How rating design influences purchasing decisions in digital marketplaces

From booking a short-term rental to selecting a babysitter, customer ratings play a central role in how people make decisions online. But new research shows that purchasing decisions isn’t shaped by rating value alone, it’s also influenced by how those ratings are presented.

Xuemei (Shirley) Huang, a faculty member in CMU’s business information systems department, studies how design choices in digital platforms affect user decision-making. Her research examines whether presenting the same rating information in different formats such as star-based graphics or numerical scores, changes how users interpret that information and whether they are willing to complete a transaction.

“Our goal was to understand whether presenting identical rating information in different ways changes how people interpret it,” Huang said. “Even small design choices can shape decision-making.”

One message, multiple formats

Online platforms commonly present ratings as stars, numbers or a combination of both. While these formats often represent identical information, Huang’s research shows that users do not always process them in the same way.

“When ratings are shown numerically, people tend to interpret them more precisely,” Huang said. “With star ratings, users may unintentionally overestimate or underestimate what the score actually reflects.”

That distinction led Huang to examine whether rating formats matter more in certain contexts than others.

When risk changes priorities

The study, published in the Journal of Decision Systems, found that rating format plays a limited role in low-risk decisions. When users were making relatively routine choices, such as booking accommodations, both star ratings and numerical ratings performed similarly.

“In lower-risk situations, people are generally comfortable as long as the rating meets their expectations,” Huang said. “Small differences do not carry much weight.”

However, the results differed in higher-risk scenarios, such as choosing a caregiver or a babysitter involving direct interaction with your family members. In those cases, graphical star ratings were more likely to introduce bias.

“When risk is higher, users spend more time reading descriptions and reviews and less time carefully evaluating the rating itself,” Huang said. “That makes the way ratings are displayed more influential.”

Designing for better decisions

Rather than recommending a single best rating format, Huang emphasized the importance of aligning design choices with the type of decision users are making.

“Platforms should consider the level of risk involved and how users prioritize information,” she said. “Thoughtful design can help users make more informed and confident decisions.”

She notes that even within the same platform, different interactions may call for different approaches to presenting rating information.

Bringing research into the classroom

Huang’s findings also inform her teaching, particularly in courses focused on data visualization and analytics.

“Visualization is not always the best solution,” she said. “Students need to think about the audience, the decision being made and the context before choosing how to present data.”

By connecting research to real-world applications, Huang helps students develop skills in critical thinking, communication and user-centered design.

As digital platforms continue to shape everyday choices, Huang’s work highlights an important takeaway for businesses and designers alike: how information is presented can be just as important as the information itself.