Commission, February 18, 2010, No M.5727

To the notifying party:

Dear Sir/Madam,

Subject: Case No COMP/M.5727 - Microsoft/ Yahoo! Search Business

Notification of 15 January 2010 pursuant to Article 4 of Council Regulation No 139-2004 (1)

1. On 15 January 2010, the Commission received notification of a proposed concentration pursuant to Article 4 of Council Regulation (EC) No 139-2004 ("EC Merger Regulation") by which Microsoft Corporation ("Microsoft", United States of America) acquires within the meaning of Article 3(1)(b) of the EC Merger Regulation control of the online web-wide algorithmic search ("internet search") and search advertising businesses of Yahoo! Inc. ("Yahoo", United States of America) including Yahoo's online search advertising platform Panama ("Yahoo search business") by way of purchase of assets.

I. THE PARTIES

2. Microsoft is active in the design, development and supply of computer software and the supply of related services on a worldwide basis. The transaction concerns Microsoft's Online Services Business division and, more specifically, its internet search platform, Bing, and its online search advertising platform, adCenter.

3. The Yahoo search business encompasses the internet search and the online search advertising ("search advertising") businesses of Yahoo, including its online search advertising platform Panama.

II. THE OPERATION

4. On 4 December 2009, Microsoft and Yahoo entered into a License Agreement and a binding Search and Advertising Services and Sales Agreement ("the Agreements").

5. The Agreements provide that Microsoft will acquire a 10 year exclusive license to Yahoo's core search technologies, and will have the right to integrate Yahoo's search technologies into its existing web search platforms. Microsoft also agreed to hire no less than 400 employees, [...] of whom from Yahoo's search advertising group and [...] from Yahoo's internet search group.

6. Furthermore, under the Agreements, Yahoo will exclusively use Microsoft's search engine on Yahoo sites. Microsoft will also become the exclusive search advertising provider used by Yahoo. In exchange, Microsoft will retain 12% of the search revenues generated on Yahoo's own and its partners' sites during the first 5 years of the agreement, paying 88% to Yahoo as a Traffic Acquisition Cost ("TAC"). Thus, under the Agreements, Yahoo will not operate a search engine or a search advertising platform of its own at least for 10 years. Yahoo will however retain the right to design the user experience when presenting internet search results received from Microsoft.

7. Once Microsoft search advertising is launched in a given country, Panama will be discontinued and Yahoo's customers will be "migrated" to adCenter, which will provide the auction processes for search advertisers for both Microsoft and Yahoo.

8. The Agreements distinguish between two types of customers: (1) self-served customers, which purchase advertising space on Microsoft's and Yahoo's search pages online via adCenter and/or Panama, without interaction from a sales force; and (2) Premium Direct Advertisers ("PDAs"), which also purchase advertising space through Microsoft's and Yahoo's advertising platforms, but in addition have a contract with Microsoft's or Yahoo's sales force in which additional services such as keyword optimisation and rebates may be agreed.

9. Post transaction, whereas self-serve advertisers will be directly served by Microsoft's adCenter, Yahoo will become the exclusive worldwide relationship sales force to PDAs for the services provided by adCenter.

10. The Agreements do not cover any other aspect of the companies' businesses, so each company's web properties and products, email, instant messaging and display advertising will remain separate. Microsoft will however enable Yahoo, on a non-exclusive basis, to integrate on all Yahoo properties and syndication properties (2) Microsoft's contextual, display, social and video advertising services, mobile search advertising as well as other forms of online advertising (for example classified, directories, and vertical listings).

11. As noted above, Yahoo will continue to determine independently the content and user interface of its sites, and may continue to use its search technology and data in other areas of its business, such as enhancing display advertising technology. The Agreements also will not affect Yahoo's right to work with current and future syndication partners.

III. CONCENTRATION

12. The Agreements entail the acquisition by Microsoft of Yahoo's internet search and search advertising businesses, including Panama. These businesses are not incorporated, and their transfer is brought about via a transfer of relevant technology, employees (including customer relations staff) and a migration of customers from Panama to adCenter.

13. The Commission finds that the Agreements entail a concentration falling under the EC Merger Regulation on the grounds that (i) the transferred assets constitute the whole or a part of an undertaking, that is a business with a market presence to which a turnover can be attributed and (ii) there is change of control over these assets which occurs on a lasting basis. The Commission further finds that Microsoft will acquire sole control over the Yahoo search business. (3)

Whole or a part of an undertaking

14. Search and advertising platform technology, human capital and advertising customers which are included in the Agreements appear to be the three most important ingredients for a search advertising business. Looking at those three dimensions of the transaction shows that it concerns a business with a market presence to which a turnover can be attributed.

15. First, and as indicated above, under the Agreements, Microsoft will obtain access to all the core technologies of the Yahoo search engine. Since the licence is also exclusive as to Yahoo, Yahoo will not be able to operate a separate search business.

16. Second, Microsoft will also hire at least 400 Yahoo employees corresponding to [...] of Yahoo's current search business staff. This proportion of employees of Yahoo's search business would be sufficient to enable Microsoft to run that business.

