Amazon Lets You See the Shopping Guesses It Makes About You
Amazon maintains a Shopping preferences page that lists guesses about shoppers based on what they bought. It appears under "Manage your information" or "About you" depending on whether you use the website or the app. TechCrunch
The page drew wider attention after Threads user @fangirlinmegan posted screenshots of her own list. It included "has flat buttocks" alongside "Shops from women's departments" and "Probably owns a Shark robot vacuum." The post passed a million views in less than a day, according to the Oct. 6 report. She said the "flat buttocks" label likely came from buying "butt scrunch leggings."
The wording is blunt. These are not marketing lines written for the public. They are internal labels, created automatically when purchase history is sorted into traits about body, household, or interests. That process explains the awkward, overly literal phrasing.
On desktop, the path is Hello, [name] > Account > Ordering and shopping preferences > Your Shopping preferences > Manage your information. In the mobile app, the path is hamburger menu > Account > Shopping preferences > About you. Both lead to the same type of data: what Amazon thinks it knows based on what was bought, browsed, or otherwise linked to the account.
A separate control covers advertising. Amazon runs an Advertising Privacy and Preferences page that includes Interest-based ad preferences and Delete ad data controls. Amazon Users can choose "Show me interest-based ads provided by Amazon" or "Do not show me interest-based ads provided by Amazon." Amazon states that changes apply to the current device unless the user signs in to apply them to the account. The company also states it participates in the Digital Advertising Alliance and the Network Advertising Initiative, and offers a "Delete ad data" option described as deleting personal information linked to the device from its ad systems.
The broader context here is the split between account-level shopping guesses and device-level ad settings. Shopping preferences reflect a lasting profile tied to identity and order history. Ad preferences, as documented, work at least in part at the device layer, with account-wide effect only after sign-in. That distinction matters in practice. Deleting ad data on one browser does not rewrite the underlying purchase history.
Looking at what this means for transparency design, showing raw labels invites ridicule and scrutiny. It also gives users a concrete item they can see and correct. Hidden personalization avoids embarrassment but offers less to check. On this surface, at least, Amazon has opted for legibility, even when the label is crude or wrong.
In my view, the useful question is not whether any single guess is accurate. Automated guessing trades precision for coverage, meaning edge cases will always look strange out of context. The useful question is reversibility. Can a user see the label, understand which purchases led to it, and limit its use in ranking and targeting. The pages described here partly answer the first and third parts. Provenance, or the link back to specific orders, remains thin. Connecting leggings to a body-shape tag required user guesswork, not system explanation.
The practical move is simple. Check both surfaces. Review shopping assumptions for stale or incorrect items, then review ad preferences separately on each device in active use, signing in where account-wide application is intended.
