Article By Can Sarıçiçek, Özlem Başıböyük Coşkun, Nadide Akdağ and Berkay Ünlüsoy

1. The Competition Board (“Board”) has concluded its investigation into DSM Grup Danışmanlık İletişim ve Satış Ticaret A.Ş. (“Trendyol”), a major player among multi-category online marketplaces. According to the allegations, Trendyol leveraged its market dominance in Türkiye by (i) manipulating algorithms and (ii) using third-party sellers’ data to benefit its retail operations, actions which restricted competition and violated Act No. 4054 on the Protection of Competition (“Competition Law”). Initiating the investigation the Board imposed interim measures, consisting of mainly withholding any actions and changes into the algorithms that may benefit Trendyol’s retail operations and withholding of use and share of any data accumulated and aggregated within the online marketplace operations for the benefit of its own retail operations until the final decision is given. Following the investigation, the Board imposed both administrative fines and various obligations on Trendyol to address the violations and restore effective competition in the market. This article examines Trendyol’s business model and sales operations in general, analyses Trendyol's anti-competitive practices, and explores the key findings presented in the investigation decision.

2. Trendyol operated solely as a retailer from its founding in 2009 until October 2017, when it also started to provide online marketplace platform services for third party sellers. This expansion transformed Trendyol into a hybrid online marketplace, simultaneously operating both as a seller and an intermediary platform service provider connecting third party sellers with consumers, within the same platform. Under the Trendyol Private Label business model, Trendyol conducts sales in the fashion category through brands such as TrendyolMilla, TrendyolMan, TrendyolKids, TrendyolModest, TrendyolShoes, and Nottis by Trendyol (“PL Brands”). The investigation’s analysis of self-preferencing behaviour focused specifically on Trendyol's activities as a seller through these PL Brands.

3. The on-site inspections conducted at Trendyol revealed evidence spanning from August 2017 to September 2021 that showed Trendyol’s systematic manipulation of algorithms through various practices and unfair use of marketplace data to benefit its own PL Brands.The decision notes that following its 2017 transition to a marketplace model, Trendyol steadily increased its market share, ultimately reaching a dominant position in the multi-category online marketplace market by 2020. The competitive pressure in the market was further evidenced by the exit of Gittigidiyor in 2022, a subsidiary of eBay, which failed to maintain its market position. While evidence of Trendyol’s self-preferencing practices dated back to 2017, the company’s legal liability began in 2020 when it achieved market dominance and continued through September 2021. This article now turns to a detailed examination of Trendyol’s anti-competitive practices: its unfair use of third party seller data and manipulation of the algorithms to benefit its PL Brands.

Trendyol's Use of Third Party Seller Data for Its Own Benefit

4. On-site inspections revealed evidence of Trendyol’s use of third party seller data to the advantage of its own operations. Among these was a document of internal communications that revealed that Trendyol had been testing an econometric model for TrendyolMilla product sales forecasts, build upon TrendyolMilla category sales data, competitor brand sales data and similar product sales data and that the model is excelled at content based predictions. Another notable document revealed internal communication about marketplace data usage. The exchange began with questions about the legal implications of accessing marketplace reports containing category and brand-based data. Within this discussion, a Trendyol employee disclosed their ongoing practice of analyzing marketplace data, emphasizing its usefulness for brand-category insights, including lingerie segment analysis. The conversation extended to another employee requesting feedback regarding sales in the modest fashion category and winter trends, referencing a similar previous analysis conducted for the children's clothing category.

5. Interviews with third party sellers revealed that due to the strong position of PL Brands in the fashion category, concerns exist regarding the data Trendyol possesses. Third party sellers expressed worry that Trendyol might produce similar products by monitoring the best-selling items, including their designs and colours. It was also indicated that Trendyol could manufacture products by examining data on seasonal trends, product costs, stock, and other characteristics in the fashion sector. Trendyol’s access to customer data, including which products are bought by specific customers at particular price points, could enable direct targeted sales and marketing initiatives. The launch of TrendyolMilla and TrendyolMan had a negative impact on third party sellers’ revenues. Several third party sellers reported that after they offered their new products, they noticed similar items appearing in TrendyolMan’s collection, at a much lower price. It was also claimed that Trendyol’s developing PL products based on their best-selling items adversely affected their sales. It was noted that Trendyol could view existing product designs and identify popular items, enabling it to create similar designs for its branded products while avoiding the design or any other business process costs and risks. By selectively producing proven successful items, Trendyol gained a competitive advantage over other brands. Third party sellers expressed that Trendyol’s comprehensive access to Seller data, when used to benefit TrendyolMilla, created an unfair competitive advantage.

Evaluations Regarding Trendyol's Alleged Use of Collected Data for Its Own Benefit

6. Data accumulated from both third party sellers and end-users, including voluntarily provided data and data aggregated throughout their use of the platform services, provides Trendyol with a broad perspective on market intelligence. This information asymmetry between Trendyol and third party sellers yields insights for Trendyol into multiple market parameters such as price points, demand elasticity, supply chain dynamics, and consumer behaviour patterns. Leveraging this position, Trendyol can strategically enter the market either as a reseller or manufacturer or expand its market presence, while circumventing traditional market entry barriers and risk assessment costs typically faced by new market entrants. It is also stated that insights as such can also provide Trendyol with the ability to easily steer its marketing and advertising expenses in the direction of the consumer preferences and benefit from targeted advertising at its peak.

