“I Really Don’t Know What ‘Thumbs Up’ Means”: Algorithmic Experience in Movie Recommender Algorithms

Abstract

Many of our daily activities and decisions are driven by algorithms. This is particularly evident in our interactions with contemporary cultural content, where recommender algorithms deal with most of their access, production, and distribution. In this context, the Algorithmic Experience (AX) design framework emerged to guide the design of users' experiences with algorithms for social media platforms. However, thus far, a framework to design specifically for AX within the context of movie recommender algorithms was lacking. To this end, the present study combines a semiotic inspection analysis of the Netflix interface with sensitized design workshops and semi-structured interviews to explore AX requirements for movie recommender algorithms. Linking the analysis with design opportunities, we shed light on AX suited for movie recommender systems. Moreover, we extend the current AX design framework with two new design areas: algorithmic usefulness and algorithmic social practices.

Middle-Aged Video Consumers' Beliefs About Algorithmic Recommendations on YouTube

“…During the interviews, participants continuously expressed how these practices influence their understanding of the algorithm. This finding echoes similar studies about Netflix [2], in which the algorithmic experience of the movie recommendation system is affected by algorithmic social practices.…”

“…Influence factors such as "My Watch History", "My Search History", "My Comments", "Others' Comments", "My User Subscriptions", "My Likes & Dislikes", and "Others' Likes & Dislikes", are specific to YouTube as a platform. Even if previous studies do not consider them explicitly, other investigations seem to reflect similar findings, e.g., Alvarado et al 's study of the algorithmic experience of movie recommendations [2], in which users ask for a better understanding and more control over the influence of all of these and similar interaction opportunities with the video recommender algorithms. Likewise, explorations of algorithmic experience in movie recommendations [2] and social media [3] portray similar results related to the "Third-Party Data-Sharing" influence.…”

“…For Alvarado and Waern, algorithmic experience (AX) is an: "analytic framing for making the interaction with and experience of algorithms explicit" [3]. In a subsequent study, Alvarado et al proposed a framework for AX dedicated to movie recommender algorithms that distinguish different design areas like profiling transparency and management, algorithmic awareness, user-control, and algorithmic social practices remembering [2].…”

Foregrounding Algorithms: Preparing Users for Co-design with Sensitizing Activities

“…It seems that this explanation improved the understanding of the participants on how algorithms produce recommendations and select specific information, facilitating subsequent co-design workshops. A follow-up study also applied a similar technique [1].…”

“…For this part, the moderator used simple visuals from the Facebook press website. 1 The workshop continued with a brainstorming exercise in break-out groups, during which the moderator instructed the participants to reflect on their news feeds and write down factors that Facebook might take into account when ranking the items. The participants later combined these insights into a single diagram via a collaborative affinity mapping activity [cf.…”