On the ScienceDirect platform, users viewing an article will see the “Recommended Articles” section on the right-hand side of the page. This feature presents relevant and related scientific publications based on the content being read.
Recommendations are generated using the following factors:
Content relevance
User reading and usage behavior
Content and social interaction signals
When a user clicks and views one of the recommended articles, this action is recorded as a use of the feature.
Time savings are estimated by comparing the time required to reach an article via the recommendation system with the time needed to find a similarly relevant publication through manual search.
Testing indicates that the article recommender saves users an average of 110 seconds per view, enabling faster and more efficient access to relevant academic resources.
Beyond time efficiency, this feature also helps users discover valuable and insightful articles they may not have found otherwise.