Amazon Closet

Amazon Closet was a personalized wardrobe feature inside the Amazon Shopping app. You could catalogue the clothes you had already bought and get style recommendations tuned to your preferences, your price range and your purchase history — a curated, editorial feel inside Amazon's apparel shopping.

I served as tech lead for Amazon Closet, shipping the feature on both iOS and Android inside the Amazon Shopping app.

How It Worked

It lived under Programs and Features. It read your clothing purchase history to generate outfit and apparel recommendations, and you could add items from Amazon's catalogue by hand to sharpen your style profile. The result was a recommendation experience tuned for fashion rather than Amazon's general "you might also like" engine.

The premise is that a recommendation engine that knows what you own can do things one that only knows what you bought cannot: build outfits out of items already in your closet, and understand a new item in the context of the rest of your wardrobe.

Before it was a feature

The research and prototyping that established whether any of this would work is written up separately as Virtual Closet — a React prototype run against real customer cohorts, and the user research that showed the genuinely hard problem wasn't recommendations at all. It was import logic: deciding which of someone's purchases are actually theirs, actually kept, and actually still in the closet.