Summary Matrix
| Attribute | Control (Version A) | Variation (Version B) |
|---|---|---|
| Role | The existing baseline experience | The experimental new experience |
| Traffic Allocation | Typically 50% of traffic | Typically 50% of traffic |
| Hypothesis Target | Represents current performance benchmark | Tests specific changes proposed in hypothesis |
Key Difference
The Control acts as the baseline benchmark to prove whether the Variation causes a statistically significant improvement or degradation.
Developer Takeaway
Clean Code Cleanup: Once an experiment concludes and a winner is declared, archive the experiment code and merge the winning variation cleanly into your main codebase. Never leave inactive A/B test toggles accumulating in production code.