A control group is a set of users who do not receive a tested change or intervention. Their behaviour is used as a baseline for comparison.
Control groups are essential for isolating impact.
Examples in Action
- Comparing test variants against unchanged experiences
- Holding back a portion of traffic from an intervention
- Measuring natural behaviour without influence
- Validating causal relationships
Typical Outcomes / Results
- More reliable experiment results
- Clear separation of signal from noise
- Reduced bias in interpretation
- Stronger confidence in conclusions
This glossary entry reflects standard experimental design principles.