Measuring the Value of Recommendation Links on Product Demand

Anuj Kumar () and Kartik Hosanagar ()
Additional contact information
Anuj Kumar: Warrington College of Business, University of Florida, Gainesville, Florida 32611
Kartik Hosanagar: The Wharton School, University of Pennsylvania, Philadelphia, Pennsylvania 19104

Information Systems Research, 2019, vol. 30, issue 3, 819-838

Abstract: Recommending substitute products on focal products’ pages on an e-commerce website can impact product sales in two ways. First, the visibility of a product as a recommendation on other products’ pages may increase its exposure and result in a greater number of its page views. Second, visibility of substitute products on the product’s page may cannibalize its own sales while resulting in greater exposure for the substitute products. The net impact of these opposing effects is unclear. We conduct a randomized experiment on a fashion apparel retailer’s website to answer the following questions: (1) what is the causal value of recommendation links from a product to its recommended products in terms of the additional sales for both the product and its recommended products, and (2) how does the value of a product’s recommendation links vary based on its network characteristics, such as its PageRank and the strength of its relationship with neighboring products? We find that as a result of a recommendation, on average, (1) the daily number of product page views increased by 7.5%, and (2) conditional on a product’s page view, its sales decreased by 1.9%, and the sales of its recommended substitutes increased by 9%. On average, recommendation links of a product result in an 11% gain in total sales of the product and its recommended substitutes. However, these gains are not evenly distributed among all products. We find that although the number of page views for a product is positively affected by the number and strength of its incoming links, its sales (its recommended products’ sales) conditional on its page view are negatively (positively) affected by the strength of its outgoing links. We conduct policy simulations to highlight how retailers and producers can apply this knowledge by engineering the recommendation network through sponsored links.

Keywords: product recommendation network; product recommendations; digital advertising; electronic commerce; randomized field experiment; sponsored product advertising; average treatment effect (search for similar items in EconPapers)
Date: 2019
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (9)

Downloads: (external link)
https://doi.org/10.1287/isre.2018.0833 (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:inm:orisre:v:30:y:2019:i:3:p:819-838

Access Statistics for this article

More articles in Information Systems Research from INFORMS Contact information at EDIRC.
Bibliographic data for series maintained by Chris Asher ().