Enhancing Mood-Based Content Recommendations for Netflix

 

Duration: 4 Weeks

Platform: TV

Design tool: Figma

Institution: Maryland Institute College of Art - Fall 2024

Role: UX Designer/ Researcher

 

Project Overview

Netflix is a leader in the streaming industry, offering diverse content to millions of users globally. Despite its success, the current content recommendation system relies heavily on past viewing history, which may fail to capture users' current emotional states. This project sought to design a solution that enables mood-based content discovery, creating a more personalized and satisfying user experience.

 

Problem Statement

The current Netflix recommendation algorithm inadequately adapts to users' dynamic moods, leading to frustrations and reduced platform engagement. Users expressed dissatisfaction with repetitive suggestions and difficulty finding content that matched their emotional state during specific moments.

 

Research Objectives

The goals were to:

  1. Investigate how users discover content and what influences their exploration behavior.

  2. Analyze the emotional impact of viewing patterns on content preferences.

  3. Design and test a "Search by Mood" feature that leverages user input for real-time mood-based recommendations.

 

Methodology

User Interviews

  • Conducted 15-20 minute moderated interviews with six participants representing varied demographics and streaming behaviors.

Persona Development:

  • Created detailed personas to reflect user types such as "Film Enthusiast" and "Casual Viewer."

Prototype Testing::

  • Designed a prototype incorporating "Mood-Based Recommendations" and conducted usability tests to gather insights and refine the feature.

 
 
 

Proposed Solution

Search by Mood: A feature integrated into the Netflix interface allowing users to select or filter content based on emotions such as "Relaxing," "Exciting," or "Feel-Good." This solution enhances discoverability and aligns recommendations with users' current emotional states.

 

Search by Mood Prototype for Netflix UX Case Study

 
 

Design Highlights

  • Mood Categories: Dynamic options like "Calm," "Energizing," and "Reflective."

  • Real-Time Adaptation: Users can adjust recommendations mid-session by updating their mood.

  • Dual Entry Points: Accessible via the home screen and search menu for ease of use.

 
 

Testing & Feedback

  • Scenarios: Users tested guided and unguided scenarios, including finding content to relax after work.

  • Results:

    • 75% of participants found the "Search by Mood" feature intuitive and engaging.

    • 50% identified mood categories that felt unclear and requested further refinement.

 

Impact

  • Improved user engagement by enabling faster content discovery aligned with emotional preferences.

  • Enhanced satisfaction through personalized and diverse recommendations, validated during prototype testing.

Takeaway

  • Mood-Based Preferences are Crucial - Emotional states significantly influence user content selection. Users often look for content that aligns with their feelings, making mood-based features essential for personalization.

  • Discoverability Drives Engagement - Users prefer easily accessible and intuitive features like dynamic filters and mood-driven categories, which reduce search effort and improve content discovery.

  • Frustration with Repetition - Participants expressed dissatisfaction with repetitive recommendations based solely on past viewing history. They desired fresh, dynamic content tailored to current moods.

  • Prototypes Validated User Needs - The "Search by Mood" prototype demonstrated strong user engagement, with 75% of participants finding it intuitive and effective during usability testing.

  • Importance of Emotional Categorization - Clear and relatable mood categories are critical for adoption. Ambiguity in mood labels can lead to confusion, requiring refinement of terminology and structure.

  • Dual Entry Points Improve Usability - Users appreciated having multiple ways to access the mood-based feature, including both the home screen and the search menu. This reinforced its discoverability and usability.

  • Personalization Needs Real-Time Adaptation - Users want content recommendations to adapt in real time based on current preferences and moods, going beyond static algorithms.

  • Large Screens Enhance Engagement - Viewing on larger screens, such as TVs, was reported to be more immersive and enjoyable, encouraging longer streaming sessions and more feature exploration.

  • User Input is Key to Success - Features like mood tagging, which allow users to provide direct feedback, were well-received and seen as a way to enhance algorithmic accuracy over time.

  • Future Opportunities for Personalization - Participants suggested further iterations, such as leveraging AI to integrate contextual cues (e.g., time of day or weather), for even more precise recommendations.