A new survey has found that related articles are still the most popular feature of scholarly websites.

The survey, “How Readers Discover Content in Scholarly Publications”, has polled researchers every three years since 2005 with questions about their behavior in the discovery of journal articles and video. The latest survey, which was conducted during January, February, and March of 2021, received over 15K responses from readers in every academic discipline and from all parts of the world. The survey was carried out by Simon Inger and Tracy Gardner, from Renew Consultants (www.renewconsultants.com) who have worked within the scholarly publishing community for over 30 years.  It was supported by TrendMD alongside a number of other societies and publishing organizations. The survey findings and all the raw survey data are freely available.

A key finding of the survey: “Related Articles functionality is the most useful feature of those tested, and has maintained its position fairly consistently over the period of study.” Furthermore, as illustrated in the following figure, Related Articles were the most popular feature in all three broad areas of scholarly research defined by the survey: Medicine, Social Sciences/Humanities, and Science/Tech/Engineering.

Lists of related Articles have been a familiar feature on publisher websites for many years, but unlike the “More like this” suggestions that are often associated with a Search function, TrendMD’s Recommended Articles widget uses an AI technique called Collaborative Filtering to offer recommendations that are most likely to be of interest to the reader, based not only on the article they are reading, but also on what they have read before, and on what other readers with similar interests have read. This technology has been shown to raise users’ overall clickthrough rate on the recommendations feature by nearly 3 times the clickthrough rate of a standard “similar article” algorithm.

More than 500 publishers who feature TrendMD’s recommendations widget on more than 5K websites in the TrendMD network benefit from additional pageviews and readership through our traffic exchange system. This reader survey helps to show that in addition to the obvious benefits that TrendMD provides to publishers, those article recommendations are appreciated by readers too!