CASE STUDY: REDESIGN
*A comprehensive refactoring of Qatar’s premier vacation rental marketplace to improve search heuristics and conversion rates.*
UX/UI Design & Frontend Dev
React, Mapbox GL JS, Figma
Initial telemetry and user session recording revealed critical friction points in the legacy architecture.
The legacy SQL queries for geospatial filtering resulted in Time to First Byte (TTFB) exceeding 2.4s, causing a 65% bounce rate on the search results page before map tiles rendered.
Touch targets on the booking calendar were below 44px, and the map interface lacked gesture handling, making it nearly impossible for mobile users (70% of traffic) to filter by location.
Listing pages lacked a clear information hierarchy. Critical amenities and cancellation policies were buried in unstructured text blocks, leading to increased support tickets regarding property details.
We migrated to a React-based frontend to facilitate instant state updates for filters. This allowed for a “no-refresh” experience when toggling between price ranges and amenities.
Replaced static maps with Mapbox GL JS, enabling vector-based rendering. Clusters now dynamically explode into individual pins based on zoom level, improving performance by 40%.
We restructured the taxonomy to highlight Qatari specificities (e.g., “Majlis” availability), creating a competitive advantage over global generic platforms.
High-fidelity prototyping of the filter logic.
Visualizing the evolution from a static directory to a dynamic application.
The previous design forced users to choose between a list or a map. The redesign implements a persistent split-screen view.
We redesigned the property page to focus on trust and clarity. The isometric view below highlights the modularity of the new “Booking Widget” which floats on desktop to ensure the Call-to-Action is always visible.
Primary CTA color #c2185b was tested against white backgrounds to ensure a ratio of 5.8:1, surpassing the AA requirement of 4.5:1.
Body text was bumped from 14px to 16px (Roboto) to improve legibility on mobile devices, with line-height increased to 1.6 relative units.
All interactive elements, including calendar dates and filter tags, were padded to meet the minimum 44x44px target size for touch interfaces.
Mobile view demonstrating accessible touch targets.
Optimization strategies for low-bandwidth environments.
Implemented Intersection Observer API to lazy load property images. LCP (Largest Contentful Paint) improved from 2.8s to 1.1s.
Utilized vector tilesets rather than raster images, reducing data transfer for the map component by approximately 75% per session.
Route-based code splitting ensures users only download the JavaScript required for the specific page they are viewing.