17. Third, once Microsoft search advertising is launched in a given country, Yahoo customers will be "migrated" to Microsoft's advertising platform adCenter, Yahoo's Panama platform will be discontinued and all search advertising services used by Yahoo will be provided by Microsoft through adCenter. Some advertisers will be redirected via an assignment of contracts although most of the smaller ones will be redirected via special migration campaign assistance (including online re-direction, migration marketing and assistance campaigns). PDAs will be required to sign new standard contracts.

18. Furthermore, for 2008, world-wide turnover of EUR [...] million can be attributed to the Yahoo search business. The Commission underlines that Yahoo will be contractually obliged to exit the search and advertising markets (see paragraph 6 above).

19. The transaction therefore concerns the acquisition of the whole or a part of a business with market presence to which turnover can be attributed.

Change of control on a lasting basis

20. The Agreements have a term of 10 years, which is a long period in a market characterised by rapid technological developments, such as internet search and search advertising. Defined events which would permit an early termination of the Agreements seem unlikely to occur and, in any event, could only take place after a significant time period. In case of termination no forced return of employees is foreseen and Microsoft would maintain the license to the core technology (although this will become non-exclusive). Therefore, for the foreseeable future Yahoo is de facto definitively divesting its ability to compete in internet search and search advertising.

21. As a result, the transaction would bring about a change of control on a lasting basis.

Acquisition of sole control

22. As the Agreements do not create a joint venture in which the parties share equally the voting rights, joint control can only arise out of (i) veto rights with respect to decisions that are essential for the strategic commercial behaviour of a joint venture; or (ii) commonality of interests.

23. Veto rights that can be relevant for establishing control are those concerning budgets, appointment of management, business plans and investments, as well as other market specific rights (4) such as choice of technology or pricing policy.

24. Following this transaction, Yahoo will be exclusively responsible for contacts with PDAs, and thus will be able to give rebates to such customers. This however is not in itself sufficient to amount to a strategic decision conferring control over a business, since rebates to advertisers seem to be a standard feature of this business, at least in Europe.

25. On the issue of commonality of interests, Yahoo is not offering a "vital contribution" to the business other than internet traffic and sales services. Therefore, there is not a sufficiently "high degree of mutual dependency" among Yahoo and Microsoft to reach the "strategic objectives" of a joint venture (5) beyond the common interest inherent to any long-term commercial agreement.

Conclusion

26. Therefore, the transaction constitutes an acquisition of sole control by Microsoft over a part of an undertaking (the Yahoo search business) by purchase of assets within the meaning of Article 3(1)(b) of the EC Merger Regulation.

IV. COMMUNITY DIMENSION

27. The undertakings concerned have a combined aggregate world-wide turnover of more than EUR 2 500 million (6) (Microsoft: EUR 39 730 million (7), the Yahoo search business: EUR [...] million (8)). In each of France, Germany and the United Kingdom ("UK"), the combined aggregate turnover of the two undertakings is more than 100 million (EUR [...] million, [...] million and [...] million respectively) and the aggregate turnover of each of the undertakings concerned is more than EUR 25 million in each of these three Member States (9). Finally, each of Microsoft and the Yahoo search business have a Community-wide turnover in excess of EUR 100 million (Microsoft: EUR [...] million, the Yahoo search business: EUR [...] million) and they do not achieve more than two-thirds of their aggregate Community-wide turnover within one and the same Member State.

28. The notified operation therefore has a Community dimension within the meaning of Article 1(3) of the EC Merger Regulation.

V. RELEVANT MARKETS

A. DESCRIPTION OF THE PRODUCTS AND SERVICES CONCERNED

29. The proposed transaction concerns internet search and search advertising services.

1. Internet search

30. An internet search engine is a tool designed to search for information on the internet. It consists of a search box in which queries can be typed. The search results of a given query are then usually presented in a ranked list of results. The information searched for may consist of text (including news), maps, images, videos or other types of content. Search engines such as those operated by Microsoft, Yahoo and Google use algorithms to search for information on the internet and rely on large indexes that try to map the content of the World Wide Web.

31. General internet search must be distinguished from vertical internet search, which focuses on specific segments of online content such as for example legal, medical, or travel search engines. Contrary to general internet search engines, which index large portions of the internet through a web crawler (10), vertical search engines typically use a focused crawler that indexes only web pages that are relevant to a pre-defined topic or set of topics.

32. Internet search must also be distinguished from site search covering only the content of one particular website. One example of such site search is so-called "social search" performed on social websites such as Facebook or MySpace. While the technology required for searching the internet is more sophisticated and powerful than the technology for searching just within one website, web-search technology can also be used for this more limited purpose. Internet search engines therefore also offer site search either as a stand-alone products or in the context of "syndication", which will be explained below.

2. Search advertising and other forms of online advertising

33. Despite the large R&D and investment costs associated with internet search, internet search services are today generally provided free of charge to users. Search engines are financed by advertising revenue generated by search advertisements ("ads") shown on the same page as the natural algorithmic search or "organic" results for a particular query. These search advertisements are known as "sponsored links". This means that, in practice, users conducting a search receive on a results page both organic search results provided by the internet search engine and sponsored links consisting of advertisements related to the query.

