• Corpus ID: 26564629

HOW DO SOCIAL RECOMMENDATIONS INFLUENCE SHOPPING BEHAVIOR ? A FIELD EXPERIMENT

@inproceedings{Grahl2013HOWDS,
  title={HOW DO SOCIAL RECOMMENDATIONS INFLUENCE SHOPPING BEHAVIOR ? A FIELD EXPERIMENT},
  author={J{\"o}rn Grahl and Franz Rothlauf and Oliver Hinz},
  year={2013}
}

• Jörn Grahl, Franz Rothlauf, O. Hinz
• Published 2013
• Economics, Business

Previous research has tried to understand the influence of user-generated content on economic outcomes, however many of these analyses avoid any suggestion of causality which would be very important for drawing managerial implications. We conduct a randomized field experiment in an online store where we exogenously manipulate the visibility of the number of people who like a product (similar to “Likes”' collected by Facebook). This design allows us to analyze the causal effect of social recommendations on shopping behavior. In a four-week experiment with new customers, we find that displaying social recommendations caused a 12.97% revenue increase, generating 13,883.74 EUR revenue in the treated group compared to 12,289.46 EUR revenue in the control group. This difference is mainly caused by significantly longer search processes when social recommendations are displayed which also lead to a 22.19% higher chance that a first-time shopper makes a purchase. We compare the social influence effect of Likes on revenue with the inherent quality effect that the Likes also capture. The social influence effect is much smaller and it is usually only responsible for less than 10% of the total effect. This emphasizes that it is difficult to draw conclusions from non-experimental studies. If user-generated content is related to economic outcomes in observational data, the estimates can be heavily biased. Overall, our results suggest however that for online stores social recommendations and Likes are intangible assets with significant business value.