Dashboard for media publishers to let them gain AI driven insights into their audience and content interactions

Abstract

This project is looking into the changing workflow of news publishers as they adopt the use of algorithmic personalization of their content to the readers of the news sites online. This report describes the initial phase of a research project of Recombee, recommendations as a service provider, which aims to provide an analytical and configuration dashboard for the AI recommendation system. This dashboard is to be used by different members of the publishers staff, from editors to marketing people. Using human centered design (HCD) methods and an iterative approach with a short feedback loop with future users this project is looking into what are the most important analytical data points that the staff needs to see in order to work well with the recommendation tool. Following the research phase a simple design prototype is proposed for the dashboard. The final version of the prototype covers two selected areas of interest, basic content performance analytics and A/B test analytics. The overall finding of the research and design work is the need to start with a very simple and minimal analytics tool to get the editors on board and then expand the features and data covered later on based on the actual usage and what the news publishing staff is missing.