Store activity and privacy
We use server-observed store activity to understand which shopping steps are reached and improve the store. The observed steps are product-page requests, buy or add-to-cart requests, redirects to sign-in, and checkout attempts.
This collector briefly processes the request IP address and browser user-agent in memory to create a pseudonymous daily bucket. It does not set analytics cookies or retain the raw IP address or user-agent in its records. Buckets are separate for each site and are not joined across days. Shared devices and networks can combine activity, and one visitor can appear in several buckets.
Stage state expires after six hours of inactivity, Redis keys for daily hashing salts after 48 hours, and aggregate counters after 730 days. When a checkout is created, we may retain an order-associated snapshot of available prior stages, with the trusted order ID, capture and stage times, purchase lane, language, chosen country shelf, device category and crawler classification for up to 90 days. Each store keeps at most 10,000 such snapshots, so older records may be removed sooner.
Order-associated snapshots are private commerce data, not anonymous records. Authorized internal tools can read aggregate paths and selected checkout outcomes, including payment status, product, currency, order value and available delivery status. This helps staff and their authorized AI tools assess the store. A missing step does not prove that a visitor left or that a sale was lost.
This collector does not record query strings, form text, contact details, dictionary entries, research queries or private application content. Reports cover supported store routes only. Collection can miss events during overload or outages.