A/B Testing
A/B(n) testing is a method that compares two or more versions of a web page, an application, an algorithm, advertising banners, etc., to determine which performs best.
An experimental method that involves comparing two or more versions (A, B, C, etc.) of a digital element – web page, application, algorithm, advertising banner, email, etc. – in order to determine which version performs best according to one or more defined metrics (conversion rate, clicks, revenue, engagement...).
The different variants are randomly presented to users, then analyzed using rigorous statistical methods (frequentist or Bayesian) to ensure that the observed results are significant and reliable.
Unlike correlative or retrospective analyses, A/B testing makes it possible to establish a causal relationship: it precisely isolates the impact of a change on user behavior, while controlling for external variables.
The A/B(n) approach is at the core of Conversion Rate Optimization (CRO) strategies, as it enables:
- continuous improvement of a site or product's performance,
- limiting risks associated with untested deployments,
- prioritizing changes that bring measurable value,
- validating decisions on empirical and data-driven grounds.
In practice:
- An A/B test compares two variants (A vs B),
- An A/B/n test can compare three or more versions (A vs B vs C...), while maintaining random distribution and consistent statistical analysis.
This method is commonly used in e-commerce, SaaS, media, or mobile app contexts, and is integrated into tools such as AB Tasty, Optimizely, VWO, Google Optimize (former), or Dynamic Yield.
Autres termes : A/B testing
View the full glossary- AA TestA type of test where two groups receive the same experience to ensure that the testing system is reliable and that observed differences are not due to chance.A/B testing
- Client-side testing (A/B test)Client-side testing modifies the content visible to the user via scripts executed in the browser.A/B testing
- DatalayerA data structure used to transmit information between a website and a tag management tool.A/B testing
- Feature FlaggingA technique for enabling or disabling features remotely without deploying new code.A/B testing
- Feature RolloutThe progressive rollout of a new feature to a subset of users before a full launch.A/B testing
- Test hypothesisA test hypothesis is a statement formulated to predict the outcome of a specific change on a website or application as part of an optimization process.A/B testing