top of page

Just In Time for the Holidays, Study Explores Pros and Cons of ‘Returnless Partial Refunds’ When You Buy Online

Writer: Kelsie Papenhausen
Kelsie Papenhausen
5 days ago
3 min read

BALTIMORE, November 11, 2026 – Online platforms face a dilemma every time a customer wants to return a product: should they require the item to be shipped back for a full refund, putting the customer through a time-consuming and frustrating process that may reduce satisfaction and future spending? Or, should they issue a full refund while allowing the customer to keep the goods, sparing the customer that hassle but bearing the financial burden themselves?

 

A “returnless partial refund policy” offers a middle ground: customers keep products while receiving claim-dependent partial refunds, thereby reducing return hassle while limiting refund payouts. A new study assessed whether such a policy creates net economic value for the platform relative to traditional returns and how that value varies with refund generosity.

 

The study, which was published in the INFORMS journal Marketing Science, found that the policy generated net transaction value gains of $1.44 per customer over 30 days and $4.41 over 90 days. By easing return hassle, the policy reduced the platform’s processing costs and boosted future customer spending, despite this same convenience leading to higher refund payouts and potentially opportunistic behavior.

 

The study, “Returnless Partial Refunds: A Large-Scale Field Experiment in E-Commerce,” was authored by researchers from Zhejiang University of Finance and Economics, the UBC Sauder School of Business, and Zhejiang University.

 

The authors evaluated the policy using a randomized field experiment involving more than 850,000 customers on a leading global cross-border e-commerce platform. Eligible customers who initiated a return were randomly assigned to either a treatment group that could receive a returnless partial refund offer, or a control group limited to the traditional return process. The experiment ran in two phases: an initial high-refund phase which averaged a 50% refund ratio, followed by a low-refund phase that averaged 30%.

 

The platform implemented the policy through a multiagent artificial intelligence (AI) system, because processing claim-dependent returnless partial refunds at scale could be prohibitively costly and time-consuming for human agents. The AI system evaluated evidence in text, image and video forms, and generated real-time offers in approximately five seconds.

 

The study found a positive overall effect. At the initial return, immediate net transaction value declined by 3.2% overall, because higher refund payouts outweighed processing-cost savings, but rose by 1.3% in the low-refund phase. The good news was that over the following 30 days, customers in the treatment group spent 2.2% ($1.92 per customer) more, with gains appearing in both phases and persisting through 90 days.

 

From an environmental stewardship perspective, the policy also reduced return shipments, producing estimated carbon savings equivalent to the sequestration of roughly 2,000 mature trees in the experimental sample.

 

These benefits were accompanied by patterns consistent with strategic behavior: subsequent return requests increased by 15.5%, while approval rates and physical return follow-through declined. These declines were most pronounced for claims more susceptible to opportunism, including those involving subjective reasons, high-priced items, and rapid repeat requests.

 

“Our findings show that a returnless partial refund can improve net platform value even when refund payouts increase,” said the study authors. “The gains come primarily from lower processing costs plus higher subsequent spending, and the AI system made it possible to evaluate multimodal evidence and extend the policy beyond low-value items at scale.”

 

“People respond to the reduced hassle by spending more later, but some also tested the policy’s limits,” said Wu. “The data reveal a clear tradeoff between better service recovery and potential opportunism, which is why refund generosity matters.”

 

“Refund generosity is a powerful policy-design lever,” said the researchers. “Firms should look beyond immediate refund costs to consider future customer value, carefully calibrating refund levels to balance better customer experience against higher refund payouts and the risk of strategic behavior.”

 

“Returns need not be just a cost center,” said the study authors. “Well-designed refund policies can become a source of stronger customer relationships, growth, and environmental gains. The same principles can also apply to other settings where firms face similar economic and behavioral trade-offs.”

 

About INFORMS and Marketing Science

 

INFORMS is the world’s largest association for professionals and students in operations research, AI, analytics, data science and related disciplines, serving as a global authority in advancing cutting-edge practices and fostering an interdisciplinary community of innovation. Marketing Science, a leading journal published by INFORMS, publishes research on quantitative marketing, consumer behavior, pricing, and strategy that informs managerial and policy decisions. INFORMS empowers its community to improve organizational performance and drive data-driven decision-making through its journals, conferences and resources. Learn more at www.informs.org or @informs.

 
 
bottom of page