7. Based on third party seller interviews and the documents obtained during the on-site inspections, the Board determined that Trendyol used marketplace data to improve sales forecasts for its own PL Brands, including its modest fashion line and TrendyolKids. The Board’s investigation revealed an intended expansion of marketplace data utilization into new product segments, mainly targeting the lingerie line. Furthermore, the Board uncovered evidence that Trendyol’s PL Brands division had systemically leveraged marketplace insights for competitive advantage. This operational strategy was justified internally by characterizing competitively sensitive marketplace data as generally available market intelligence. Accordingly, the Board concluded that Trendyol had positioned marketplace data accumulated through third party sellers for immediate analysis and production purposes. This system, in return, allowed for a generation of customized reports from this data, which could be utilized to benefit Trendyol’s own brands.

8. Internal communications obtained during the on-site inspections at Trendyol included employee-expressed concerns regarding the illegality of such data usage. In this respect, the Board considered these documents to reveal that any company employee could reasonably recognize that using this data to forecast or improve PL Branded products would undermine fair competition conditions. Considering the documentation demonstrated Trendyol’s extensive data collection and analysis capabilities regarding third party seller and end-user activities, the Board interpreted this marketplace-derived data capability as potentially creating competitive implications when applied to private label operations would lead to earning of unfair competitive advantages for Trendyol and thus could restrict competition in the marketplace.

9. Analyzing the evidence and third party seller testimonials, the Board determined that Trendyol’s marketplace infrastructure allowed Trendyol to see which categories and products, the third party sellers, namely competitors in the marketplace sold the most, allowing it to closely and thoroughly monitor the current and individual sales data of the market-leading third party sellers in the relevant category/product to be then used for forecasting and/or designing its own products and as an input for other commercial or strategic decisions.

10. The Board's assessment indicated that Trendyol's intermediary position provides access to commercially valuable consumer behavior data, including product preferences, demographic patterns, and consumer feedback - information not equally available to third party sellers. This information asymmetry created by data accumulation abilities, combined with Trendyol's dual role as both marketplace operator and seller, enables strategic decision-making in its private label operations including marketing and advertising decisions. While Trendyol shares certain reports with third party sellers, the Board found this insufficient to eliminate the inherent information disparity.

Trendyol's Intervention in Algorithms and Product Rankin

11. Evidence related to algorithm and ranking interventions obtained during on-site inspections at Trendyol revealed that Trendyol employees made requests that the software is designed to boost the ranking of PL Brands, with these implementations on the algorithms being documented in system records after getting fulfilled. Trendyol also concealed the low follower counts of its proprietary brands like TrendyolMilla and TrendyolMan to maintain their credibility. This preferential treatment extended to the algorithmic architecture itself, where Trendyol established an exclusive category for its private label offerings. A structure known as "torpil adjustment" was identified by the case-handlers as a mechanism that systematically amplified the visibility of the company's brands. By implementing higher multiplication coefficients for its own products' scores, Trendyol created a ranking system that inherently favored its offerings over competing brands on the platform, thus pushing their products to the forefront in rankings. In a Slack conversation within the company, the Board noted that a brand selling in the children's category had fallen behind in organic rankings despite receiving ranking support, and it was requested that this brand to be moved up in brand listings. Trendyol artificially manipulated the follower counts of both its brands and third-party brands operating on the marketplace to increase these follower numbers. When evaluating the documents comprehensively, it was determined that since 2017, Trendyol aimed to highlight its own brands in the product rankings that included all the third party sellers and/or brands, put differently its competitors, in the marketplace.

12. Third party seller interviews during the investigation highlighted the critical importance of securing top positions on product listing pages, as prominent placement drives buyer engagement. In the e-commerce environment, product visibility directly correlates with click-through rates and conversion potential. Products displayed at the top receive greater exposure and conversion rate compared to lower-ranked items, while also benefiting from enhanced brand visibility. Given that Trendyol searches generate thousands of results and users typically limit their scrolling, top positioning becomes particularly significant.

Evaluations Regarding Trendyol's Intervention in Algorithms and Product Ranking

13. The Board evaluated that Trendyol possessed full control over both the design and operational principles of its search ranking algorithm. Evidence collected showed Trendyol to be acting in a way that brings its PL Branded products to the forefront/top rankings in the results displayed in response to searches. Recognizing that users usually gravitate towards top searches, the Board found documentation in Jira records and the ranking algorithm specifications revealing the implementation of the “torpil adjustment” mechanism. This system artificially increased the product scores of Trendyol PL products by multiplying them with higher coefficients, thus preparing the necessary infrastructure for these PL products to be shown at the top of the rankings.

