Amazon has more than 1 million third-party sellers in the United States alone, making it hard to stand out in the crowd. A search query for “televisions,” for example, produces more than 1,000 results, while a prompt for “school supplies” offers up more than 100,000.
That’s where the buybox comes in.
The buybox is an algorithm that curates multiple listings of the same product and displays the best value as the default option. It makes shopping easier for both customers and vendors, which boosts sales. Research from Wharton’s Santiago Gallino found that when a major online marketplace introduced a buybox for a subcategory of electronics, conversion rates increased by 23% and transaction volume went up by a hefty 78%.
“A buybox creates opportunities for customers and sellers,” said Gallino, a professor in the operations, information and decisions department. “Customers benefit from lower prices and higher quality, while sellers benefit from a standardized process that helps them get access to more customers.”
In his co-authored paper, “Algorithmic Assortment Curation: An Empirical Study of Buybox in Online Marketplaces,” Gallino and his colleagues concluded that a carefully designed, well-managed buybox is a powerful tool with positive results. The profound effect of a buybox explains why it’s so popular with marketplace sellers around the world — from Amazon and Walmart in the U.S., to Mercado Libre in South America, to Rakuten in Japan.
“As traditional retailers increasingly create marketplaces that feature third-party sellers, we anticipate that the use of buybox-like tools will expand beyond global marketplaces to include traditional retailers,” the authors wrote in the paper. The co-authors are Nil Karacaoglu, operations and business analytics professor at Ohio State University’s Fisher College of Business, and Antonio Moreno, business administration professor at Harvard Business School.
“If you are marketplace, the cleanest way for you to run a buybox is if you don’t participate as a seller, too.”— Santiago Gallino
A Buybox Sorts a Dizzying Array
Gallino likened the experience of shopping an online marketplace to looking for tomatoes at the farmers market, where 20 different vendors are selling similar tomatoes. After a while, the customer tires of comparing the cost and quality of each seller’s tomatoes, and what should be a low-stakes decision becomes frustratingly difficult.
“The same thing happens when you go to a digital marketplace. If you search for basic things, like a green iPhone, you see 20 options. It’s choice overload for the consumer,” Gallino said. “The buybox is an action the marketplace can take, because it puts a default that shows the best option for the consumer, not necessarily the cheapest.”
Working with an extensive dataset provided by the unidentified online marketplace, the professors made some key findings about buyboxes:
- The positive effects are more pronounced on mobile channels, which typically have higher search frictions because of small screens and more difficult navigation. Similarly, a buybox has larger effects on weekdays and during work hours, when customers have less time for searching.
- As more sellers participate in the buybox for a product, the average price paid for the product decreases, and the average seller quality of the product sold increases. This suggests that the buybox induces competition.
- A buybox helps sellers by automating tasks, such as creating product descriptions, which reduces the labor required to participate in the marketplace.
While the dataset was for a subcategory of electronics, the results can be generalized to any category of precisely defined products, such as auto parts, Gallino said. “If you move to apparel, it’s hard to find a situation where sellers have exactly the same thing,” he said. “A red dress is a red dress, but it’s not the same thing.”
“As a marketplace, you need to be careful to keep your two customers happy, and those are the vendors and the shoppers.”— Santiago Gallino
How to Optimize the Buybox
The professors emphasize that buyboxes are beneficial when designed objectively. They draw a distinction between a buybox and other kinds of algorithms such as sponsored search, in which sellers pay for the ranking; platform endorsement systems, in which the marketplace operator pushes products it approves; or recommender systems that aid customers in product discovery.
In 2023, the Federal Trade Commission and 17 state attorneys general filed an antitrust lawsuit against Amazon, alleging the retail giant uses unfair practices — including biased search results — to stifle competition and punish vendors. Gallino said the lawsuit has garnered interest in the paper’s findings.
“If you are a marketplace, the cleanest way for you to run a buybox is if you don’t participate as a seller, too,” he said.
The point of the buybox, Gallino said, is to help both customers and sellers through fair competition rather than selecting a “winner.” The algorithm should be inclusive and probabilistic, rather than preferential. Returning to the farmers market analogy, he said, if only one tomato seller shows up week after week, customers will stop coming.
“As a marketplace, you need to be careful to keep your two customers happy, and those are the vendors and the shoppers,” he said. “That is where the buybox algorithm can be relevant.”
