Standard Variance

Description

Standard Variance measures how significantly a metric deviates from its configured threshold, offering a normalized measure of variation. It helps you assess the severity of deviations across different insights by standardizing the impact relative to the threshold. A higher value indicates a greater deviation.

Formula

% Standard Variance = ABS( (Insight Metric – Threshold)/ (Insight Metric + Threshold) )

Where,

• Insight Metric is the numerical value used to measure deviation for a given insight.
• Threshold is the reference value against which the insight metric is evaluated.

The definition of Insight Metric and Threshold varies based on the insight type.

For example, in Anomaly Insights, the metric is the Anomaly Value and the threshold is the Historical Mean.

Example

For a given Forecast Variance & Bias Insight, if MAPE = 10% and Threshold = 8%, then:

Standard Variance = ABS ((10-8) / (10+8)) = ABS (2/18) = 0.11 or 11%

Note:

In the Insights list, Impact is equal to Standard Variance.