1. Introduction
On 4 June 2026, the Stockholm District Court’s Patent and Market Court (the Court) set aside a fine imposed by the Swedish Competition Authority (SCA) on Min Doktor, digital primary care provider, for agreeing to refrain from bidding on keywords containing its competitor’s trademarks on Google Search (a so-called ‘non-brand bidding agreement’, NBBA).[1] While NBBAs have been subject to antitrust scrutiny before,[2] this judgment marks the first time that an EU court has addressed the question of whether NBBAs between competitors restrict competition “by object”. The Court’s judgment has significant relevance for advertisers and overlaps with pending investigations into similar NBBAs by the European Commission and the Swiss competition authority.
2. Background
The case originates from an investigation opened by the SCA in 2022 in response to a leniency application by Kry, a digital primary care service provider. The investigation concerns agreements Kry had entered into with three of its competitors: Min Doktor, Dokter.se, and Doktor24.[3] In early 2020, Kry separately agreed with each of these companies to refrain from bidding on each other’s brand names as keywords on Google’s search advertising platform, Google Ads.
According to the SCA, the agreement was designed to restrict consumers’ ability to become aware of alternatives when Googling a specific primary care provider’s name names, meaning that the parties had stopped competing with one another in such situations. The SCA thus found that these agreements constituted “by object” restrictions under Article 101 of the Treaty on the Functioning of the European Union (TFEU) and its national equivalent and imposed fines of SEK 26 million (approximately EUR 2.3 million) on the three companies. Min Doktor, which was fined SEK 6 million (approximately EUR 594,000) appealed the decision before the Court.
3. Summary of the ruling
The Court’s central analysis focused on whether the agreement was, by its very nature, so harmful to competition that no assessment of its effects was required, i.e. on whether it restricts competition “by object” within the meaning of the jurisprudence of the Court of Justice of the European Union (CJEU)’s on Article 101 TFEU, most recently in Tondela.[4]
The Court began by noting that the agreement did not fall within the category of conduct traditionally regarded as most harmful to competition, such as price-fixing, restrictions on production capacity, or market sharing. Moreover, the agreement at issue in this case differs from those previously addressed by the European Commission and Bundeskartellamt respectively in Guess and Asics: those cases concerned restrictions on distributors’ ability to bid on the trademarks of their supplier as part of a broader set of anticompetitive restrictions within a distribution system. Thus, the Court concluded that a thorough examination of the agreement’s content, objectives, and crucially, legal and economic context would be necessary to conclude it constitutes a by-object restriction.
a. Content and objectives
On the content and objective aims of the agreement, the Court found that parties had agreed to refrain from appearing as a paid advertiser on Google Search when the other party’s brand name was used as a search term. This meant that, in effect, consumers searching for “kry” would not be shown paid advertisements for Min Doktor, and vice versa.
The parties, however, implemented the agreement differently: Kry added “min doktor” as a negative keyword, meaning that Kry’s advertisements would not be displayed on any Google search queries which included “min doktor”. A consumer who searched for “min doctor” would thus not be presented with Kry’s advertisements, even if the search query included other terms (e.g. “min doctor back pain”). Min Doktor on the other hand merely paused keywords containing “kry”, with the result that Min Doktor’s ads could still appear in searches which combined “kry” with other terms. A consumer who searched for “kry” would thus not be presented with any of Min Doktor’s advertisements, but if they searched for “kry back pain”, they might still be presented with an advertisement for Min Doktor’s services.
The Court accepted that the agreement had an objectively anti-competitive purpose in the sense that the parties refrained from competing for potential patients in this specific marketing channel. Interestingly, the court also recognised that the primary motivation of the parties was to better allocate search advertising costs, which is in itself a legitimate purpose. However, the Court did not consider this to have any significance for competition law, since these savings were not passed on to consumers, meaning that this did not constitute a pro-competitive purpose.
b. Context: bidding on trademarks was not an important competitive parameter
The Court’s analysis of the legal and economic context largely revolved around the question of whether the agreement concerned a significant parameter of competition. SCA had found that this was the case, as Min Doktor’s advertisements on “Kry” garnered more impressions than most other advertisements, indicating that it was a significant parameter of competition.
The Court disagreed with the SCA’s assessment. According to the Court, advertisers largely based their search engine marketing decisions on clicks, conversions, and the cost-per-click (CPC). Impressions are thus not the most important factor for advertisers when allocating marketing budgets.
The search term “kry” ranked among the top keywords in terms of ad impressions, clicks, and conversions compared to other search terms. The click-through rate on “kry”, however, was only 3.36%, and conversions as a proportion of impressions amounted to merely 0.36%. This was very low compared with (i) Min Doktor’s own brand name and (ii) with generic healthcare terms. According to the Court, this indicates that “kry” is a “navigational search” term, meaning that consumers who search for “kry” have generally already decided to use Kry, and are therefore less likely to engage with a competitor’s advertisement. Thus, according to the Court, it was less important from the consumer’s perspective that parties advertise on search queries for one another’s trademarks.
