Yuzuu
Yuzuu
Role
Product Designer (Sole Design and UX Lead)
Scope
Search Architecture, Filtering Systems, Interaction Design, Visual Specs
Status
Archival case study, platform inactive
Overview
Overview
The London rental market is fragmented. Renters kept five to ten tabs open at once, switching between Rightmove, Zoopla, OpenRent, SpareRoom and independent agents, fighting duplicate listings and search criteria that meant something different on every site.
The London rental market is fragmented. Renters kept five to ten tabs open at once, switching between Rightmove, Zoopla, OpenRent, SpareRoom and independent agents, fighting duplicate listings and search criteria that meant something different on every site.
Yuzuu indexed every London rental in real time on one discovery canvas. Filters were applied once, and routing out to the originating platform to apply stayed seamless.
Yuzuu indexed every London rental in real time on one discovery canvas. Filters were applied once, and routing out to the originating platform to apply stayed seamless.
Approach
One canvas, every source
One canvas, every source
Three problems had to be solved before a single index could feel like one product rather than a scraper.
Three problems had to be solved before a single index could feel like one product rather than a scraper.
Universal filters
Universal filters
A normalisation layer mapped disparate listing tags from every platform onto one set of criteria: bills included, furnished status, commute distance, outdoor space, pet policies. Price, bedrooms and location stayed in a persistent top bar; everything granular moved into an expandable drawer.
A normalisation layer mapped disparate listing tags from every platform onto one set of criteria: bills included, furnished status, commute distance, outdoor space, pet policies. Price, bedrooms and location stayed in a persistent top bar; everything granular moved into an expandable drawer.
Source routing
Source routing
Every card carried a source badge, so it was always clear which platform a listing originated on, and clear visual signifiers marked whether an action would open an external window or keep the user on the map view.
Every card carried a source badge, so it was always clear which platform a listing originated on, and clear visual signifiers marked whether an action would open an external window or keep the user on the map view.
Zero states
Zero states
Over-restrictive filter combinations returned nothing, so the search offered relaxations instead of a dead end: no exact matches within half a mile, showing twelve listings within one mile.
Over-restrictive filter combinations returned nothing, so the search offered relaxations instead of a dead end: no exact matches within half a mile, showing twelve listings within one mile.
To make searching across millions of listings effortless, we built a smart autocomplete engine that maps natural language text or partial keywords directly to complex geographic boundaries. Instead of forcing users to navigate dropdown menus or manually draw shapes, typing a simple phrase like “Shoreditch” or “near King’s Cross” instantly resolves into precise coordinate bounds, multi-postcode clusters, and surrounding transit zones behind the scenes—turning a chaotic multi-site search into a single, lightning-fast text input.
To make searching across millions of listings effortless, we built a smart autocomplete engine that maps natural language text or partial keywords directly to complex geographic boundaries. Instead of forcing users to navigate dropdown menus or manually draw shapes, typing a simple phrase like “Shoreditch” or “near King’s Cross” instantly resolves into precise coordinate bounds, multi-postcode clusters, and surrounding transit zones behind the scenes—turning a chaotic multi-site search into a single, lightning-fast text input.
