What is an A/B test (or split test)?

• A. An online test comparing two variants to determine which performs better, usually based on impressions
• B. An online test comparing two variants over several campaigns
• C. An online test comparing two variants to determine which performs better, usually based on conversions
• D. An online test comparing two variants to determine which performs better, usually based on the number of clicks

Answer: (D)

A/B testing is an online experiment where two variants are shown to similar audiences to see which one yields better results on a chosen action. The reason the metric of clicks is the best fit here is that it provides a direct, observable signal of user engagement that’s easy to measure across variants. It focuses on which version prompts more user interaction, which is the core goal of a split test: determine which variant drives a stronger immediate response. While conversions are also a meaningful outcome in many contexts, the essence of an A/B test is comparing behavior across variants using a measurable action, and clicks are a clear, common example of that action. Impressions don’t reflect performance, testing across multiple campaigns can introduce confounding factors, and basing on conversions narrows the metric to a specific end result rather than the general comparison of two variants.