Hyper-Media Search and Consumption
51 Pages Posted: 27 Jun 2013
Date Written: June 25, 2013
Hyper-media, owing to its ability to embed and link to information from other sources, has become an important modality for consuming news and information. Yet there remains a dearth of empirical research seeking to explain and forecast consumer behavior in this context.
We develop a structural model of hyper-media search and consumption that accounts for features unique to this environment, including: 1) the rapid turnover of information (and consumers’ concomitant uncertainty about its relevance and availability), 2) the redundancy of information across sites, and 3) the important signaling role played by linked excerpts. We estimate the model using panel data on consumer celebrity blog browsing and information scraped from sites regarding the link structure between them.
We find that blogs are differentiated horizontally by their degree of sexually-oriented content and that links are informative of the linked sites’ content. Results indicate links can decrease (increase) visitations to the linked (linking) sites.
A counterfactual experiment related to fair use indicates that sites with mainstream (niche) content lose more (less) traffic when excerpting is prohibited. A net neutrality counterfactual shows that consumers who strongly prefer a high bandwidth site might be better off after its bandwidth is allocated to less preferred sites.
Keywords: Structural models, Dynamic programming, Learning models, Bayesian estimation
JEL Classification: M31, D83
Suggested Citation: Suggested Citation