A platform with strong content and a poor system for surfacing it
Weallnet is a Vietnamese media and content platform that combines news, entertainment video, chat messaging, and promotional content in a single app. It has real editorial investment and a meaningful user base — but the product had been built with speed as the priority, and UX as something to fix later.
By the time I joined, "later" had arrived. Engagement was declining. Users were spending less time per session. The product team knew the content was good; they didn't know why people weren't staying. My 3-month engagement was to find out and fix it.
4
Content sections restructured: news, entertainment, chat, promotions
3 mo
Full-time research, design, and iteration cycle
↑
Session duration and content consumption post-launch
Content overload is a design failure, not a content problem
Auditing the existing app revealed a pattern I've seen before: every surface was trying to do too much at once. The home screen mixed news, entertainment, promotions, and chat entry points with no visual hierarchy between them. Primary and secondary content competed equally for attention. Navigation mixed high-frequency actions with rarely-used settings. The result was an app that felt exhausting to use — not because of bad content, but because there was no system for deciding what mattered most in any given moment.
Four specific problems
- No personalisation — every user saw the same content in the same order regardless of their behaviour
- Navigation mixed primary flows (news, video) with secondary actions (promotions, settings) at the same visual weight
- Chat was buried under media content, reducing the social layer to an afterthought
- No clear information hierarchy on the home screen — everything competed for immediate attention
The content wasn't the problem — the architecture was. Weallnet had good editorial and strong media. What it needed was a system for getting the right content to the right user at the right moment.
Users wanted the depth — they couldn't navigate to it
I ran user interviews, surveys, and usability testing sessions across the engagement. The dominant finding was consistent: users weren't leaving because the content was bad. They were leaving because finding the content they actually wanted required too much effort on every visit.
The secondary finding was about the social layer. Users who used the chat features were significantly more retained than those who didn't — but the chat entry point was so buried that many users didn't know it existed. This was a discovery and navigation problem, not a feature problem.
A third finding shaped the personalisation approach: users had strong preferences across content categories, but those preferences were invisible to the system. Two users opening the same home screen saw identical content despite having completely different interests. The data to power personalisation already existed in user behaviour — it just wasn't being used.
Restructure first, personalise second, iterate continuously
Information architecture before any screen
I rebuilt the IA from scratch — separating news, entertainment, chat, and promotions into distinct, clear sections with dedicated navigation entries. The home screen became a curated surface (highlights + personalised recommendations) rather than a dump of everything. Secondary content types were moved out of the primary navigation into contextually appropriate locations.
Personalisation layer
Collaborated with the data team to implement behaviour-based content recommendations. Users' reading history, category preferences, and engagement patterns informed what appeared in their feed — without requiring explicit preference settings. The system learned and adjusted silently, producing a feed that improved over time.
Chat elevation
Chat was moved from a buried secondary tab to a persistent, clearly accessible primary navigation entry. This alone significantly increased chat engagement — and, predictably, overall session length, since social interactions extend time on platform.
What shipped
Personalised Feed
Behaviour-driven content recommendations adapting to each user's category preferences and reading history — no preference settings required
Restructured Navigation
Four content sections (news, entertainment, chat, promotions) clearly separated in primary navigation with distinct visual identity per section
Immersive Media Player
Full-screen video and image experience with minimal UI chrome — designed for content consumption, not interface exploration
Elevated Chat
Chat moved to primary navigation with persistent access — end-to-end encrypted messaging integrated without breaking the content flow
Home Screen Hierarchy
Clear visual hierarchy separating featured editorial content from personalised recommendations and promotional content
Longer sessions, more content, better discovery
The redesign produced measurable improvement across the key engagement metrics. Session duration increased as users who previously bounced after a few screens now navigated deeper into content categories. Content consumption per session rose — a direct result of personalisation surfacing more relevant content early. Chat engagement increased significantly once it was visible and accessible.
User feedback after launch consistently highlighted the same things: the app felt easier to use, content felt more relevant, and finding specific categories was no longer confusing. None of these were aesthetic changes — they were structural ones.