Which description best captures the effect of a search-based split on user distribution?

• A. It evenly distributes users between existing and experimental campaigns
• B. It assigns all users to the experimental variant
• C. It assigns all users to the control
• D. It ignores user allocation and focuses on budgets

Answer: (A)

The idea behind a search-based split is to allocate traffic between a current (control) campaign and a new (experimental) campaign so you can compare how the changes perform. In this approach, a portion of users or search impressions goes to the experimental variant while the rest continues with the existing setup. That balanced distribution, often near 50/50, gives enough data in both groups to detect real differences and attribute any lift to the changes rather than random variation. If all users went to the experimental variant, there’d be no baseline to compare; if all stayed with the control, there’d be no test at all; and focusing only on budgets ignores the measurement needed to evaluate the variant. So the description that best captures the effect is distributing users between the existing and experimental campaigns.