Study Shows Algorithm Bias Increases Consumer Purchases by 10%

A study by South Korea's Fair Trade Commission has revealed that algorithm-driven self-preference practices by online platforms can distort consumer product choices. Most online shoppers tend to purchase top-ranked products, indicating that manipulation of algorithms can significantly alter purchase rates.

On June 28, the Fair Trade Commission announced the release of a report titled "Consumer Behavior Experiment on Algorithm-Based Self-Preference Practices of Platforms." This research aimed to verify how consumer choices change when platforms preferentially display their own products through search, recommendation, and ranking algorithms.

The commission created a virtual shopping mall that accurately replicated actual online shopping interfaces and conducted a randomized controlled experiment with 3,072 consumers. The study also analyzed the effectiveness of informational corrective measures, such as labeling.

The findings indicated that consumers heavily rely on the rankings presented by algorithms. A significant 51.7% of purchases were concentrated on the top five products, and 94.6% of consumers completed their purchases on the first page. Only 25.2% of consumers changed the default sorting order, and 83.8% did not use filtering options, demonstrating a tendency to follow the platform's default rankings.

The study also examined the effects of self-preference based on changes in order and ranking. In a second shopping round, a product that was identical to a previously lower-ranked item but priced 10% higher was placed at the top of the search results. This led to a purchase rate increase of approximately 34 percentage points for that item.

Conversely, the purchase rate for existing top-ranked competing products dropped from 52% to 20%, illustrating how simple ranking manipulation can significantly alter consumer choices.

The effectiveness of informational corrective measures to reduce choice distortion was found to be limited. A "warning label" attached to self-preferred products actually increased their purchase rate by about 4.5 percentage points and discouraged active exploration by consumers. Only 10.7% of consumers confirmed the sorting criteria.

However, among some consumer groups that did check the disclosures, a trend was observed where the purchase rate for self-preferred products decreased by approximately 18.4 percentage points.

This research may serve as a basis in legal disputes involving major platforms. The Fair Trade Commission is currently engaged in legal disputes with platforms such as Kakao Mobility and Coupang regarding self-preference practices.

A commission official stated, "Given the confidentiality and opacity of algorithms in the platform market, proving the causal relationship between actions and market outcomes is challenging. Therefore, experimental methodologies like randomized controlled experiments are expected to be useful analytical tools for competition policy research and law enforcement."