Purpose: The purpose of this paper is to introduce methodology and findings of a trend study in the field of e-learning. The overall interest of the study was the analysis of scientific e-learning discourses. What comes next in the field of academic e-learning? Which e-learning trends dominate the discourse at universities? Answering such questions is the basis for the adaptation of service strategies and IT-infrastructures within institutions of higher education. Design/methodology/approach: Which e-learning formats dominate the current scientific discourse? To answer this question, a trend study based on a content analysis was performed. The abstracts of 427 scientific articles of leading German-speaking e-learning conferences Gesellschaft für Medien in der Wissenschaft and E-Learning-Fachtagungen der Gesellschaft für Informatik e. V. (GMW and DeLFI)--published from 2007 to 2013--were examined. A category scheme was derived from the Horizon Report. The category scheme then was gradually expanded and adapted to the data material during the investigation. Findings: This paper found that the detailed analysis of the frequency distribution over the seven years reflects the intensity of scientific discussion towards e-learning trends within the investigation period, and conclusions about the didactical or technical potentials of innovations can be drawn because both conferences are different in terms of their objective. The authors also classified the life stages of selected innovations based on the Gartner hype cycles, and the striking findings of the study will be formulated in the form of assumptions, which reflect the development potential of learning management, mobile learning, virtual worlds, e-portfolio, social media and Massive Open Online Courses in German Higher Education. Research limitations/implications: Only abstracts of the selected contributions were investigated. Errors in the category allocation due to unclear terminology cannot be excluded. Organisers of the investigated conferences often define the (main) topics. This influenced the spectrum of represented topics overall, as well as the focus of individual contributions. The above-presented study was conducted at German-speaking conferences and, therefore, reflects the situation in Germany, Switzerland and Austria. No conclusions about actors or institutional relationships can be made, in contrast to the original assumptions of discourse analysis. The categorial classification of contributions does not allow any conclusions about the quality of the discourse. Originality/value: The study shows how proceedings of scientific conferences can be used for trend studies. It became clear that discourse analytical studies can be used complementary to other methods of future studies. The advantage of this methodology lies mainly in the easy access to the text material, as conference proceedings are mostly available online. In addition, the analysis of large amounts of data (or texts) can be greatly facilitated by use of digital technologies (e.g. by automatic analysis of keyword). This paper makes an important contribution to the diffusion of digital media in higher education.