The Challenge

Hypothesis: Netflix recommendations are strong for history-based viewing, but they don’t adapt to a user’s changing moods in the moment. This creates frustration when suggestions feel repetitive or misaligned with how users feel.

As a junior designer, my goal was to explore a feature that could help users find content faster and more meaningfully by mood.

Research and Discovery

Validating the hypothesis:

Investigated how users discover content and what factors influence their choices.
Conducted seven user interviews (15–20 minutes each) and uncovered that emotional state was a key driver in content selection.
Learned users wanted shows that matched how they felt after work, on weekends, or when relaxing.
Developed personas - Film enthusiast, the Casual viewer and the On-the-go viewer/ Binge watcher to capture behavioral patterns.

DEsign Approach

Based on user feedback:

Sketched and prototyped a “Search by Mood” feature.
Built categories like “Be Amused”, “Get inspired”, “Feel Empowered”.
Integrated feature in both the home screen and the search menu for easy access.

Testing and Iteration

Ran usability testing with six participants.
Key Insights
- 75% found the feature easy and fun to use.
- 50% said some mood labels were unclear and needed refinement.
Iterated by:
- Clarifying mood tag language.
- Improving hierarchy for visibility.

Outcomes

Users reported finding content faster when searching by mood.
The feature increased satisfaction by making recommendations feel personal and relevant.
Highlighted the value of personalization beyond viewing history.

What I Learned

Clarity matters: Mood labels need to be intuitive for users to trust them.
Multiple entry points help adoption: Placing the feature in search and on the home screen gave users more confidence in finding it.
Feedback loops improve design: Iterating on vague mood categories showed me the importance of refining language with user input.
I finished this project in November 2024 and Netflix released their mood based discovery in April 2025.

Next Steps

If developed further,

Explore AI and contextual signals (like time of day or weather) for deeper personalization.
Expand testing to more diverse user groups.
Partner with researchers and engineers to validate feasibility, synthesize insight, improve feature adoption and measure impact on engagement.

See this Link for Hi-fi Prototype