Amazon's Shopping Profiles Include Hilariously Specific Guesses About Who You Are

Amazon's personalization engine isn't just tracking what you buy—it's making surprisingly personal inferences about who you are as a person. A viral thread on the social platform Threads has exposed the detailed descriptions Amazon maintains in users' profiles, revealing assumptions that range from hilarious to uncomfortable.

What Amazon Thinks It Knows#

The discovery began when one user found that Amazon's systems had categorized her as someone with "flat buttocks," sparking a wave of others sharing their own profiles. The inventory of assumptions is staggering in its specificity. Users have documented being labeled as people who "don't trust easily," "find it hilarious when toys get stuck to ceiling tiles in the classroom," "cuss when scared," "cannot keep real plants alive," "have no friends," and "treat cats as biological children."

Some of these micro-profiles veer into the comical—the sheer randomness of ceiling-tile toys or the oddness of a butt-focused category suggests Amazon's machine learning is pattern-matching rather than actually understanding human psychology. Yet others hit surprisingly close to reality.

How to Check Your Profile#

Amazon makes this data visible to users, though it's tucked away. Head to your account settings by clicking your profile icon in the top-right corner, then navigate to Your Account > Your Shopping preferences > About You. There you'll find the full roster of behavioral tags Amazon has assigned to you. Better still, you're not stuck with these labels—each one can be edited or deleted directly.

The Accuracy Question#

For some users, the accuracy is genuine. One reviewer's profile correctly identified habits like maintaining aquarium systems, drinking coffee regularly, and playing video games—all verified by their actual behavior. Others found the descriptions were fairly generic marketing categories. The real question is whether these intimate guesses affect ad targeting and product recommendations in ways users should be concerned about.