What Role Do Automated Product Pickers Play?

It’s rare today to encounter an e-commerce website that isn’t making use of automated tools to help identify products we might like.  What began on the likes of Amazon is now commonplace across the web, and so new research from the University of Texas at Dallas is interesting as it explores the impact of these systems on our buying habits.

“Recommender systems have become ubiquitous in e-commerce platforms and are touted as sales-support tools that help consumers find their preferred or desired product among the vast variety of products,” the researchers say. “So far, most of the research has been focused on the technical side of recommender systems, while the research on the economic implications for sellers is limited.”

The researchers focus on the informative role played by recommender systems, and in particular their role in telling us about products we might otherwise have been oblivious to.  It’s a system that seems attractive to sellers because it expands our awareness of their stock and can be considerably cheaper than traditional advertising might have been.

Changing trade

Indeed, the authors highlight how recommender systems have been shown to boost sales on marketplaces by between 35% on Amazon to 60% on Netflix.  This exposure might not always be profitable, however.

The study shows that sellers typically advertise less on their own while the competition from other sellers prompts them to lower their prices.  What’s more, any benefits that do emerge from recommender systems only tend to do so when the system is precise.

“This means that sellers are likely to benefit from the recommender system only when the recommendations are effective and the products recommended are indeed consumers’ preferred products,” the researchers say.

As such, while the increase in exposure can be undoubtedly positive, the fall in profitability may overshadow any positive effects that emerge.  As a result, the authors suggest that sellers don’t stop advertising in the false belief that recommender systems will do all the work for them as advertising can better target customers.

“Free exposure turns out to not really be free,” they explain. “To mitigate such a negative effect, sellers should strive to help the marketplace provide effective recommendations. For example, sellers should provide accurate product descriptions, which can help recommender systems provide better matching between products and consumers.”

Consumer changes

Recommender systems also have an impact on consumer behaviors.  For instance, on the plus side, they introduce consumers to new products and can benefit them when sellers have to compete intensely for their custom.

They can also, however, end up paying more in terms of increased prices than the additional value they receive for the personalized recommendations.

“Consumers should embrace recommender systems,” the researchers say. “However, sharing additional information, such as their preference in the format of online reviews, with the platform is a double-edged sword. While it can help recommender systems more effectively find a product that a consumer might like, the additional information can be used to increase the recommendation precision, which in turn can reduce the competition pressure on sellers and can be bad for consumers.”

The researchers argue that while considerable effort is going into the technical aspects of recommender systems, much less is known about their economic implications.  They hope that their research will be the first step in redressing the balance.

Facebooktwitterredditpinterestlinkedinmail