Experimentation
Foundationexperimentation
Optimizely is the industry leader in experimentation, offering both client-side visual testing and server-side feature experimentation. This section explains how experimentation works, the differences between Web and Feature Experimentation, and how to choose the right approach for your use case.
Experimentation in Optimizely Understand why experimentation matters, how Optimizely's experimentation products work, and how to choose between Feature and Web Experimentation.
Feature Experimentation Understand how Feature Experimentation works — feature flags, progressive rollouts, server-side experiments, and SDK integration across platforms.
Web Experimentation Understand how Web Experimentation works — visual A/B testing, multivariate tests, personalization campaigns, and the Stats Engine.
SDK Architecture Understand how Optimizely's server-side SDKs and client-side snippet evaluate experiments locally for sub-millisecond decisions.
Event Dispatching and Analytics Learn how conversion events flow from SDKs and snippets to Optimizely analytics, including batching, retry logic, and custom dispatchers.
User Context and Bucketing Understand how Optimizely consistently assigns users to experiment variations using MurmurHash3, traffic allocation, and sticky bucketing.
CMAB: Contextual Multi-Armed Bandit Learn how CMAB automatically optimizes variation assignment using machine learning to maximize conversions during experiments.
The Visual Editor Understand how the visual editor lets marketers create experiment variations without code, and where its boundaries are.
Snippet Deployment Models Compare CDN hosting, custom domains, and self-hosting for the Optimizely snippet, and understand sync vs async loading trade-offs.
Stats Engine: How Statistical Decisions Work Understand sequential testing, always-valid p-values, false discovery rate control, and CUPED variance reduction in Optimizely's Stats Engine.
Performance Edge Learn how Performance Edge eliminates client-side flicker by moving variation assignment to the CDN edge.
Experiment Types: When to Use What Compare A/B tests, multivariate tests, multi-armed bandits, and CMAB to choose the right experiment type for your scenario.