14. Although these documents date back to 2017, the Board determined that such self-preferencing behaviour towards Trendyol PL Brands and the related algorithmic intervention could have lasting implications for future search rankings since the algorithm will keep functioning on the base once it was provided with and has the ability to self learn. The Board listed two main implications for future search rankings. Firstly, since the third party sellers are downgraded, they will receive less clicks, less comments and less conversion rate, which will create a negative feedback loop, and they will rank lower due to less comments less clicks and less conversion. Second, Trendyol will have the ability to keep strengthening its position in the platform and will be protected from any competitive pressure. This will also have far reaching implications that Trendyol will lose the incentives to innovate, both for the products and both for the online platform services, Trendyol will also lose the incentives to supply new, better quality and/or better priced products than its competitors.

15. Among the documents obtained during the on-site inspection, case handlers discovered a task explicitly detailing plans to modify Smartlisting algorithm’s raw scores based on delivery options. The documentation revealed that the goal was to move products with fast delivery options up in the rankings by multiplying their listing scores with higher coefficients. Accordingly, it was instructed that when TrendyolMilla offered a fast delivery option for a product, the fast delivery coefficient would be taken and multiplied by the product’s raw score, whereas for another brand offering a fast delivery option, this value would be taken as the default unless otherwise specified. Further evidence also revealed that while competitor brand follower counts were artificially inflated by a factor of three, Trendyol's own brands (TrendyolMilla, TrendyolMan, and TrendyolKids) received a five-fold multiplication of their actual follower numbers.

16. In this context, the Board performed an extensive effects analysis and demonstrated that third party sellers’ sales on Trendyol were distinctively higher than their sales on other online marketplaces, emphasizing the indispensability of Trendyol for them. Furthermore, the Board examined the growth in sales volume of Trendyol’s PL products and third party sellers’ products, in macro and micro levels. The comparison provided the insight that, while Trendyol’s PL products showed a consistent growth scheme, competitors’ products failed to do so.

17. Therefore, it is considered that Trendyol’s self-preferencing behaviour towards its PL Branded products in the fashion category through data and algorithms contributed to maintaining its revenues in the multi-category online marketplace market. In other words, Trendyol’s anti-competitive actions, allowed it to finance other categories and/or its overall online marketplace activities, potentially causing adverse effects in the competition in the broader online marketplace sector. Trendyol’s anti-competitive actions, were also considered to restrict the third party sellers from reaching out to the end users which will have a direct impact on their sales volumes.

18. When the evidence, documents and opinions from the sector stakeholders regarding algorithm and ranking interventions are evaluated comprehensively, it is seen that Trendyol has aimed to highlight its own brands in the product rankings that included all the third party sellers and/or brands in the marketplace since 2017. Consequently, from the documents obtained during the on-site inspection within the current file, it has been concluded that Trendyol intervened in the algorithm in a way that favoured its brands, TrendyolMilla, TrendyolMan, and TrendyolKids, over their competitors in the fashion category, making it difficult for other third party sellers operating on the Trendyol marketplace to compete and/or leading to their exclusion, thus harming intra-platform competition.

Conclusion

19. The Board concluded that Trendyol held a dominant position in the multi-category online marketplaces market and exploited this position through algorithmic manipulation and use of marketplace seller data to gain unfair advantages in its retail operations, thereby restricting competition. As a result, an administrative fine of TRY 61,342,847.73 (approx. EUR 5,858,915.734), based on its annual gross revenue at the end of the 2021 fiscal year, was imposed on Trendyol. Additionally, to establish effective competition, the Board decided that Trendyol must take various measures and fulfil certain obligations, with some Board members presenting different justifications and unanimous votes.

20. The obligations Trendyol must fulfil include removing exclusivity clauses from the agreements signed with the third party sellers, measures to provide transparency for sellers and end users and limiting the advertising and marketing budget up to a percent of its commission revenues in order to terminate the predatory pricing practices via advertising expenditures.

21. The dissenting opinions of two Board members highlighted the key reasons for their disagreement with the imposed penalty on Trendyol. First, they argued that setting the base fine rate for Trendyol at a symbolic level fails to account for the substantial harm caused by its market power and anti-competitive behaviours, is neither fair nor consistent with the Regulation on Fines to Apply in Cases of Agreements, Concerted Practices and Decisions Limiting Competition, and Abuse of Dominant Position (“Regulation”). They pointed out that in precedent cases involving digital platforms, the Board had applied significantly higher base fine rates at 2-3%.

22. Additionally, the dissenting members stated that Trendyol had reached a dominant position in the “multi-category online marketplaces market” by 2020, as established by the Board. Considering that the most recent document obtained during the on-site inspections is dated September 2021, they reasoned that Trendyol should be held responsible from 2020 until September 2021. Consequently, they argued that the fine imposed on Trendyol was not in accordance with the Regulation and that the duration should be set to longer than one year but less than five years, with the base fine rate increase of 50%.

23. In light of the precedent cases and the long duration of the violation, the dissenting members concluded that the fine amount should be TRY 552,085,629.57 (approx. EUR 52,730,241.602), which is nine times the base amount. This amount, they contended, would better address the competitive harm incurred and align with the Regulation, and thus they opposed majority’s decision.