34. Search engines seek to maximise their revenue by selling "inventory" on their search results pages to advertisers. Advertisers aim to have ads placed in the best possible positions on the search engine results pages to obtain the greatest visibility and therefore the highest number of clicks.

35. Advertisers buy search advertising space for those sponsored links on the basis of keywords included in search queries of users. Keywords are "auctioned off" to advertisers through advertising platforms such as Panama, adCenter and Google's adWords where advertisers bid for keywords. Each advertising platform allocates a ranking to an ad based on how much advertisers bid per click for a given keyword query and how many clicks the platform expects each ad to generate (ad relevance). Depending on the outcome of the auction, certain ads will be "served" on the advertising search results part of the search results page in an order determined by the advertising platform algorithm. Advertisers only pay when the user clicks on the hyperlinked search directing the user to a web page determined by the advertiser ("landing page").

36. Search advertising is thus a type of online advertising that uses the internet for delivering marketing messages to attract customers. Online advertising comprises a wide number of techniques and formats that can be divided according to selection mechanism, their appearance or format, user device, or pricing mechanism (discussed below). All these different categorisations and criteria may lead to different permutations. It is for example possible to show sponsored text ads and/or display ads next to a list of video search results. Also, there seems to be a degree of convergence between the different forms of online advertising and new hybrid forms emerge frequently.

Distinction of online ads according to the selection mechanism

37. A first criterion to categorise online ads is the selection mechanism which is the way the ad is selected to appear on a user's screen. On this basis, there are two main categories: search ads and non-search ads.

38. Search ads are served next to the results of search queries entered by internet users into an internet search box. They are selected inter alia on the basis of the search keywords chosen by the user.

39. Non-search ads can appear on any web page and they can either be contextual ads, selected according to the content of the web page on which they appear, or non-contextual ads.

40. A growing number of both search and non-search ads are also behaviourally targeted. That means that in addition to other targeting criteria such as the search keywords used, the content of the visited web page or the geographical location of the internet user, the ads to be served are selected according to the individual's previous web-browsing behaviour. Information on that user behaviour is collected by using so-called "cookies".

41. A separate third category of ads (according to the selection mechanism) are classified ads. These ads are grouped within specific web pages under headings classifying the products or services being offered (for example houses for sale in particular suburbs, plumbers active in a specific town). These ads are different from both search and non-search ads because they are neither targeted to the content of the web page nor to any characteristic of the user. Instead they themselves constitute the main content of the page on which they are shown and they are shown to each visitor of that page alike.

Distinction of online ads according to their appearance or format

42. A second criterion to categorise online ads is their appearance or format. Online ads can for example take the form of text ads, display (graphical) ads or video ads. Ad inventory (the space provided by publishers where ads are placed) is rarely, if ever, restricted to a specific type of ad appearance.

Distinction of online ads according to the device on which they are shown

43. A third criterion to categorise online ads is the device on which they are shown. Most online advertising is designed for desktop or laptop use. By contrast, mobile ads are specifically designed for mobile devices such as smartphones. Mobile ads take the form of short text ads on which searchers can click to get to the advertisers' page or to call them directly with their mobile phone.

Distinction of online ads according to the pricing mechanism

44. A fourth criterion to categorise online ads is based on the pricing mechanisms for the online ads.

45. The so-called "cost per click" ("CPC") pricing mechanism is used mainly for search and contextual ads: the advertiser generally aims at a direct response from the user and pays accordingly only for such a direct response.

46. By contrast, the "cost per thousand impressions" pricing mechanism ("CPM") is mainly used for non-contextual and display ads. The advertiser pays the publisher when an ad has been displayed to users a number of times. Each showing of an individual ad is one "ad impression".

3. Main actors and distribution channels in search and search advertising

47. Internet Search engines such as www.search.yahoo.com, www.bing.com or www.google.com operate two-sided platforms serving both search users (for "free") and advertisers (for remuneration).

48. In order to be successful, a search engine operator will try to attract as many participants on both sides of the platform as possible. As part of this, search engines try to reach participants through various distribution channels on both sides of the platform.

Distribution channels on the side of search users

49. Users can access the search engines' internet search services in a variety of ways which are sometimes referred to as "entry points". Search engines try to maximise entry points to their engine in order to attract users' searches. Search ad providers can conclude two main categories of agreement in order to get more traffic on their search engine via these entry points.

50. Firstly, they can use so-called "distribution agreements" with hardware manufacturers (PC original equipment manufacturers, "OEMs"), Independent Software Vendors ("ISVs"), and Internet Service Providers ("ISPs") to distribute toolbars and establish default settings that direct user searches to the search engine. Searches start from toolbars, browser search boxes, auto-search (DNS search), widgets, and default home pages and the results of searches are displayed on the search engine's site (11).

51. Secondly, in addition to these types of entry points which, when used, cause the user to land on the search engine's website itself, users can also often access search services (and related search ads) indirectly on third-party websites who have concluded syndication agreements with a search engine and its advertising platform.

52. Syndication enables third party publishers to use a search engine's technology and an ad platform's pool of ads to deploy internet search services and search ads (as well as contextual text ads if the publisher so wishes) onto their websites. Publishers get paid a traffic acquisition cost ("TAC") by the search engine which usually is a percentage of the revenue the search engine makes from users' clicks on the ads shown on the publisher's website.

