How to use experimentation in the back-end
summary
Server-side experimentation yields bigger wins because it enables complex code changes, feature flags and staged rollouts instead of only cosmetic front-end tweaks. Data shows 51% of experiments use 1–5 lines of code, while more lines double the chance of a winner. Running 5+ variations increases win rate by 75% versus plain A/B. A server-side algorithm swap produced 24% higher click-to-sale, 30% add-to-cart and 210% conversion lift in one example. Use painted-door tests, kill switches and targeted rollouts.
AI-written summary · about
chapters (10)
- 0:00Start
- 20:49Painted Door Tests
- 24:00Painted Door Cautions
- 27:14Feature Flags
- 28:14Feature Tests
- 30:11Non-Negative Experiments
- 31:41Staged Rollouts
- 32:28Dark Launches
- 33:08Canary Releases
- 33:30Kill Switch
related talks
24mEffective Product Experimentation: Using Data-Driven Techniques for Iteration
35mExperimentation for Engineers
28mExperimenting your way to a better product
36mA/B Testing in Production
28mOptimizing your A/B testing roadmap
27mExperimentation in BigTech Validating Hypotheses and Driving Innovation
33mA/B testing: performative experimentation at scale
01h 15mExperimentation at scale: Fuel your application’s growth through A/B testing