Article ID Journal Published Year Pages File Type
4942549 Electronic Commerce Research and Applications 2016 18 Pages PDF
Abstract
Recommender systems are commonly used by firms to improve consumers' online shopping experiences, with the secondary benefit of increased sales and profits. Prior research has demonstrated that a trade-off between relevance and profit exists, and that recommendations' manipulations and biases may hurt the credibility of recommender systems, and thus reduces customer trust. While many of the proposed designs suggest simple heuristics to bias recommendations toward higher-margin items, very little is known about consumers' reactions (in terms of purchasing behavior and trust) to recommender algorithms that balance recommendations' relevance and profitability or the drivers of this behavior. We aim to fill this gap. Data from an online randomized field experiment showed that balancing recommendations' accuracy and profit has a positive effect on consumers' purchasing behavior and does not affect their trust. We also found that the profit made during our experiment was due to a balance of several variables.
Related Topics
Physical Sciences and Engineering Computer Science Artificial Intelligence
Authors
, , ,