The Experiment Journal is an independent editorial desk for people who run A/B tests and care about statistical honesty.
Every entry is written as a lab notebook. It records the same six fields up front — the hypothesis, the metric under study, the test design, the sample-size consideration, how to read the result, and what threatens validity — so readers can line two experiments up side by side instead of comparing a headline number to a feeling.
We describe statistical and experimental concepts carefully, avoid invented statistics and named sources, and stress that outcomes depend on your own setup and sample size. Everything here is general education, not business, financial, or engineering advice for any specific situation.