Using the confidence index in an experiment

The confidence index percentage indicates how confident you can be when evaluating the results of an experiment. Knowing how confident you can be with the experiment results is important before setting the winning variant for all the traffic or making other data-driven decisions.

A high confidence index percentage has a greater chance of containing the true value. Marketers typically find an interval representing a 95% likelihood of capturing the true value to represent a good compromise between sensitivity and error. The Confidence Index % that displays on the Performance screen is only meaningful after the minimum sample size is reached.