The Relative Effect of the Convergence of Product Recommendations from Various Online Sources
Information Systems
Xu, Jingjun David; Benbasat, Izak; Cenfetelli, Ronald T.
Published in Journal of Management Information Systems, November 2020
In recent years, e-commerce has continued to proliferate, and a large number of product choices are available online. A lot of people aren’t just looking at reviews, or expert opinions alone. Websites are increasingly providing product recommendations from multiple sources to help their customers reduce product uncertainty and choose a suitable product from a wide variety of alternatives. Remarkably, there hasn’t been much research into this very real-world behaviour.
This study evaluates an increasingly common behaviour, people seeking recommendations from multiple sources, and explains how and why people integrate these different sources to make product choices. What recommendation source or sources combined has a greater influence on shoppers’ purchase decision? Opinions from fellow consumers, product experts who possess recognized product knowledge and provide objective product opinions, or online recommendation agents (RAs) that assist online shoppers by eliciting their product preferences and making product recommendations that satisfy such preferences?
Dr David Xu at the Department of Information Systems and co-authors carried out three online experiments with over 800 subjects from North America.
“We found that recommendation convergence between RAs and experts leads to the greater acceptance of the jointly recommended tablet products,” said Dr Xu.
“This convergence was greater than that between experts and consumers, between RAs and consumers, or the recommendations from any single source.”
This finding challenges the traditional belief that shoppers would pick a product that is recommended by a greater (vs. lesser) number of sources. In fact, it is based on how many dimensions of product uncertainty are best reduced collectively by these sources. The convergent recommendations of RA and experts triumph over others and are comparable to convergent recommendations among all three sources. The reason is that RAs best reduce fit uncertainty, and experts best reduce description and performance uncertainties. Experts and RAs complement each other by reducing all three dimensions of product uncertainty.
As one practical implication, online merchants are advised to incorporate multiple sources into their websites, as about 80% of the participants in the study tend to access two or three recommendation sources. The findings can also inform online merchants on which combinations of two sources are most important. Another final advice for practitioners is that experts should be solicited from publicly available and reputable channels such as CNET and Consumer Reports to make the expert recommendations valuable.
