We all know that the big American retailers, such as Costco, Amazon and Target, all make great use of personalized recommender systems. When you search for example for an e-reader once, the next time you visit Amazon, you will see different kind of e-readers, recommended for you. Once you buy the e-reader, the next day you will see on Amazon all different books and accessories recommendations. But what about the companies whom do not use personalized websites, do they also benefit from these recommender systems?
In the paper of Ghosal, Kumar and Mookerjee, the impact of recommender systems on competition among personalized and non-personalized websites is discussed. The main question in this paper is the fact that how these recommender systems have an impact on the used prices and profits for both the companies.
Recommender systems help customers to find their ideal product and reduce their fit costs. These are the costs that come along with the search through the big stack of products to find your ideal product. Yet these beneficial recommender systems are not used by all companies, due to their high cost. Especially small and medium sized companies, for example Buybooksonline.com, still does not use these systems.
In many cases, customers search their preferred product with the recommender system on the big retailer websites and then decide whether they want to buy this product from this website, or from a smaller retailer. Typically, the prices on websites using these personalized websites are higher, due to the high costs of the system. These assumptions will be tested in the study. Recommendations are also based on the customers past purchases, so when the customers buy at another firm, the recommendation quality from the personalized website will declines. To reduce this risk and reduce the future fit cost, customers may decide to distribute their purchases among personalized and non personalized websites.
In the first part of the study the focus relies on the choice of the customer between buying products form a personalized website or an identical substitute from a non personalized website, when using the recommender system from the personalized website. The second part of the study focusses on the pricing game between the two firms.
The authors find interesting results regarding the strategic behavior of the customers, showing that they indeed make a trade off between the quality of the recommender systems and the the lower price between from the non-personalized website. They also find that non-personalized websites can free ride on the recommender systems, and gain additional revenue. But when fit costs increase, due to lower quality recommendations, the revenue of the non-personalized firm decreases.
What about you guys? With this knowledge, would you still purchase at the cheapest website, or go for the more expensive website with better recommendations and lower fit costs? I’m really interesting in your opinions!
I also recommend you guys to read the full article, it is really interesting and close to customer centric digital commerce.
Ghoshal, A., Kumar, S., & Mookerjee, V. (2015). Impact of Recommender System on Competition Between Personalizing and Non-Personalizing Firms. Journal of Management Information Systems, 31(4), 243-277.