We conduct a field experiment on a movie-recommendation platform to investi-
gate whether and how online recommendations influence consumption choices.
Using a within-subjects design, our experiment measures the causal effect of recommendations on consumption and decomposes the relative importance of two economic mechanisms: expanding consumers’ consideration sets and providing information about their idiosyncratic match value. We find that the informational component exerts a stronger influence – recommendations shape consumer beliefs, which in turn drive consumption, particularly among less experienced consumers.
Our findings and experimental design provide valuable insights for the economic
evaluation and optimisation of online recommendation systems.