Topics
Discover how guardrails can make experimentation faster and safer.Understand the 4 levels of safety you can configure with guardrail metrics.Master how to protect the gamification of metrics by pairing a primary metric with a guardrail.Learn from case studies at Airbnb and Google for using guardrails.Audience
This talk is for you if:
You are worried about hurting crucial metrics while running tests.You know about guardrail metrics but feel you are not using them properly in testing.Leading metrics of your product show improvement, but still main business KPIs do notYou want to set up a system that allows your team to run more experiments.After this talk, you will be able to think creatively on how you can use guardrail metrics for different purposes in your experimentation program.
Hi! I am a statistician, a teacher and the founder of Thinking Bell.
I am known for simplifying experimentation with analogies, doodles, and games that make decisions sharper and science practical.
I have led data science research at VWO and I have authored the VWO Stats Blog. In 2025, I taught a statistics bootcamp for a team of 60 marketers, product managers and data scientists.
Join me to master experimentation with clarity, depth, and fun.
Stat-Talk #10
4 Ways To Select Guardrail Metrics
Automatically Stop Harmful Tests Early
Curious to Dive Deeper?
See Our Entire Course Curriculum
Join the 30-student cohort and master A/B testing stats through live classes, games, and community.