Overview

Amazon has refreshed the “Manage Your Experiments” (MYE) interface, adding pre‑filled optimal settings, automatic termination when statistical significance is reached, and instant publishing of the winning variant. The overhaul removes manual configuration steps and speeds up the rollout of successful changes, a development that directly affects brand‑registered sellers who rely on A/B testing to improve listings.

Key Points

• Pre‑Selected Optimal Settings — New experiments launch with Amazon‑recommended parameters already applied, so sellers no longer need to set traffic split, duration, or confidence thresholds manually.
• Run‑to‑Significance Logic — Tests end automatically once enough shopper interactions have been collected to declare a clear winner, eliminating the need to wait for a preset end date.
• Auto‑Publish Winning Content — The variant that meets the significance criteria is applied to the live listing without any additional seller action, cutting down post‑test latency.
• Advanced Settings Toggle — Sellers who prefer full control can switch to an advanced mode to define custom test length, traffic allocation, and conclusion rules.
• Higher Test Velocity — Because experiments no longer occupy a fixed eight‑week slot, multiple sequential tests can be run on the same ASIN, accelerating optimization cycles.
• Reduced Setup Friction — The streamlined workflow requires only the new content upload, allowing sellers to start a test in minutes rather than hours.

How the New Manage Your Experiments Workflow Operates

1. Create Experiment with Default Parameters — The seller selects the ASIN, uploads the B‑version (e.g., a new main image), and clicks “Launch.” Amazon automatically assigns a 50/50 traffic split, a 95 % confidence level, and a dynamic duration that will end when significance is achieved. Example: A seller testing a new hero image for a kitchen gadget simply uploads the image and launches; no dropdowns for traffic or end date appear.
2. Adaptive Data Collection — As shoppers view the product, Amazon continuously monitors conversion, click‑through, and other performance metrics. When the algorithm determines that the observed difference meets the confidence threshold, the test stops.