2 Industrial organization of media markets and competition policy

Data (r)evolution: the economics of algorithmic search and recommender services

“…They have to trust the system and believe in the suitability of the recommendation. Experience and credence goods come with imperfect information; if one market side (either demand or supply side) systematically lacks information, there are information asymmetries (Akerlof, 1970;Budzinski & Kuchinke, 2020). Automatized and algorithmic recommendations are based on (big) data; consumers do not (perfectly) know how recommendations are created.…”

“…From an economic point of view (a) search and (b) recommendation systems are individualized services and similar, yet not completely the same (Budzinski & Kuchinke, 2020). (a)…”

“…Using algorithmic SRS is beneficial and profitable for the companies (Budzinski & Kuchinke, 2020). By providing individualized search rankings and recommendation services and thus (better) meeting the consumers' individual preferences, companies can intensify the users' transactions and consumption time with the respective service which in turn increases demand and turnover.…”

“…), and "advanced" consumer data regarding stated and revealed preferences (e.g. stated: like-button on Instagram; revealed: videos I am actually watching, products I am actually buying) (Budzinski & Kuchinke, 2020). By combining those two data types and through connecting with data of people with similar behavior and interests so called "derived" data arise which algorithms then altogether can process.…”

“…Three stylized types of business models build on the idea of (consumer) data as payment (Budzinski & Kuchinke, 2020): (i) user data analysis for third parties, e.g. user-behavior analyses for upstream markets or targeting advertising content on the behalf of companies, (ii)…”