Hi, I'm Shan.
I'm a data scientist who enjoys solving difficult problems, building useful things and exploring ideas. My work spans pricing, forecasting and experimentation, with a focus on making data useful in the real world.
A/B Testing Is a Decision Problem
How much evidence do we need to make a decision, and what does it cost us to be wrong?
Read the articleLatest writing
Experimentation in the Real World A three-part series
- Part 1 of 3Beyond peeking: Making decisions during A/B tests False-positive risk, outcome maturity and commercial opportunity when an A/B test shows an early signal.
- Part 2 of 3Fixed-power designs: How much data do we really need? How a fixed-power design adapts the sample requirement as mature outcomes reveal the actual variability.
- Part 3 of 3Can we roll out early and keep learning? How a protected comparison can keep learning while a separate population receives a promising treatment.