A research team at the University of Tsukuba analyzed regional pricing for PC games sold on Steam and classified publishers' pricing settings into four patterns. Even though Valve provides recommended prices for each market, publishers do not adopt them uniformly. Comparing 268 companies and 7,375 titles across eight markets in 2023, the researchers found a range—from companies that closely follow the recommended prices to those with large variations by country or title.
However, this is an observational study capturing pricing practices at a single point in time. It does not prove that corporate affiliations caused the price differences, nor does it test which pricing strategy boosts sales. It's important to distinguish what the four categories do and do not explain.
How the 7,375 Titles Were Compared
Tomohiro Murakami and Mitsuo Yoshida used pricing data collected from Steam in August and September 2023. The study covered eight markets: in the Americas, the United States and Brazil; in Europe, France and the United Kingdom; and in Asia, China and India, along with Japan and South Korea. To avoid the effects of temporary sales, the researchers limited their analysis to regular prices before discounts. They also kept only publishers handling 10 or more titles, in order to capture stable pricing trends for each company. As a result, the final dataset covered 268 companies and 7,375 titles.
Because publishers with fewer than 10 titles were excluded, the pricing behavior of companies selling only a small number of games is not reflected in the analysis. The figure of 7,375 refers to the number of game titles, not the number of buyers or transactions. This selection criterion sets a boundary for extending the findings to all sellers or consumer behavior on Steam as a whole.
The benchmark for comparison was the region-specific recommended prices that Valve provided at the time, in 2023. For markets other than the United States, the research team calculated a "price ratio" by dividing the actual selling price by the recommended price. A ratio of 100% means the two match; a ratio of 130% means that if the recommended price were set at 100, the actual selling price was 130. This is not a rate of increase tracking price changes over time.
Next, the researchers used features such as the mean, median, maximum, and minimum price ratios for each company in each market, and applied hierarchical clustering. This is a statistical method that groups objects with similar characteristics, not a method that measures responses by changing experimental conditions. Nor is it the result of a theoretical model simulation. It is an analysis that explores groups of companies with similar pricing settings based on publicly available price distributions.
The unit of analysis is the publisher, whose pricing characteristics were aggregated. This is not an experiment in which consumers were assigned as subjects, nor a survey asking about purchase intent or willingness to pay. While pricing patterns can be observed, this design cannot reveal whether consumers who saw those prices actually made a purchase.
How to Interpret the Four Pricing Categories
The hierarchical clustering found that the largest group, with 116 companies, was the "Adjusted" type, which makes small adjustments to the recommended price in each market while keeping variation across markets relatively low. Next came the "Emerging Market Premium" type, with 69 companies, which tends to set prices above the recommended level in markets like Brazil and India, followed by the "Benchmark Follower" type, with 54 companies, whose prices stay close to the recommended level. The "High-Variance Portfolio" type, with large differences by country or title, comprised 29 companies. Together, these four groups add up to the 268 companies analyzed.
| Type | Number of Companies | Characteristics Shown in Public Materials |
|---|---|---|
| Benchmark Follower | 54 companies | Prices stay close to recommended levels in each market, with low variation |
| Adjusted | 116 companies | Small adjustments from recommended prices; relatively low variation across markets |
| Emerging Market Premium | 69 companies | Tends to exceed recommended prices, particularly in Brazil and India |
| High-Variance Portfolio | 29 companies | Large variation across markets and titles |
This table classifies how companies used a common benchmark. Because the study did not analyze sales volume, revenue, or profit for each group, it is not possible to determine which of the four types is superior. Nor is it guaranteed that these four clusters would appear in the same form at other times or on other platforms.
What Brazil's 128% and India's 130% Actually Show
By market, the average price ratios were approximately 111% for China, 128% for Brazil, and 130% for India. For Brazil, this means that if the recommended price were set at 100, the average actual selling price was about 128. India comes out to about 130 on the same scale. This difference should not be read as a rate of price increase, since the study did not measure price changes over time.
In France and the United Kingdom, average prices tended to stay relatively close to the recommended price, with low variation in price ratios across titles. Japan and South Korea showed the same tendency. However, the publicly released press materials do not list the average ratio, absolute prices, or confidence intervals for each of these four markets individually. The name of the statistical test or effect size used to compare differences between groups also cannot be confirmed from the public materials. What can be read without supplementing these missing details is the contrast that, in the four mature markets, prices stayed close to the recommended level, while China showed somewhat greater variation. Brazil and India showed both higher price ratios and greater variation.
The reasons behind these regional differences were not directly measured either. Purchasing power and exchange rates, as well as regulatory factors and the risk of cross-border purchasing, provide context for thinking about international pricing, but the study did not identify what specifically drove the ratios seen in Brazil and India. Without surveys of the people who set the prices, or experiments that manipulate conditions, it is not possible to determine companies' intent from the observed differences alone.
19 Affiliated Companies and the Skew Toward High-Variance Pricing
The research team also examined where the 19 companies confirmed to have affiliations or connections to major global gaming companies fell within the four groups. Eleven of them were classified in the High-Variance Portfolio group, accounting for 37.9% of that group's 29 companies. Across the entire dataset, these 19 affiliated companies represented 7.1% of the 268 companies analyzed.
In the other groups, one affiliated company appeared in the Benchmark Follower group (1.9% of 54), four in the Adjusted group (3.4% of 116), and three in the Emerging Market Premium group (4.3% of 69). These concrete figures support the observation that affiliated companies were more heavily represented in the High-Variance Portfolio group.
Even so, no causal relationship can be drawn showing that corporate affiliation caused the pricing variation. The research team did not directly measure companies' internal price-setting structures or pricing capabilities, and this was not an intervention study controlling for factors other than company size. It also does not support the conclusion that larger companies have more sophisticated pricing strategies. What could be confirmed was a skew in affiliation, nothing more.
2023 Data Versus the 2026 Price Table
On March 27, 2026, Valve updated Steam's price conversion data to reflect global market conditions. At the same time, it introduced three methods for converting U.S. dollar prices into 37 currencies and four regional groups: a method based purely on exchange rates, a method based purely on purchasing power, and a method that combines purchasing power, prices of comparable entertainment products, and exchange rates.
Publishers still make the final pricing decisions today. Valve's official documentation explains that developers can choose which conversion method to use, or set their own custom prices. The coexistence of a common benchmark and corporate discretion, which the study captured, still applies to the current system, but the approximately 128% and 130% figures were calculated using the recommended prices that existed before the update as their denominator. They cannot be repurposed as comparison values against the current recommended prices.
The paper, co-authored by Murakami and Yoshida, was published online ahead of print on August 21, 2026, in SAGE Publications' Journal of International Marketing. The authors' official list of publications includes it as a peer-reviewed academic paper. However, this is a single study just published, and whether the four categories would be reproduced at other points in time or on other distribution platforms remains unconfirmed.
To further validate these findings, it would be necessary to repeat the same analysis after the March 2026 conversion method update and track how companies chose among the three conversion methods. If sales volume and revenue could also be linked to this analysis, it would move beyond simply establishing the existence of four categories, allowing for the first assessment of how each pricing pattern relates to business outcomes.