Credibility by automation : Expectations of future knowledge production in social media analytics

Social media analytics is a burgeoning new field associated with high promises of societal relevance and business value but also methodological and practical problems. In this article, we build on the sociology of expectations literature and research on expertise in the interaction between humans and machines to examine how analysts and clients make their expectations about social media analytics credible in the face of recognized problems. To investigate how this happens in different contexts, we draw on thematic interviews with 10 social media analytics and client companies. In our material, social media analytics appears as a field facing both hopes and skepticism—toward data, analysis methods, or the users of analytics—from both the clients and analysts. In this setting, the idea of automated analysis through algorithmic methods emerges as a central notion that lends credibility to expectations about social media analytics. Automation is thought to, first, extend and make expert interpretation of messy social media data more rigorous; second, eliminate subjective judgments from measurement on social media; and third, allow for coordination of knowledge management inside organizations. Thus, ideas of automation importantly work to uphold the expectations of the value of analytics. Simultaneously, they shape what kinds of expertise, tools, and practices come to be involved in the future of analytics as knowledge production.Peer reviewe