Distribution channels on the side of advertisers

53. Search engines distribute search advertising through two main channels, namely through automated search advertising platforms ("self-served advertisers") as well as through their own sales forces (human-assisted sales) which sell search advertising space to both advertisers and advertising agencies.

• Self-served advertisers

54. Self-served advertisers are usually the smaller advertisers in terms of spending thresholds who run and manage their ad campaigns entirely through the automated web interface. These advertisers are generally offered (largely technical) assistance during the bidding process through a call-centre or e-mail.

• Premium Direct Advertisers and media agencies

55. Microsoft's and Yahoo's sales forces (providing human-assisted sales services described hereafter) are primarily used to serve large advertisers (so-called Premium Direct Advertisers ("PDA")), who account for a substantial fraction of the parties' advertising revenues. Thus local sales teams assist large and mid-tier advertisers in creating and optimising their ad campaigns, through a team composed of sales specialists, trade marketing advisors, search specialists, etc.

56. The sales specialists discuss with the PDA how to increase the performance of search advertising, help upload the relevant keywords on the system, inform them of promotional opportunities and also serve as a knowledge centre to address questions relating to search advertising.

57. Very large online advertisers have strong internal expertise and, therefore, rely less on search engines' sales forces when designing their campaigns. Some PDAs use the automated search advertising platforms for performing updates for which it would not be efficient to go through human-assisted sales.

58. Media Agencies represent an important share of Microsoft's and Yahoo's human-assisted sales. They act as intermediaries for the advertisers and have employees specialised in search ads that assist advertisers in developing strategies for creating and managing their online advertising campaigns to maximise outcome, for example by helping them interpret search behaviour and by determining the relevant keywords to bid for and monitoring the performance of these keywords. Some agencies even offer account managers who carry out daily optimisation to ensure maximum efficiency.

59. Microsoft and Yahoo often assist the agencies in the creation of campaigns for the advertisers that use the agencies and in developing strategies maximizing the outcome of a campaign.

B. PRODUCT MARKET DEFINITION

60. The notifying party submits that the relevant product market encompasses the provision of online advertising services and, specifically, the provision of search advertising space.

1. Online advertising

61. In its recent Google/DoubleClick decision, the Commission established a distinction between online and offline advertising markets (12). In the present case, only one respondent to the market investigation submitted that online advertising and offline advertising belong to the same product market. The Commission therefore considers that online advertising is a distinct market from offline advertising.

a. Search advertising and non-search advertising

62. In Google/DoubleClick (13), the Commission considered whether search and non-search advertising are separate markets, but did not reach a conclusion in this regard because under any alternative product market definition, the transaction did not raise any competition concerns. In the present case, the notifying party argued that search advertising is a separate product market.

The notifying party's arguments

63. Microsoft has submitted the following arguments in support of its view that search advertising is a separate product market.

64. First, from the user's perspective, search advertising allows targeting of ads based on a real-time expression of his interest, disclosed by his search query. This is a key feature that distinguishes search advertising from all other forms of online advertising. Furthermore, Microsoft believes that search advertising is purchased by advertisers for different purposes than non-search advertising and its pricing is not significantly constrained by the pricing of other online advertising formats.

65. From the advertisers' perspective, search advertising offers a unique value proposition by ensuring a direct targeting to respond to the search user's immediate expressed intent, whereas, for instance, display advertising offers an inferior targeting which is not always based on the real-time needs of users.

66. Display advertising is, according to Microsoft, geared towards creating demand and building brand awareness rather than addressing the immediate and real-time requests of users.

67. Furthermore, the format of search ads is unique. Search ads typically consist of written text, while display (and contextual) advertising frequently include a variety of textual, graphical and video media formats.

68. In Microsoft's view, search advertising also differs from non-search advertising with respect to pricing. Search advertisers typically pay only when there is a click on their ads. This distinguishes search advertising from online display advertising, which are generally sold on a CPM (Cost-per-(thousand) impression) basis. The cost-per-click approach allows a much closer tracking between the cost of advertising and its benefits.

69. In addition, search advertising is sold through an auction system which is not the case for other types of online advertising and it is purchased separately by advertisers.

70. Incidentally, Microsoft claims that contextual advertising is to be distinguished from search advertising. As neither Microsoft nor Yahoo currently provides contextual advertising in the EEA, this question can be left open.

Results of the market investigation

71. The results of the market investigation on the product market definition are mixed. While a majority of respondents agree with the statement that all types of online advertising compete because advertisers mix and match different ad formats and targeting technologies (14), a significant number of respondents believe that search advertising is a market in its own right that cannot be grouped together with any other online advertising activity (15).

72. A large number of those respondents that view search as a separate market have mentioned that search advertising ensures direct targeting in response to the search user's expressed intent whereas other methods, such as display ads, aim primarily at increasing or creating brand awareness.

73. However, a majority of respondents does not specifically confirm the other arguments submitted by Microsoft (differences in format, pricing mechanism, size of clients or mode of sale) as reasons to make a distinction between search and other types of online advertising.

