Carolin Kaiser, Lara Enzingmueller, René Schallner
A multitude of photos is posted each day on social media. These photos do not only reflect people’s social lives but also their touchpoints with brands, representing a rich source of knowledge for market research. Since most of the early software tools focus on text, this wealth of information has long been neglected. Building on the new advancements in computer vision, this paper presents an automated approach for analyzing the temporal development of brand-related social media photos. The continuous analysis allows to detect consumer trends, to measure the influence of marketing campaigns and to examine the correlation with sales. Thus, marketers are in the position to identify opportunities and risks early on and to take appropriate actions. The approach is illustrated by a case study from the confectionary industry.