In addition, the Court noted that Min Doktor reallocated the money it saved to bid on other keywords, and that the companies continued to compete with one another for other relevant key words, such as generic healthcare terms. Moreover, the costs associated with bidding on “kry” made up less than 1% of Min Doktor’s total marketing costs.
c. Economic evidence on the matter is inconclusive
The Court also considered economic research submitted by Min Doktor. It noted that studies by other competition authorities and academic researchers did not provide a clear picture that bidding on competitors’ trademarks is beneficial from a competition perspective, nor that agreements restricting such bidding are inherently harmful to competition. On the contrary, academic literature suggested that the competitive effects of such agreements vary significantly depending on the circumstances of the individual case.
d. Impact of trademark and marketing law
The Court also examined whether the NBBA could be justified on trademark or marketing law grounds as part of its assessment of the legal context of the agreement. While the Court acknowledged that using a competitor’s trademark in keyword advertising can raise trademark concerns, it found no evidence that either party had damaged the other’s trademark functions, caused dilution, or gained unfair advantage. Similarly, although Min Doktor argued that healthcare patients are particularly vulnerable to misleading advertising and that the agreement could have addressed this, the Court found no evidence that the marketing environment at the time was characterised by such problems. As a result, neither trademark law nor marketing law provided a valid justification for the agreement, and these arguments carried no weight in the broader competition law assessment.
4. Analysis
This judgment illustrates the importance of rigorous contextual analysis when assessing whether an agreement can be classified as a restriction of competition “by object”, especially when dealing with new forms of potential anticompetitive agreements that have not been labelled before. The Court reached its conclusion that the NBBA in this case had objectively anti-competitive purpose relatively easily. Nonetheless, since the economic literature did not clearly indicate that NBBAs are anti-competitive and search queries on the names of competitors did not actually appear to be an important marketing tool for the competitors, but more a means for Google to replace organic search results with paid results that trigger a payment to Google if clicked, the Court considered that the agreement in this case was not sufficiently harmful to merit being classified as a restriction of competition by object.
This approach aligns with the CJEU’s case law as also explained in a recent judgment in Tondela, in which the CJEU reaffirmed that even categories of agreements which are “particularly harmful to competition” can under certain circumstances nonetheless not restrict competition by object.[5] The Court moreover emphasises the importance of prior enforcement experience in determining the degree of scrutiny which must be applied to a given form of coordination, reflecting Advocate General Emiliou’s opinion in Tondela,[6] and the CJEU’s ruling in Budapest Bank.[7] Taken together, this suggests that competition authorities and private parties should conduct thorough contextual analysis before trying to put new theories of harm into the ‘by-object box’, or risk their decisions being overturned.
The judgment also mentions, though in the authors’ view still underrates, the structural issues faced by businesses in relation to online search: Google’s dominance in search and search advertising. As Min Doktor noted, Google operates a de-facto monopoly in online search, as a core platform for businesses to be found, not just in Sweden. Google is seeking to exploit such a monopoly by increasing the share of end users’ clicks on paid results, rather than on organic results. Creating a mechanism by which business users of Google Search are compelled to place ads for their own brand names is one of those tactics. Google’s auction-based pricing model means that the more keywords advertisers bid on, and the higher they bid, the higher the cost per click for all participants, and Google’s search revenues. Min Doktor had argued that the SCA should have considered this context when assessing the agreement, an argument which the Court surprisingly did not address.
In fact, the way Google has set up its brand-keyword bidding to extract rents from businesses that are dependent on being found if end users search for their names appears central to the case. Advertisers have long complained that Google does not just tolerate but actively facilitates the bidding on competing trademarks, even though the auctions are designed to leave businesses with no other choice but to bid on their own registered trademarks, and to raise their own bids in Google Ads to outbid competitors if they wish to defend the top-ranking position for their own trademark. Importantly, the Court’s judgment recognises that consumers’ search queries containing trademarks tend to be navigational searches, suggesting that most of the advertising budget spent on such bids by trademark holders is better spent elsewhere.
With the Digital Markets Act, the EU legislator has aimed to prevent tech firms with important positions in the online economy from taking advantage of their position to the detriment of dependent business users. Given Google’s own enormous market power, as well as the lack of economic evidence that NBBAs are harmful, this raises the question of whether competition authorities’ limited resources could be better used investigating Google, rather than the business users it exploits.
[1] Stockholm District Court, Patent and Market Court, Case number PMÄ 8348-25, final decision issued 4 June 2026.
[2] In the EU NBBAs have previously been assessed by the European Commission in Guess (Case AT.40428) and by the Bundeskartellamt in Asics (Case B 2-98/11) in the context of selective distribution agreements as well as by the Dutch Authority for Consumers and Markets (ACM) in a study about their use by hotel online travel agents. (ACM Working Paper, “Working paper Price effects of non-brand bidding agreements in the Dutch hotel sector”, 7 June 2019, Available at: https://www.acm.nl/sites/default/files/documents/2019-06/working-paper-acm-price-effects-of-search-advertisement-restrictions.pdf). Outside the EU, the New Zealand Commerce Commission fined Moola, a loan provider, for entering into NBBAs in 2021,[2] and the US Federal Trade Commission (FTC) filed a complaint against contact retailer 1-800 Contacts regarding such agreements in 2016, though this was ultimately overturned (See 1-800 Contacts, Inc., Docket No. 9372, Available at: https://www.ftc.gov/system/files/documents/cases/160808_1800contactspt3cmpt.pdf).
[3] SCA Press Release, “Online healthcare companies fined for anti-competitive agreement on advertising digital health services”, 3 April 2025, Available at: https://www.konkurrensverket.se/en/news/2025/online-healthcare-companies-fined-for-anti-competitive-agreement-on-advertising-digital-health-services/.
[4] Judgment of the CJEU in Case C-133/24, Tondela (2026) ECLI:EU:C:2026:361.
[5] Judgment of the CJEU in Case C-133/24, Tondela (2026) ECLI:EU:C:2026:361.
[6] Opinion of AG Emiliou in Case C-333/24, Tondela (2025) ECLI:EU:C:2025:364.
[7] Judgment of the CJEU in Case C-228/18, Budapest Bank (2020) ECLI:EU:C:2020:265.
Leave a Reply