https://www.mdu.se/

The recommender system has become an intergral part of any online experience. Previous research on recommender systems effects have focused on the user as the single stakeholder by evaluating the utility of the system for end users in a short term perspective. This thesis is building on the current research on recommender systems by expanding the number of stakeholders with the business as an explicit stakeholder and by evaluating the long-term effects of recommender systems on both the business and the customers.

Subscription-based business models are increasing in popularity, outgrowing nonsubscription-based businesses. Many companies want to explain themselves as “a Spotify for X” focusing on highlighting the similarity to the well-known music streaming service. We arguethat there are distinct differences between subscription businesses that need to be articulatedand aim at providing a common language to use when talking about subscription-basedbusiness models. Supported by business model theory, with a specific focus on profitability,we explore and exemplify how customer activity relates to variability in revenues and costs.The subscription typology illustrates meaningful differences between subscription businessesin their variance in customer activity, and to what degree customer activity is linked torevenue and cost. Through the profit equation, customer activity is linked to profitability.Profit variability is shaped by volume- or choice-driven cost variability and upgrade- orinvestment-driven revenue variability. In this study, the expectation that the customerinteract with the product is seen as key for a subscription business. We define subscriptionsas recurring, advance payments for a product or service with which the customer activelyengages, highlighting the customer behavior aspect.