Uplift

numeric

Sample size for a 1-point conversion lift

Experimentation & A/B Testing · Design & Power · medium · Original Uplift question · 8 min

Baseline conversion is 10%. You want to detect a 1 percentage-point absolute lift (10% -> 11%) at 80% power and 5% significance (two-sided). Roughly how many users per arm do you need? Use the rule of thumb n ≈ 16 · p(1−p) / δ².

Evaluation criteria

  • Uses the correct method or distribution
  • Shows the calculation
  • Rounds as the prompt asks

Detailed grading feedback appears after you submit.

Source

Original Uplift question based on a common interview pattern.

Your answer

users per arm

Commas, currency symbols and a trailing percent sign are fine. Answers within ±1500 users per arm of the expected value are accepted.

You get one graded attempt here without an account. Create a free account for unlimited attempts, saved progress and community write-ups.

Answers stay hidden until you ask. Reveal the full model write-up when you are ready to compare.

Community

How other people solved this

There is rarely one right query or one right framing. Read how others approached it, and share your own once you have submitted an answer.

Reading is open to everyone. Sign in to share your approach or mark a write-up as helpful.

Loading write-ups…