anomalous data points. The Microscope alerts you if the anomaly you're investigating falls on a holiday.

## Configure your analysis

Amplitude automatically starts scanning with the thirty most-queried properties. If you'd like a specific event property to be included in future scans, click _Configure_ to open the _Configure Analysis_ modal. Then click _Select property..._ and select the property you're interested in from the list.