## Metadata - Author: [[marton-trencseni|Marton Trencseni]] - Full Title:: Five Ways to Reduce Variance in a/B Testing - Category:: #🗞️Articles - Document Tags:: [[Hypothesis testing|Hypothesis Testing]], - URL:: https://bytepawn.com/five-ways-to-reduce-variance-in-ab-testing.html - Finished date:: [[2024-09-30]] ## Highlights > it's worth knowing that in terms of the best lift measurement (lowest variance), the most optimal split is the even split of 50%-50%, and a 20%-80% split is better than a 10%-90%. We can see this with simple math. ([View Highlight](https://read.readwise.io/read/01j8zp0x3fynbfyshp9056zm8t)) >  ([View Highlight](https://read.readwise.io/read/01j8zp0je3m08x8tqmn1qzh55q)) >  ([View Highlight](https://read.readwise.io/read/01j8zp0jf0k2698da1b61mdv88)) > 1. [Winsorizing](https://en.wikipedia.org/wiki/Winsorizing), ie. cutting or normalizing outliers. ([View Highlight](https://read.readwise.io/read/01j8zp3pppk6dcyeavt8htwtwn)) > 2. Measuring a metric further down the funnel. ([View Highlight](https://read.readwise.io/read/01j8zp3sxfv8c13ztcpt3749mx))