On June 26, 2026, the Korea Fair Trade Commission (“KFTC”) published a research report titled “Consumer Behavior Study on Algorithmic Self-Preferencing by Online Platforms” (“Report”). The Report is particularly significant because it represents the KFTC's first attempt at empirically examining the effects of algorithmic self-preferencing on consumer choice through a consumer behavior experiment designed and conducted directly by the KFTC.
Online platforms often perform a dual role by simultaneously operating a marketplace and competing within such marketplace through their own products or services. This creates incentives for the platforms to design search, recommendation, or ranking algorithms in ways that favor their own offerings. However, establishing the competitive effects of such behavior under traditional competition law has not been easy, because algorithmic self-preferencing is difficult for consumers to detect and challenging for competition authorities to observe directly.
Based on a controlled experiment involving 3,072 consumers, the Report found that consumers placed considerable reliance on default search rankings, that even relatively small changes in ranking can significantly influence purchasing decisions, and that labels or disclosures alone may not be sufficient to correct these effects.
Although the Report does not represent the KFTC’s formal legal position, it offers valuable insight into how the KFTC may approach future competition law enforcement involving digital platforms, particularly through the use of behavioral economics and experimental methods.
The key findings of the Report and their practical implications for businesses are summarized below.
I. OVERVIEW OF THE STUDY
1. Background
Recent competition law enforcement actions concerning alleged algorithmic self-preferencing by digital platforms, including Google Shopping and Amazon cases in the European Union, as well as the NAVER Shopping and Coupang cases in Korea, have brought increased attention to this issue.
Self-preferencing generally refers to practices whereby a platform uses search, recommendation, or ranking algorithms to give preferential prominence to its own products or services over those of competing businesses. Such practices may arise because platforms often act both as intermediaries and as competitors within the same marketplace, creating potential conflicts of interest.
The Report notes that consumers often perceive search rankings as objective indicators of product quality or relevance. Against this backdrop, the study was designed to empirically examine whether algorithmic self-preferencing influences consumer purchasing decisions.
2. Experimental Design
To isolate the effects of algorithmic self-preferencing, the KFTC created a simulated online shopping platform designed to closely resemble a real e-commerce marketplace. The experiment involved 3,072 consumers with prior online shopping experience and was structured as follows:
-
Initial shopping session: Participants were presented with standard search results that did not involve self-preferencing.
-
Second shopping session: The platform artificially moved its own product (priced 10% higher than competing products) to the top of the search results, while keeping all other product attributes unchanged.
-
Randomized treatment groups: Participants were randomly assigned to one of four groups: (i) a control group, (ii) a group shown a label identifying the platform’s own products, (iii) a group provided with information explaining the ranking criteria, and (iv) a group receiving both the label and the disclosure.
-
Behavioral tracking: The experiment recorded participants’ browsing behavior, including clicks, scrolling, page navigation, use of filters, and changes to sorting options.
By holding all factors constant except the search ranking, the experiment was designed to isolate the effect of algorithmic self-preferencing on consumer purchasing decisions.
3. Key Findings
(1) Consumers rely heavily on search rankings
The study found that consumers relied heavily on default search rankings when making purchasing decisions.
-
51.7% of all purchases were made from the top five search results.
-
94.6% of purchases were completed on the first page of search results.
-
74.8% of participants did not change the default sorting option.
-
83.8% of participants did not use the filtering functions.
According to the Report, these findings suggest that consumers tend to treat default search rankings as objective indicators of product quality or relevance, rather than actively comparing or exploring alternative products.
(2) Even a simple change in ranking significantly influenced consumer choice
When the platform’s own product (which was priced 10% higher than competing products) was moved to the top of the search results:
-
purchases of the platform’s own product increased by approximately 34 percentage points; and
-
purchases of competing products that had previously ranked near the top declined by approximately 32 percentage points.
The study suggests that consumers place considerable weight on search rankings as a signal of product quality, and that ranking manipulation alone may substantially reduce competitors’ opportunities to compete for sales.
(3) Labels and disclosures had only a limited corrective effect
The study also examined whether disclosure measures could mitigate the effects of self-preferencing.
A label identifying the platform’s own product actually reduced the consumers’ search behavior and resulted in a further increase of approximately 4.5 percentage points in purchases of the platform’s own product.
Similarly, only 10.7% of participants viewed the disclosure explaining the platform’s ranking criteria. While consumers who reviewed the disclosure appeared somewhat less likely to purchase the platform’s own product, the overall corrective effect on consumer choice remained limited.
(4) Consumers often failed to recognize the harm
Perhaps the most striking finding was that consumers did not recognize that they had made a less favorable purchasing decision, even when purchasing the platform’s own product at a price higher by 10%.
According to the Report, this suggests that consumers may not recognize the potential harm arising from algorithmic self-preferencing, even where it results in objectively less favorable purchasing outcomes.
II. POLICY IMPLICATIONS
The principal significance of the Report lies not so much in its individual empirical findings as in what it signals about the future direction of the KFTC’s enforcement approach toward digital platforms.
First, the Report suggests that the KFTC intends to make greater use of behavioral economics and experimental methods in algorithm-related competition cases. Given the inherent opacity of algorithms and the resulting difficulties in demonstrating anticompetitive effects through conventional analytical tools alone, experimental methodologies, such as randomized controlled trials, may increasingly serve as important complementary evidence alongside the traditional economic analysis.
Second, the Report indicates that the influence of search rankings on consumer choices may become an increasingly important consideration in assessing competitive effects. At the same time, it should be noted that the Report is a policy study rather than a statement of the KFTC’s formal legal position. In any actual enforcement action, the assessment of legality would continue to depend on a comprehensive analysis of all relevant factors, including market power, anticompetitive effects, and any efficiency-enhancing justifications.
III. PRACTICAL IMPLICATIONS
In light of these developments, platform companies should consider regularly reviewing whether their algorithms could result in any preferential treatment of their own products or services, and maintaining appropriate documentation of the principles governing algorithm design and subsequent modifications. They may also wish to ensure that product ranking criteria remain objective, implement appropriate safeguards to manage potential conflicts of interest, and integrate algorithm governance into their broader competition law compliance framework.
The significance of the Report lies not in establishing the illegality of self-preferencing itself, but in demonstrating the analytical tools and evidentiary approaches that the KFTC may increasingly rely upon in future platform investigations. The extent to which behavioral economics and experimental evidence will be incorporated into any actual administrative enforcement and judicial review remains to be seen. The increasing use of behavioral economics and experimental evidence is nevertheless likely to play an important role in shaping the development of Korean competition law in the digital platform sector.
* * *
For any inquiry or questions regarding the content of this newsletter, please contact us.
[Korean version]