74. A number of responses have highlighted that in recent years online advertising has been characterised by a degree of convergence between formats. Several respondents argue that the "classical" distinction between search and non-search on the basis of the distinction between direct response marketing with search ads and the building of brand awareness with display ads is increasingly blurred in practice. Cited reasons include the growth of behavioural targeting and the fact that all the main players (Google, Microsoft and Yahoo) offer all types of advertising.

75. In any event, the exact product market definition as regards the possible distinction between search and non-search advertising can be left open since under any alternative market definition the proposed transaction would not raise serious doubts in the EEA.

b. Mobile search advertising

76. In the EEA, both Microsoft and Yahoo offer limited mobile search advertising services. The Agreements contemplate the future possibility for Yahoo to source its mobile search advertising, mobile search, and/or mobile mapping services from Microsoft on a nonexclusive or exclusive basis.

77. Similar to PC-based search advertising, mobile search advertising is designed to generate leads for mobile advertisers using sponsored search links that appear within mobile search results.

78. Mobile search ads are typically short, text based ads on which mobile searchers can click to get to the advertiser's landing page or to call the advertiser directly from their mobile phone. Similar to PC-based online search, mobile searchers can also get directions, traffic, weather, movie and other information by entering their search query into the search box on the relevant mobile search service from their mobile phone or personal digital assistant (16). Such mobile search services are accessible using a mobile browser to navigate to the mobile web-based version of the mobile search service as well as through mobile search applications, which are small software applications that reside on the mobile device itself and enable the mobile searcher to access the mobile search service directly.

79. A number of respondents to the market investigation have identified mobile advertising as one of the most important recent developments in online advertising (17). Also, some respondents have noted that mobile advertising presents some distinguishing features both technically (for example size of the ads) and commercially (for example it is especially appropriate to advertise outlets near to the actual location of the holder of the smartphone).

80. However the market investigation has also revealed that a majority of the respondents do not consider that mobile search advertising is a market in its own right (18).

81. In any event, the exact market definition can be left open because under any alternative market definition the transaction would not raise serious doubts in the EEA.

2. Intermediation in online advertising

82. In Google/DoubleClick the Commission distinguished between "direct sales" and "intermediated sales" by publishers of web space to advertisers and concluded that a "separate market for intermediation in online advertising can be defined" (19). Online advertising platforms act as intermediaries between advertisers and publishers when pooling advertisers' demand for web space and acquiring web space through syndication agreements with publishers. In Google/DoubleClick the Commission left open whether the market for intermediation in online advertising could be further sub-divided according to whether the intermediation concerns search ads or non-search ads. (20)

83. Both Microsoft and Yahoo are active as intermediaries in online advertising. The proposed transaction would result in Microsoft becoming the joint technical platform for both Microsoft's and Yahoo's intermediation businesses. It has been investigated whether this will have an impact on Yahoo's ability to continue competing as an independent intermediary (see sections VI.E.3. and VI.E.4). The exact product market definition as regards intermediation can be left open because the proposed transaction would not raise serious doubts in the EEA under any alternative market definition.

3. Distribution agreements on entry points

84. The question could be raised whether there is a separate market for distribution agreements on entry points to search engines. As the proposed transaction does not raise serious doubts in the EEA under any alternative market definition, this question can be left open.

4. Internet search

85. A question has been raised during the investigation as to whether a separate market for internet search, i.e. the provision of search results to internet users, could be defined.

86. However, the Commission's first phase market investigation has not assessed the question whether internet search constitutes a separate market. Moreover, even if such a separate market could be defined, the market would still have to be investigated because forms of user searches that do not require a search engine (for example online directories or even offline search formats) may also belong to the same product market. In any case, the transaction would not give raise serious doubts in the EEA under any alternative market definition. It is therefore appropriate to leave open the question whether internet search constitutes a separate market.

Conclusion on product market definition

87. In line with the merger decision Google/DoubleClick, the Commission has found that online advertising constitutes a relevant market. Whether segments of that market constitute relevant markets in their own right can be left open because the transaction would not give rise to serious doubts as regards its compatibility with the common market in the EEA under any such narrower market definition. For the same reason, it can be left open whether intermediation, internet search and distribution agreements on entry points to search engines also constitute relevant markets.

C. GEOGRAPHIC MARKET DEFINITION

1. Online advertising and its segments

88. The notifying party submits that the relevant geographic market(s) for online advertising may be defined alongside linguistic borders within the EEA. In most instances, this corresponds to markets national in scope.

89. The results of the market investigation in Google/DoubleClick demonstrated that, both for advertisers and for publishers, there are many factors which, from a commercial point of view, point to a distinction of national or linguistic sub-markets (21). The supply or the purchasing of advertising space is differentiated on the basis of national preferences, languages, and cultural specificities. In general, support and sales networks are locally present in different countries.

90. In Google/DoubleClick, it was therefore concluded that the markets for online advertising space are divided alongside national or linguistic borders within the EEA. The same conclusion would apply to the alternative hypothetical narrower markets for search and non-search advertising.

91. The market investigation in the present case has confirmed this conclusion. Virtually all respondents to the market investigation agree that the provision of search advertising space may be defined along linguistic lines within the EEA in terms of geographic scope. In most instances, this will correspond to a market national in scope. The same would also apply to non-search advertising (22).

