Privacy Regulation and Online Retailer Resilience: The Roles of Customer-Acquisition Dependence and Traffic Architecture

This study investigates whether the share of online retailers’ traffic and sales attributed to Meta fell after Apple’s App Tracking Transparency (ATT) restricted identifier-based advertising, and whether revenue fell with it. A continuous-exposure difference-in-differences design on a balanced monthly retailer panel shows that, after enforcement, retailers with greater pre-shock Meta reliance drew a smaller share of sessions, orders, and revenue from Meta-mediated traffic than less reliant retailers. Because the data record shares and do not separate paid from organic referrals, the decline does not by itself show that fewer customers were acquired through Meta. The decline widened over time and is stable across post-surge window starts, exposure measures, and category and size shocks; its attribution to ATT rests on the parallel-trends assumption, which the short pre-enforcement period tests only in part, and which retailer-specific trends and sensitivity bounds leave open. Among retailers active throughout the pre-shock year, revenue and orders in months with sales show no statistically detectable relative decline; equivalence tests bound these responses within a few percent but cannot distinguish full replacement of the contracted channel from a smaller undetected loss. The larger revenue decline reported in prior work, which compares above- and below-median exposure on a transformed outcome across all firms, arises in the same data from months without sales among mostly small retailers. The desktop-web share does not detectably moderate the decline in Meta shares.