Treatment Group

Treatment Group is the set of users in an A/B test who are exposed to the new variation or experimental condition being tested, as opposed to the control group which sees the original version.

Meaning & Context

Also called the experimental group or variant group, the treatment group receives the intervention you're testing—whether that's a new design, different copy, altered pricing, or any other change. In experiments with multiple variations, you may have several treatment groups, each experiencing a different version. The treatment group's performance is compared against the control group to measure the impact of the change you've implemented.

Why It Matters

Properly defining and isolating treatment groups ensures you can accurately attribute performance differences to the changes being tested rather than external factors. Treatment groups must be assigned randomly and simultaneously with control groups to avoid selection bias and temporal effects. The size of your treatment group affects statistical power and how quickly you can reach significant results, with larger groups providing more precise effect estimates.

Example

In testing a new product page layout, you randomly assign 50% of visitors to the treatment group who see the redesigned page with larger product images and repositioned reviews, while the other 50% (control group) see the current layout. After two weeks, you compare conversion rates between the groups.