92. Also, a large number of advertisers and media agencies have replied that a majority of their advertising campaigns (measured in budget) is conducted on a national basis (23).

93. Therefore, the scope of the relevant geographic market for online advertising and its segments is national or linguistic.

2. Intermediation in online advertising

94. On the one hand, syndication agreements normally cover all web sites on which the publisher wants to monetize inventory, regardless of their target audience. Many intermediation providers are active on an international basis. On the other hand, the syndicated content will likely be in the language of the publisher's content offering which, in view of the online advertising market being national in scope, might point to a more narrow market definition. However, in the present case, the exact geographic market definition may be left open because the transaction would not raise serious doubts in the EEA under any alternative market definition.

3. Distribution agreements on entry points

95. There are distribution agreements on entry points that are worldwide, regional (EEA-wide) or national in scope. Section VI.B.4. below discusses in more detail the effects of the transaction in relation to distributors. On this basis, the exact geographic market definition may be left open because under any alternative market definition the transaction would not raise serious doubts in the EEA.

4. Internet search

96. The geographic scope of a potential market for internet search could potentially be wider than national or linguistic markets because the largest search engines like Google, Yahoo or Microsoft operate on a global basis. Those search engines strive to index the whole internet and are accessible from anywhere in the world. Also, fluency in English (the language of most of the websites) is increasingly extended across the world.

97. Nevertheless, seen from the demand side, many users require access to a search engine and to search results in their own language.

98. In any event, the exact market definition can be left open because even under the narrowest possible definition of the relevant geographic market (national basis), the transaction would not raise serious doubts in the EEA.

VI. COMPETITIVE ASSESSMENT

A. MARKETS CHARACTERISTICS

Overview of online advertising

99. Sales of online advertising in Europe were approximately EUR 13 000 million in 2008 (24). Online advertising related to search queries were the primary source of revenues for advertising platforms which in 2008 are estimated at approximately EUR 5 500 million in Europe. Display advertising was approximately EUR 3 800 million while classifieds accounted for EUR 3 300 million. Email advertising accounted for EUR 200 million. Furthermore, search advertising is increasingly becoming more popular, relative to non-search advertising, enjoy a growth rate of around 25% (higher than other formats of online advertising) in Europe.

Two-sided platform

100. As discussed above, a search engine is a matchmaker connecting advertisers and users. Therefore a search engine is a two-sided platform where the demands on the two sides are interdependent. Advertisers aim to reach a large audience and monetise their investment in advertising. Users value the relevance of the internet search which includes the organic (or algorithmic) and advertising (or sponsored) results. The Commission therefore will examine the position of search advertising platforms on their ability to generate search traffic as well as their ability to sell that search traffic to advertisers.

User side

101. Search engines attract users by offering searches free-of-charge. Competition for the users mainly takes place on the basis of the quality of the search results (i.e. their relevance to the users but also speed of returning results) and the user interface. The relevance of search results is not only based on the results of the organic search but also on the ads that will be displayed on the result page.

102. Microsoft claims that users tend to "single-home", i.e. conduct searches using one single search engine. It submits that less than [...]% of users multi-home exclusively between Microsoft and Yahoo in France and the UK. Microsoft has submitted a study (25) based on comScore data for September 2009. By defining a single-homer as a user that performs more than 90% of his queries on a single platform within a month, approximately [...]% of users in the UK and [...]% in France single-home on Google. Less than [...]% and [...]% single-home on Yahoo in France and in the UK respectively. Approximately [...]% of users single-home on Microsoft in both countries. Similar results are recorded in Germany, Italy and Spain. Therefore, these figures tend to confirm that the great majority of users single-home (as defined in the study).

103. The notifying party has also provided data for September 2009 that evaluate the reach of different search engines as computed by comScore. The reach evaluates the probability that users will visit a particular search engine within a month. According to these reach data only [...]% of search users in Europe did not perform a search on Google in September 2009. Moreover, [...]% of search users did not perform searches on Microsoft and [...]% did not use Yahoo. These figures are consistent across Europe whereas worldwide and in the US Google has a relatively lower reach of users. Finally, Yahoo has a much smaller reach in Europe than in the US ([...]% versus [...]%). These figures illustrate the importance of Google in Europe and that a relatively small share of search users considers Microsoft and Yahoo to run their searches.

Region Google Microsoft Yahoo Ask Yandex AOL Baidu

Europe [90-100]% [20-30]% [10-20]% [10-20]% [5-10]% [5-10]%

USA [80-90]% [30-40]% [50-60]% [30-40]% [30-40]%

Worldwide [80-90]% [20-30]% [30-40]% [10-20]% [0-5]% [10-20]% [10-20]%

104. Furthermore, Microsoft has provided data based on comScore on the queries share by entry points. On the basis of this analysis, Microsoft claims that [...]. This is illustrated by the fact that most search queries are not entered at the homepage bing.com. Instead, they are often made at Microsoft's portal msn.com or for example through an internet browser's search box or tool bar.

