Tracking Back-Talk in Consumer-Generated Advertising

ABSTRACT

The advent of inexpensive hardware (video cameras) and free video-production and -editing software has enabled almost anyone to produce a reasonably competent video. When this is coupled to free video-hosting sites such as youTube, individual consumers can produce content—and many do so—in the form of ads about the brands they love, hate, or simply want to comment on. This means that advertising no longer is strictly under the control of marketers and their advertisers' agencies. It also means that many of the tried-and-trusted tools of advertising research do not work well in the age of consumer-generated content. Much of the feedback on consumer-generated advertising is in the form of ad hoc comments and discussion on video-hosting sites rather than data collected by means of formal structured survey. Yet it may be critical, in many cases, for those who manage advertising to understand it well. The authors introduce and demonstrate two approaches that may be used to make sense of the conversations that surround consumer-generated advertising—correspondence analysis of the word structure in consumer comments and a new form of Bayesian machine learning-based content analysis that iteratively “learns” concepts and their relationships. Managerial implications are identified, the limitations of the research acknowledged, and avenues for future research outlined.