New York, NY — Between Black Friday and Cyber Monday, millions of consumers visited websites like Amazon, Best Buy, Target, and Walmart to buy gifts for the holiday season sifting through thousands of product reviews to try to determine the right fitHowever, there’s a cost to this searching: consumers lose money if they end up making the wrong purchase, and sellers lose future sales if customers have a bad experience — struggling to find what they want. And, looking for the right product takes time. A new paper from Columbia Business School examines how platforms can strategically decide how to prioritize which products to display so that consumers can learn about product quality and make informed decisions quickly, while companies do not sacrifice sales and revenue opportunity in the process. 

Companies introduce new products all the time, like the ones promoted on Black Friday. Neither platforms nor consumers know the true quality of these products in absolute, and relative to competing alternatives. We all have our own experience comparing product alternatives on a website, balancing the price with the product features and looking at customer feedback. Platforms and consumers need to balance profits and their surplus against product experimentation and learning about competing alternatives. 

Columbia Business School Dean Costis Maglaras proposes that platforms address these challenges by deciding the order of how products appear to an arriving consumer. In the research, Social Learning from Online Reviews with Product ChoiceMaglaras, and co-authors Professor Marco Scarsini at the University of Rome Department of Economics and Finance, Assistant Professor Donwook Shin at HKUST Business School, and Stefano Vaccari of Global Data Hub, Global Digital Solutions, and ENEL Global Services find that the optimal order combines quality products that have been highly rated with new products that have fewer reviews. While many sites assign the top position to the most popular products at the moment, focusing on short term profits, this de-incentivizes experimentation for consumers to try and learn about other products that may be more attractive to them, to the sellers and the platform. So, focusing on short-term sales may slow down learning and harm long-term sales and profitability. 

The authors study a marketplace model where consumers evaluate product alternatives incorporating their price, their perceived quality which is educated by past consumer reviews and balancing search costs — which is the time spent looking for the desired product. Different ranking algorithms balance revenues, learning, and consumer search costs. The greedy and easiest solution is to place at the top of the search results products that we currently think are best – this optimizes short-term sales but may forego learning about the quality of other products that may in fact be better than the ones for which we have information now. This problem is accentuated in settings with dozens or hundreds or even thousands of alternatives. An alternative set of ranking policies forces a degree of experimentation, adding some under-explored products among the top ranked alternatives, which can be done in a way that does not jeopardize significant short-term revenues, but ultimately improves long term performance for the platform, for the sellers and for the consumers.

“As companies begin considering how to best market their products online for the holidays, these findings offer a playbook for alleviating consumer regret rather than maximizing immediate revenue. This will ensure that companies are making revenue from the products they sell, and consumers are satisfied with the online shopping experience,” said Dean Costis Maglaras. “Further refinement of users’ online shopping experience could alleviate the type of targeting we see in social media platforms by ensuring, or even requiring, that users are shown a more diverse set of content as opposed to retargeting individuals repeatedly with one type of information, ads, or news that they most recently selected.”

To learn more about cutting-edge research being conducted, please visit Columbia Business School.


About the Researcher

Dean Costis Maglaras

Costis Maglaras

Dean's Office
David and Lyn Silfen Professor of Business
Decision, Risk, and Operations Division