This research focuses on the analysis of the recommendation algorithms employed by Google Discover, utilizing data from two e-commerce platforms operating in Poland.
The study uses the information obtained from Google Search Console in a time span of 17 months. The examination of Google Discover focuses on the number of displays, clicks and click-through ratio, from the viewpoints of content publishers and web users.
The results suggest that user engagement positively influences a website’s efficiency in Google Discover, yet the algorithm also considers variables such as the popularity of similar content on other websites, user location and content update frequency. Thus, a website may be excluded from Discover despite a substantial click count.
There is a lack of studies on how Google Discover is perceived by users based on real data. We offer a quantitative perspective, which has not yet been done. This study offers an overview of the history and evolution of Google Discovery, an overview of data we used to show the perception of the service, and two unique perspectives on recommender service, users and publishers.
