Surprise me! A longitudinal user study on serendipitous interface design in news recommender systems

Serendipitous encounters in news recommender systems offer users both pleasant surprises and opportunities to engage with a wider range of news content. Existing research has focused predominantly on algorithmic strategies for promoting serendipity, yet little is known about how users perceive and respond to such elements. This study shifts the focus from algorithmic performance to user experience, introducing a “surprise box” feature as an interface affordance for serendipitous news exploration. We conducted an online experiment (𝑁 = 118) comparing user feedback between a random news recommender and a personalized, similarity-based recommender system. Quantitative results indicated that users exposed to the personalized recommender were more likely to interact with the “surprise box”, suggesting their willingness to opt out of monotonous news feeds and seek serendipitous content. Qualitative feedback further revealed user awareness of the lack of diversity in the personalized feed; both groups, however, appreciated such an interface affordance for serendipity encounters. Our findings highlight that news users are not passive recipients but active agents seeking diverse content and showcase the effectiveness of interface affordances for promoting serendipitous engagement in the context of news recommendations.