Interactive Content Diversity and User Exploration in Online Movie Recommenders: A Field Experiment (2309.13296v1)

Abstract: Recommender systems often struggle to strike a balance between matching users' tastes and providing unexpected recommendations. When recommendations are too narrow and fail to cover the full range of users' preferences, the system is perceived as useless. Conversely, when the system suggests too many items that users don't like, it is considered impersonal or ineffective. To better understand user sentiment about the breadth of recommendations given by a movie recommender, we conducted interviews and surveys and found out that many users considered narrow recommendations to be useful, while a smaller number explicitly wanted greater breadth. Additionally, we designed and ran an online field experiment with a larger user group, evaluating two new interfaces designed to provide users with greater access to broader recommendations. We looked at user preferences and behavior for two groups of users: those with higher initial movie diversity and those with lower diversity. Among our findings, we discovered that different level of exploration control and users' subjective preferences on interfaces are more predictive of their satisfaction with the recommender.

• The paper demonstrates that allowing users to adjust exploration levels enhances user satisfaction and engagement.
• It employs interviews, surveys, and a large-scale field experiment to reveal varying preferences for recommendation diversity.
• The study shows that adaptive interface designs can effectively balance personalized suggestions with broader exploratory options.

The paper, "Interactive Content Diversity and User Exploration in Online Movie Recommenders: A Field Experiment," tackles the challenge of balancing personalization and diversity in movie recommender systems. Recommender systems often face difficulties in striking a balance between providing personalized suggestions that align closely with users' existing preferences and offering diverse recommendations that can help explore new interests.

Key Findings:

1. User Preferences on Recommendation Breadth:
   • Interviews and surveys revealed a mixed sentiment towards recommendation breadth. Many users appreciated narrower, highly personalized suggestions, viewing them as useful. However, a notable minority expressed a desire for recommendations with greater diversity.
2. Impact of Exploration Control:
   • The paper highlights the importance of exploration control. Users' satisfaction with the recommender system is significantly influenced by the ability to control the degree of exploration offered through the interface. Different user groups, characterized by their initial diversity of movie preferences, showed variations in how much control they preferred.
3. Field Experiment and Interface Design:
   • The authors conducted a large-scale online field experiment testing two new interface designs. These interfaces aimed to broaden the range of recommendations accessible to users. The changes targeted increasing user engagement by allowing them to explore beyond their usual preferences.
4. User Groups and Diversity:
   • Users were divided into two groups based on their initial diversity in movie preferences: those with higher diversity and those with lower diversity. The experiment revealed that the levels of exploration desired and the subjective preferences regarding interface design varied between these groups. Users with initially diverse tastes reacted positively towards interfaces offering broader exploration.

Implications:

The paper suggests that recommender systems can benefit from personalized interfaces that allow users to adjust the diversity of recommendations. This personalization not only enhances user satisfaction but also potentially increases engagement by catering to individual preferences related to exploration. By enabling users to control the extent of recommendation diversity, systems can better cater to a broader range of user expectations and satisfaction levels.

This research contributes to the understanding of how interface design and user control over exploration can play critical roles in improving the effectiveness of recommendation systems, especially in the context of the entertainment industry like movie recommenders.