ABSTRACT
The digitization provides mass media with various means to personalize their online offerings. By better meeting user preferences, personalization theoretically allows mass media to counteract declining circulations and revenues. While past research has extensively addressed approaches to personalization, it is unclear to what extent mass media use personalization – a gap we address in this paper. Against empirical evidence from a one-week experiment simulating browsing behaviors on five German newspaper websites, we find little personalization. We discuss this counter-intuitive finding and derive various suggestions for future research. We conclude with a summary and an outlook on personalization systems in mass media.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1. Due to technical difficulties causing incorrectly captured screenshots, we repeated the screenshots for the SZ sports article and SZ travel article on March 3, 2020 (from 1:18 pm to 1:38 pm).
2. For instance, on BILD, the Taboola feed used the headlines ‘More from BILD’ and ‘This could also be of interest to you’. The latter indicated personalized content based on users’ interests and reading habits. However, refreshing the page several times showed that the articles appeared randomly and thematically did not relate to the respective user account’s interest. Similarly, the Taboola feed on the SZ website appeared to be random. In contrast, for WELT, the Taboola feed is related to the content of the respective article. This suggests that on the WELT website, the Taboola selection was based on Taboola’s content logic. This means that Taboola applies different recommendation logics.
3. 43 min with A1, 43 min with B1, 51 min with A2, 50 min with B2 – see Appendix for detailed, daily times spent on every news website with every account per article (in min).