The client
MyFriendSeZ is a recommendation platform built on a simple but powerful insight: the best recommendations come from people you actually trust. The app allows users to discover and share trusted recommendations — restaurants, plumbers, mechanics, doctors, products — strictly within their personal network of friends and family, eliminating the anonymous and often fake reviews that plague public platforms like Google Maps and Yelp.
The challenge
MyFriendSeZ had strong core logic but three significant problems were limiting its usefulness and adoption:
- No Google search integration: When a user tapped on a recommended business, they got only what their friend had written — no address, no opening hours, no photos, no ratings from the broader world. Users had to leave the app and search Google separately, breaking the experience.
- Recommendation visibility logic was broken: Recommendations were not being surfaced to the right people within a user's network. A recommendation from a close friend would be buried while a distant connection's post appeared prominently. The algorithm was not reflecting the trust hierarchy the product was built on.
- App performance was poor: Slow load times, janky scrolling, and delayed network responses were creating friction at every point in the user journey. Users were abandoning sessions before seeing any recommendations.
Our solution
Google Places API integration: We integrated the Google Places API directly into the recommendation detail view. When a user taps on a recommended business, they now see the full Google Places profile inline — address, opening hours, photos, contact details, ratings, and a directions link — without leaving the app. The trusted friend's recommendation is shown prominently above this data, preserving the personal trust layer while adding the factual context users need to act on it.
Recommendation visibility logic overhaul: We audited the existing algorithm and rebuilt the feed ranking logic to correctly weight recommendations by network closeness. Direct friends' recommendations surface first, followed by second-degree connections, with clear visual indicators showing how the recommender is connected to the viewer. Recommendations now reach exactly the people they're most relevant to.
Performance overhaul: We profiled the app end-to-end, identifying the key bottlenecks: unoptimised database queries, missing indexes, redundant network calls on screen load, and uncompressed image assets. We rebuilt the data layer with efficient queries and caching, implemented lazy loading for images and off-screen content, and reduced initial screen load time substantially. The app now feels fast on mid-range devices — the majority of the user base.
Design revamp: Alongside the functional improvements, we updated the UI — cleaner information hierarchy, better use of whitespace, and improved visual separation between the personal recommendation and the Google data layer, making both immediately readable.
Results
- Google Places data now surfaces inline — users no longer leave the app to search manually
- Recommendation feed now correctly prioritises close network connections
- App performance improved substantially — faster load times across all screen types
- Design revamp improved clarity between personal recommendations and contextual business data
- User session length increased as friction points were removed from the core discovery flow