105. Microsoft believes that the demographics of its own, Yahoo's and Google's users are broadly similar. The Commission could not confirm that the profile of users among the main search platforms is similar. Microsoft and Yahoo however have provided data on total searches split by type of content in the UK and France. The analysis of these data shows that Yahoo's portfolio has a higher share of keywords related to [...]. Microsoft's portfolio however has a relatively higher share of types of miscellaneous keywords not directly linked with any predefined categories ([...]% of keywords that are classified as "other").

Advertisers' side

106. Competition for advertisers is principally based on the return on investment (ROI) that a search platform can deliver. One determinant of ROI is the "conversions" into sales that ad clicks generate (conversion rate). Another key measure is the "Click Through Rate" (CTR) which measures the likelihood that users will click on ads. As explained above, for online search, advertisers pay per click (CPC), i.e. each time a user clicks on their ad. The price that an advertiser pays each time a user clicks on its ads is determined by a real-time auction organised by the ad platform. Search advertising offers several benefits to advertisers such as the potential to target advertisements to a group of users that is likely to have an interest in purchasing their particular products. Also advertisers are better able to monitor the ad effectiveness and their ROI.

107. Microsoft and Yahoo have submitted data on multi-homing of advertisers between the Microsoft and Yahoo platforms in the UK and France. Multi-homing is defined as the extent to which advertisers are currently advertising on more than one platform. For advertisers that amounts to 90% of the turnover generated in each platform, approximately more than [...] of them advertise on both platforms.

108. Microsoft and Yahoo have also provided data on advertisers' revenue for the UK and France as well as the distribution of their advertiser base according to the size of spending and the advertising content for the years 2008 and 2009 (until October). In the UK, the distribution of Yahoo's and Microsoft's advertisers is [...]. Regarding the split of advertisers by type of content (for example travel, financial services, etc.), the analysis of the last 11 months (from November 2008 to September 2009) shows that, both in France and the UK, Yahoo and Microsoft [...].

Innovation

109. Another important dimension of competition between search engines is their ability to innovate. Search engines not only try to develop new services (leap-frog innovation) but also constantly strive to innovate incrementally on existent services in order to be able to deliver better services to both advertisers and users. Search engines spend enormous efforts in order to continuously improve the search algorithm and their ability to match users and advertisers.

110. The high degree of innovation in these markets is proven by the rate at which new innovation has been introduced in the past (for example introduction of the auction mechanism, quality rank, analytics, spell check, etc.) and by constant experimentations that search engines undertake daily on several fronts.

Entry barriers

111. Barriers to entry into this market appear to be high and include, among others, hardware, cost of indexing the web, human capital, cost of developing and updating the algorithm and IP patents. Microsoft estimates that the capital expenditure required to enter the market is approximately USD 1 000 million in hardware and USD 1 000 million in human capital. On top of that, Microsoft estimates that a new entrant would have to spend several billions of dollars to develop and update the algorithm. Finally, Microsoft explains that there are very significant costs that a new entrant would have to bear related to the necessity to have a large database.

B. MARKETS SHARES

User side

112. The internet search markets in the EEA countries are very concentrated. Google is by far the platform that enjoys the highest market shares. In 2009, Google has market shares above [90-100]% in most national markets whereas Microsoft and Yahoo both have market shares below [5-10]%. The only exception in the EEA where Google is not the number one search platform is in the Czech Republic where Seznam is the market leader (26). Also, there are some smaller platforms, namely Ask and AOL, that are active in several Member States and also worldwide.

113. Microsoft and Yahoo run their Bing and Yahoo search engines in several EEA countries. Both search engines are active in Austria, Denmark, Finland, France, Germany, Ireland, Italy, Netherlands, Norway, Spain, Sweden and the UK.

114. Paid search data from comScore for September 2009 confirm that the Microsoft and Yahoo have low market shares in all Member States and that the transaction does not lead to any national affected market. Paid search refers to the subset of searches initiated by internet search users that trigger a sponsored link (27). The highest market shares of Microsoft and Yahoo in terms of paid search queries are in the UK and France (28).

Country/Region Google Microsoft Yahoo Ask Yandex AOL Baidu

Austria [90-100]% [0-5]% [0-5]%

Belgium [90-100]% [0-5]% [0-5]%

Germany [90-100]% [0-5]% [0-5]% [0-5]%

Denmark [90-100]% [0-5]% [0-5]%

Finland [90-100]% [0-5]% [0-5]%

Spain [90-100]% [0-5]% [0-5]%

France [90-100]% [0-5]% [0-5]%

Ireland (29) [90-100]% [0-5]% [0-5]% [0-5]% [0-5]% [0-5]%

Italy [90-100]% [0-5]% [0-5]% [0-5]%

Netherlands [90-100]% [0-5]% [0-5]%

Sweden [90-100]% [0-5]% [0-5]%

UK [80-90]% [0-5]% [0-5]% [0-5]% [0-5]%

Europe [80-90]% [0-5]% [0-5]% [0-5]% [0-5]% [0-5]% [0-5]%

USA [60-70]% [5-10]% [10-20]% [0-5]% [0-5]% [0-5]%

World [70-80]% [0-5]% [5-10]% [0-5]% [0-5]% [0-5]% [5-10]%

115. In response to a request from the Commission, Microsoft submitted data on market shares on search queries by Member States. The market shares of Microsoft and Yahoo are limited and the highest overlap arises in the UK were Microsoft and Yahoo have each a [0-5]% market share.

