Bazinga!Bazinga!

The idea

Bazinga applies a swipe interface to fashion retail: browse a feed of clothing, swipe to save or dismiss, and build a picture of someone's taste from their gestures rather than from a preferences form they will never fill in.

Why swipe is harder than it looks in commerce

Swiping is effortless, and that is exactly the problem. A gesture that costs the user nothing also tells you very little - people swipe quickly, inattentively, and often on a colour rather than a garment. Treating every swipe as a strong preference signal produces a recommendation model that confidently learns nonsense.

We weighted signals by the effort behind them. A long look at a product detail page counts for far more than a fast dismissal; an add-to-bag outweighs a hundred swipes. The feed stays effortless while the model underneath listens mostly to deliberate actions.

What we built

• Swipe-driven discovery feed with a preference model behind it
• Product detail, sizing and bag flow
• Saved items and wishlist persisting across sessions
• Image pipeline tuned for mobile data
• Checkout with guest purchase

The detail that mattered most

Image loading. A fashion feed is almost entirely photography, and on a mobile connection a feed that stutters is a feed people leave. Progressive loading, careful size negotiation and prefetching the next few cards did more for engagement than any change to the recommendation logic.