A/B Testing Interview Questions

A/B Testing Interview Questions

A/B testing questions are central to data science and product analytics interviews at companies like Meta, Google, Netflix, and Airbnb.

Expect questions on experiment design, randomization units, sample size calculation, multiple comparisons, and metric selection.

Interviewers evaluate your statistical rigor, practical judgment, and ability to communicate experiment results.

871Questions
85Companies
Easy 102 • Medium 423 • Hard 346Difficulty mix

Common A/B testing interview patterns

  • Designing an experiment for a product change
  • Calculating sample size and experiment duration
  • Choosing between one-sided and two-sided tests
  • Handling multiple comparisons and peeking
  • Interpreting results with novelty or primacy effects
  • Network effects and interference between test groups

A/B testing interview questions

Filter and sort
All SQL questions

Common mistakes in A/B testing interviews

  • Not specifying the randomization unit (user vs session vs page)
  • Peeking at results before reaching the required sample size
  • Ignoring practical significance when statistical significance is achieved
  • Not considering guardrail metrics
  • Failing to account for novelty effects in short experiments

How A/B testing questions are evaluated

Structure your experiment design: hypothesis, metrics, unit, sample size, duration.

Discuss what could go wrong and how you would detect it.

Show ability to make a recommendation even when results are ambiguous.

Related analytics concepts

A/B Testing Interview FAQs

How do you determine the sample size for an A/B test?

Use a power analysis with inputs: baseline metric, minimum detectable effect (MDE), significance level (alpha, usually 0.05), and power (usually 0.80). Larger effects need fewer samples. For small MDE on rare events, you may need millions of users.

What is the difference between statistical and practical significance?

Statistical significance means the observed difference is unlikely due to chance (p-value < alpha). Practical significance means the effect is large enough to matter for the business. A statistically significant 0.01% lift may not be worth the engineering cost to ship.