'Personalised advertising', consisting of custom-made messages, addressed to the individual user, is getting extremely popular on social networking sites (SNS) such as Facebook. Governments struggle with the possibilities and implications of this new phenomenon. Policymakers and advocacy groups are increasingly expressing concerns on privacy in SNS. The present project examines how SNS, and personalised advertising on SNS in particular, raise concerns about personal data privacy. The central research question is: "What is the impact of different forms of personalisation of SNS advertisements on the SNS users' privacy concern and on the advertisements' effectiveness?" Three subprojects approach this research question from different angles.
(i) The first subproject hypothesizes that the effectiveness of personalised ads in follows a curvilinear relationship with the degree of personalisation. Thus, an optimal level of personalisation, in which privacy and effectiveness are balanced, exists.
(ii) The second subproject aims to clarify, through a factorial survey, the impact of specific factors (i.e. position of the ad, level of personalisation, data sensitivity, organisation type) on privacy concern related to personalised SNS advertising.
(iii) The final subproject uses an experiment to examine technically feasible, future applications of personal data in personalised SNS advertising. This approach is necessary to ensure applicability and relevance of the results in the current fast changing digital environment.
In sum, the project is innovative through its methodological approach and its orientation towards future applications. Results will contribute to the existing privacy and advertising literature. A critical evaluation can inspire governmental action related to privacy protection and empowerment towards persuasive attempts on SNS. Moreover, it can provide a framework for advertisers to create privacy-friendly ads. The project thus not only has a scientific, but also an important professional and societal relevance.