Helping 1.2 million monthly shoppers compare prices faster across the web while improving revenue and product click-through.
Realry helps 1.2 million monthly shoppers compare prices across the web. I redesigned the journey to make that value easier to reach.
My focus was navigation and conversion: understand why users were dropping off, make product discovery feel familiar, and move more shoppers confidently toward a partner store.
Users were leaving before reaching products because the AI feature dominated the hero while the core shopping experience sat below the fold.
Search terms unexpectedly launched an AI conversation, interrupting a familiar shopping behavior and creating a major drop-off point.
A non-standard category structure made browsing feel uncertain and repeatedly sent shoppers to dead ends.
Realry's most valuable feature was difficult to see, while prescriptive recommendations actively discouraged conversion.
6 weeks
A full platform redesign ahead of an investment pitch.
Web + iOS
Traffic was split almost evenly across desktop and mobile.
Built in parallel
Reusable components had to support both products from day one.
Analytics review
Honey, ShopBack
Daily CEO reviews
IBM Carbon foundation
Close developer pairing
Measure and iterate
I moved the AI experience into a floating entry point so it could stay visible without blocking shopping. The homepage gained intent-based collections and a larger brand showcase, giving users more relevant ways into the catalog.
Product search returned to a familiar, persistent pattern. Suggestions now surface products and stores directly, turning a point of friction into an additional path to partner revenue.
A persistent navigation menu replaced ambiguous category pills. The overlay makes location and hierarchy clear, while a new store directory supports shoppers who browse by retailer rather than product type.
Best-price badges and a complete list of offers made comparison immediate. I replaced blocking recommendations with neutral market context, helping shoppers understand the tradeoff and choose for themselves.
Measured through a three-week A/B test.
+17%
+23%
+39%
+50%
+28%
-10%
Average session duration fell by 15% while click-out increased. For a shopping aggregator, helping people find the right option faster is a better signal than keeping them on the site.
I designed a detailed advanced filter, but heatmaps showed fewer than 4% of users touched it. A smaller MVP would have tested the need sooner and protected design time.
Popular products drove more click-outs, but some users returned within seconds when inventory was unavailable. The next product opportunity is real-time stock visibility.