Streaming Platform
Redesign
How we reduced interactions to play from 7 to 3 and lifted engagement 34% through mood-based content discovery.
7→3
Interactions to play
+34%
Engagement in 90 days
−18%
Drop-off rate
+22%
Daily active users
The problem
The path to play was a maze
The client’s streaming platform had 12 million subscribers but was losing ground on engagement. Average session length had declined 22% over 18 months. Analytics pointed clearly: users were spending too long deciding what to watch, then giving up.
The navigation required seven distinct interactions to reach the play button from the home screen: Home → Browse → Genre → Sub-genre → Show page → Season → Episode → Play. On TV remotes, this was acutely painful. 68% of sessions that reached the third navigation level ended without a play. Users were lost inside a warehouse of 40,000 titles with no compass.
"Users aren’t browsing because they want to. They’re browsing because we haven’t given them a better way to find something to watch."
Original navigation flow — 7 interactions from home to play
Research
Understanding decision fatigue at scale
We ran 24 user interviews across three platforms, recruiting users who opened the app but didn’t watch within 15 minutes. This “browsing without playing” cohort was our primary target.
Key findings
- Mood-first decisions
82% of users described their decision as starting with a feeling, not a genre. They wanted something “light” or “intense” — not “drama” or “thriller”. Our taxonomy was built around the wrong axis.
- Paralysis by choice
When shown more than 8 options simultaneously, completion dropped dramatically. The homepage surfaced 60+ titles with no hierarchy or confidence signals.
- Recommendation distrust
Carousels were widely dismissed. Users felt the platform showed them what it wanted them to watch — not what they would actually like. Trust in the algorithm was critically low.
- Context collapse
The same title felt right in one mood and wrong in another. Genre alone couldn’t capture this. We needed a new organising principle.
Design approach
Mood-based discovery as the primary axis
The central insight: reorganise discovery around emotional intent, not content category. We mapped a “mood taxonomy” through card sorting sessions with 60 users, landing on 8 primary moods that covered 94% of described intentions: Unwind, Energise, Laugh, Think, Escape, Inspire, Feel, Share.
Each title received a mood score across all 8 dimensions based on a combination of editorial tagging, completion rate data, and watchtime patterns by time-of-day. A film watched predominantly at 11pm on weeknights scores differently to the same film watched on Sunday afternoons.
Mood taxonomy card sorting — affinity mapping output
The new flow: 3 interactions to play
Home → Mood selection → Curated results → Play. The home screen presents the 8 mood tiles directly. Selecting one surfaces a curated set of 12 titles specifically selected for that mood, current time and user history. The play button is above the fold on every title card.
We eliminated the genre browser as a primary navigation item, demoting it to an advanced filter within mood results. 90% of plays happen through mood paths. Genre now serves power users who want to refine further, not the default path for everyone.
New home screen — mood tile grid
Mood results page — 12 curated titles
The system
A component system designed for three platforms
The redesign required a unified component system that worked across web, iOS/Android and TV (Samsung Tizen, LG webOS, Apple TV). TV UX has unique constraints: 10-foot distance, D-pad navigation, no hover state, no touch. Every interaction had to work with only Up, Down, Left, Right, Select and Back.
We designed 48 core components in Figma with full platform variants. The title card was the most complex: it needed to communicate mood score, match quality, genre tags, duration, rating and play CTA — within a fixed 16:9 thumbnail area — without relying on hover states.
The solution: a layered disclosure system. Resting state shows title and a single mood badge. D-pad focus expands the card to reveal the full metadata row. Select plays immediately. The disclosure sequence respects the hierarchy: the most important action (play) is always one click from focus.
Outcomes
Measurable impact within 90 days
7→3
Interactions to first play, all platforms
+34%
Average session length, 90-day post-launch
−18%
Homepage drop-off rate
+22%
Daily active users, returning cohort
The mood discovery path accounted for 87% of all plays in the first month — substantially higher than projections. Genre browsing dropped from the primary path to a secondary refinement tool used by approximately 12% of sessions. The trust signal improvements (transparent match percentage, editorial mood tags) were cited by users in post-launch research as the most noticed change.