Protocol for Auditing Algorithmic Bias in the Recommendation Systems of Shopify Stores (v2.2)

Open protocol of AVBR Lab for the external (black-box) audit of product-recommendation blocks on Shopify stores: sampling rules, seed selection (A/B/C), capture procedure, the visibility rule (unit of observation), metric definitions and the coding scheme for visual representation and accessibility. Version 2.2, September 2026 (edited 15 October 2026; RQ2 wording of 2 November 2026). Price results are reported descriptively (see the lab's update of 31 October 2026: a within-store null model reproduces the percentage gradient, so no steering is inferred). Released together with the papers of the research program; figures cited elsewhere are pre-publication. Author of the methodology: Polina Hneletska, Founder & Research Director, AVBR Lab (American Visual & Behavioral Research Lab Inc., a California nonprofit public benefit corporation). License: the text is CC BY 4.0. The audit method is the subject of pending U.S. patent applications (Nos. 64/040,345 and 64/041,084); commercial use of the method requires a license from AVBR Lab; research and verification use is free. Contact: [email protected].