Evaluating social media privacy settings for personal and advertising purposes Rob Heyman, Ralf De Wolf and Jo Pierson Rob Heyman, Ralf De Wolf and Jo Pierson are based at iMinds-SMIT, Vrije Universiteit Brussel, Brussels, Belgium. Abstract Purpose – The purpose of this paper is to define two types of privacy, which are distinct but often reduced to each other. It also investigates which form of privacy is most prominent in privacy settings of online social networks (OSN). Privacy between users is different from privacy between a user and a third party. OSN, and to a lesser extent researchers, often reduce the former to the latter, which results in misleading users and public debate about privacy. Design/methodology/approach – The authors define two types of privacy that account for the difference between interpersonal and third-party disclosure. The first definition draws on symbolic interactionist accounts of privacy, wherein users are performing dramaturgically for an intended audience. Third-party privacy is based on the data that represent the user in data mining and knowledge discovery processes, which ultimately manipulate users into audience commodities. This typology was applied to the privacy settings of Facebook, LinkedIn and Twitter. The results are presented as a flowchart. Findings – The research indicates that users are granted more options in controlling their interpersonal information flow towards other users than third parties or service providers. Research limitations/implications – This distinction needs to be furthered empirically, by comparing user’s privacy expectations in both situations. On more theoretical grounds, this typology could also be linked to Habermas’ system and life-world. Originality/value – A typology has been provided to compare the relative autonomy users receive for settings that drive revenue and settings, which are independent from revenue. Keywords Online social networks, Twitter, Facebook, LinkedIn, Advertising, Privacy settings Paper type Research paper 1. Introduction Studies related to social media and privacy refer to at least two types of privacy interchangeably. The latter is problematic, as both interpretations downplay important actors in social media (Karahasanovic et al., 2009). Users are either portrayed as cattle generating money for platform owners (Cohen, 2008; Fuchs, 2012a; 2012b) without a choice or users are lauded as empowered agents writing themselves into being free from constraints (Boyd, 2007). We will maintain two types of privacy throughout the paper to elaborate on their differences. The first type, “privacy as subject”, can be summarised as the management of information about one’s identity vis-a ` -vis the other users. The latter type of privacy has been called lateral or social privacy. In “privacy as object”, users are not seen by other users. Algorithms sort their behaviour and user-generated content (UGC) for economic benefits derived from big data. Both perspectives have a blind spot. The micro-level research exploring benefits for users should not underestimate the limits imposed on their system for commercial reasons, which are driven by decisions taken on an aggregated level. The surveillance and critical theory studies have better tools to conceptualise these challenges, but it is impossible to understand why users join these platforms of exploitation in the first place. Coté and Pybus Received 24 January 2014 Revised 2 April 2014 Accepted 17 April 2014 This research was funded by EMSOC. EMSOC (User Empowerment in a Social Media Culture) (www.emsoc.be) is a four-year project (1 December 2010-30 November 2014) in the SBO programme for strategic basic research with societal goal, funded by IWT (government agency for Innovation by Science and Technology) in Flanders (Belgium). The research project is a cooperation between Vrije Universiteit Brussel (IBBT-SMIT & LSTS), Universiteit Gent (IBBT-MICT & C&E) and Katholieke Universiteit Leuven (ICRI & CUO), coordinated by IBBT-SMIT. PAGE 18 info VOL. 16 NO. 4, 2014, pp. 18-32, © Emerald Group Publishing Limited, ISSN 1463-6697 DOI 10.1108/info-01-2014-0004