Who should assemble a mystery box? UF research has an answer

When you buy a mystery box, whether it contains a Labubu collectible, Funko Pop toy or surprise book from Barnes & Noble, you’re participating in a market that’s expected to grow to $25 billion by 2034. New research from the University of Florida Warrington College of Business finds that who assembles mystery boxes, retailers or suppliers, can affect whether the strategy is profitable.

“Who assembles a mystery product can matter just as much as what’s inside the box,” said Amy Pan, William R. Hough MBA Faculty Fellow Associate Professor. “Our research finds that the decision over who assembles probabilistic products can shape whether the strategy is profitable and, under certain conditions, can benefit suppliers, retailers, and consumers.” 

Mystery or blind boxes are examples of what is known as probabilistic selling – when consumers purchase a product at a set price but don’t learn exactly which product they will receive until after they’ve paid. The strategy allows companies to package products with different qualities or values into a single offering without revealing the exact product in advance. 

Pop Mart’s Labubu and Funko Pop Disney mystery boxes are examples of suppliers assembling these products, while Barnes & Noble’s mystery book boxes and Hotwire’s car rental services illustrate retailers taking on that role. 

Using a game-theory model, Pan and her co authors, Warrington graduates Zhechao Yang (Ph.D. ‘25) of George Mason University and Hongseok Jang (Ph.D. ‘21) of Tulane University, compared situations in which the retailer assembles probabilistic products with those in which the supplier does. They found that probabilistic selling can make a supply chain more efficient by helping companies reach more customers, better match products with what customers are willing to pay, and, in some cases, reduce the pricing pressures that can arise when a supplier and retailer set prices separately. 

“The strategy is generally most attractive when the additional transaction cost of assembling and fulfilling probabilistic products, such as inventory, logistics, accounting, and fulfillment, is below a certain threshold,” Pan explained. “But our findings also show that supplier assembly can make probabilistic selling more likely than retailer assembly under some conditions.” 

The findings also suggest that retailers may benefit from letting suppliers handle the assembly of probabilistic products. Supplier assembly can make the strategy profitable under a wider range of conditions than retailer assembly, allowing both sides of the supply chain to benefit. 

The availability of high-quality products changes the strategy, too. 

When high-quality products are readily available, the model shows that probabilistic products can be offered alongside those products when transaction costs are relatively low. As those costs rise, companies may instead offer probabilistic products alongside both high- and low-quality products.

When high-quality products are in short supply, the strategy changes, the researchers find. Rather than selling those products individually, the model shows that all available high-quality products can be combined with lower-quality products to create probabilistic products. Those mystery box offerings are then targeted toward consumers who value higher-quality products, while lower-quality products are offered separately to other consumers.

Under the model’s limited-capacity conditions, probabilistic selling is profitable only when the supplier assembles the products.

“It’s under these limited-capacity conditions that there can be a ‘win-win-win’ outcome in which the supplier, retailer and consumers all benefit,” Pan said.  

The research, “Probabilistic Selling for Vertically Differentiated Products in a Decentralized Channel,” is published in Manufacturing & Service Operations Management. 

Researchers:

• Zhechao Yang – George Mason University Costello College of Business
• Hongseok Jang – Tulane University Freeman School of Business
• Amy Pan – University of Florida Warrington College of Business