This is part two of a series on Stats Accelerator. In the first part, we explained the when, why, and how of Stats Accelerator. In today’s installment we will discuss a major roadblock to successful productionization of bandits in the A/B testing context and how we eventually overcame it. This is a high-level overview. For […]
This content originally appeared on Medium. What is Stats Accelerator and why use it? When a customer sets up an A/B/n test using Optimizely, they choose what percentage of visitors (or samples) will be exposed to each of the test’s variations. This is the test’s traffic allocation. The traffic allocation stays the same, unless a customer manually changes […]