The ratings panel from participation to personification: Nielsen and the big-data reconfiguration of the audience

Abstract

The defining feature of the Nielsen ratings has always been that they derive from a statistically sampled panel of viewers. Even amid the firm’s turn to a panel/big data-integrated currency, Nielsen claims its viewer panel is the thing that sets them apart from their third-party competitors because of its capacity to ground and calibrate the disarray of platform-derived proprietary data. In this article, I discuss Nielsen’s new panel/big data integrated currency, which integrates viewer-panel data and big data with AI modeling to create a composite of US viewership. More specifically, I analyze how Nielsen has discursively reframed the value of their viewer panel in terms of its capacity to assist data personification. The word ‘personification’ describes the process of mapping human qualities back onto big data via machine learning systems. In reality, the process of personification scales and abstracts data in such a way that the output no longer bares a direct tie to a person. To illustrate the socio-cultural stakes of this, I detail three of the machine learning processes that Nielsen characterizes as ‘personification’: ID Graphing, Demographic Modelling, and Scaling Persona Profiles. I also analyze the particularly neo-liberal rhetorical work that Nielsen does to mythologize these AI processes of personification as socially beneficial.