A Recommendation on Recommendations
Online consumers are inundated not only with choice, but also with recommendations. According to new research, without careful curation these recommendations can actually lower the likelihood of a purchase.
Look for a rock hammer and you also find a chisel. Look at a rain jacket and you’ll see rain pants. Look up a toy ninja sword and the ninja outfit appears below. Recommendation engines are pervasive in online retail; customers who viewed this item also bought… But does presenting a set of additional products affect whether consumers purchase what they originally sought?
In short: yes.
With colleagues Ravi Dhar from the Yale School of Management and Jennifer Savary from the University of Arizona, Yale SOM doctoral candidate Elizabeth Friedman recently investigated this question, publishing results in the Journal of Consumer Research. “Researchers [had] not yet explored whether and how the type of alternatives considered affects preference for the target option,” she writes. Nor had research examined “the psychological mechanisms underlying these effects.”
To dig into this issue, Friedman and her colleagues ran a series of experiments grounded in one general design that manipulated the context in which a product is displayed, showing suggestions or alternatives that were either similar or dissimilar to the product being viewed. In one case, for instance, participants were presented with a wireless speaker set selling for $25. Below this, two other items were labeled by a recommendation engine: “Customers Who Bought This Item Also Bought.” One group of participants saw similar items (headphones and more speakers) while the other group saw dissimilar items (two button down shirts). Participants were then asked how likely they were to buy the wireless speakers.
Across ten studies, Friedman and her colleagues ultimately found that considering dissimilar alternatives decreases intent to purchase a target option more than considering similar alternatives. This holds across a range of particulars—whether the good is hedonic or utilitarian, whether the consumer thought of alternatives or marketers presented them, and so on. By tweaking specific aspects of the experiment, they were also able to reveal that so-called “focal goals” drive this result.
Consumers set out to purchase products with a goal in mind. If they are looking for speakers, perhaps they want to upgrade their current audio setup; perhaps they want the option to play music in a room that doesn’t currently have speakers. Whatever the goal, it weighs heavily on consumer behavior. Given this context, Friedman and her coauthors found that presenting consumers with products that are dissimilar to the original target activates a competing goal, which, in turn, decreases the importance of the original goal. In the end, this leads to lower purchase rates. (The authors did find one important exception: amplifying commitment to a focal goal “shields” that goal from influence when consumers consider other products representing other goals.)
On the academic side, this work is the first to weave together two foundational, but distinct, strands of research: how people consider opportunity costs, and how goals factor into consumer decision-making. Equally important, though, are the implications for marketers. When recommending one product as a complement to another, marketers should carefully consider how purchase goals might connect the two products. Are they likely to be aligned in the mind of a consumer? If not, is there at least a clear way to amplify a purchase goal, and so shield it from adverse influence? “If the marketer’s goal is to sell a particular product,” they write, “any alternatives present in the choice environment should serve the same goal as the target option.”