116. Worldwide algorithmic search markets are also rather concentrated with Google accounting for more than [70-80]% of paid search queries. Apart from the players already mentioned it is interesting to note the importance of some national players such as Baidu in China, Yandex in Russia and Naver in South Korea which are the leading search platforms in their respective countries. The Chinese search engine Baidu has a significant market presence worldwide with higher market share than Microsoft, and lagging only slightly behind Yahoo.

117. Also, the degree of concentration in the market is higher in the EEA than in the US where both Microsoft and Yahoo have higher market shares in terms of paid search queries (around [5-10]% and [10-20]% respectively). Similarly, Microsoft's and Yahoo's position is stronger worldwide than in Europe. It appears that Europe is a Google stronghold since Google has lower market shares both in the US and worldwide than in Europe.

118. Therefore, Microsoft and Yahoo have small market shares in the EEA as a whole and in each Member State where their activities overlap. Also, Microsoft's and Yahoo's presence and market shares in the US and worldwide, which might be more indicative of the incentive of Microsoft and Yahoo to innovate, are limited.

119. On this basis, the transaction does not lead to affected national or EEA markets for internet search when market shares are defined as the shares on the volume of search queries and paid search queries. However, given that competition in this market takes place in terms of quality and innovation, and the entry barriers are high, the Commission undertook a thorough analysis of the possible effects of the transaction on the search market.

Advertising side

120. A similar picture emerges when one considers the market for search advertising in terms of market shares. Microsoft and Yahoo account for a small share of search advertising in France and in the UK, the two countries where they both have their own independent advertising platform. Microsoft and Yahoo also currently have a vertical relationship with each other in nine Member States. In Austria, Denmark, Finland, Germany, Italy, the Netherlands, Norway, Spain and Sweden Microsoft's search ads are powered by Panama.

121. In France, the combined market share of Microsoft and Yahoo was approximately [10-20]% in the first semester of 2009. Yahoo's market share in France has fallen sharply from almost [20-30]% in 2006 to less than [5-10]% in the first semester of 2009. Microsoft's position is also marginal and has risen from [0-5]% in 2006 to [0-5]% in the first semester of 2009. Google has been gaining grounds from slightly less than [80- 90]% to approximately [90-100]% in the first semester of 2009.

122. In the UK, the situation is rather similar with Microsoft and Yahoo holding combined market share of [5-10]% in the first semester of 2009, [5-10]% for Yahoo and [0-5]% for Microsoft. While Yahoo's position, as in France, has been falling over the last years, Microsoft's shares have also been falling from [0-5]% in 2006 to their current level.

123. Microsoft and Yahoo have also provided data on their revenues on a daily basis, which indicate that Yahoo's position in terms of revenue has weakened in recent years.

124. Therefore, the transaction does not lead to affected markets for search advertising. However, given the complexity of the market under consideration, the high entry barriers and the reduction of the number of players from 3 to 2, the Commission has undertaken a market investigation and has thoroughly analysed the effects of the transaction.

Other segments within online advertising

125. The transaction does not lead to affected markets also on wider markets encompassing i) search and online advertising, ii) search and contextual advertising or iii) search and mobile advertising.

126. In a market for online advertising, Microsoft's and Yahoo's activities do not lead to any substantial overlap. Furthermore, as explained in the section on product market definition, display advertising, which accounts for a large part of total online advertising turnover in the EEA, is not included in the agreement between Microsoft and Yahoo. Also, Microsoft and Yahoo are not active in contextual advertising in the EEA and only to a very limited extent in mobile advertising. Microsoft has also submitted market shares for an overall market for online advertising that confirm that the transaction does not lead to any affected market in the EEA.

127. Concerning contextual advertising, Microsoft and Yahoo submit that Microsoft is not active in that market while Yahoo has halted its activities. Therefore, considering a wider product market encompassing search and contextual search would only dilute the Microsoft and Yahoo's market shares in search advertising.

128. In a market encompassing search and mobile advertising, Microsoft and Yahoo do not have higher market shares. In the EEA they offer limited mobile search advertising services. Yahoo only offers mobile search advertising in the UK, where it generated revenues of GBP [...] million in 2008. Similarly, Microsoft only offers such services in the EEA in France where it generated revenues of EUR [...] million in 2008. Compared to the size of the search advertising markets in the UK and France these activities represent a small portion and therefore Microsoft's and Yahoo's market shares would not be increased as a result of the proposed transaction.

Publishers' and distributors' side

129. Microsoft currently has no syndication agreements with publishers anywhere in the EEA and, according to Microsoft and Yahoo, Yahoo's market share in the EEA as regards syndication is comparable to its overall market share for search advertising. The transaction therefore does not lead to any affected market in the EEA with regard to intermediation in online advertising. Market shares with respect to distribution agreements on entry points are difficult to measure. The estimates produced in the course of the investigation show that the transaction does not lead to any affected market in the EEA in this respect.

Conclusion

130. Market share data indicate that the combined market shares of Microsoft and Yahoo are well below the safe harbour of the horizontal merger guidelines (30). The transaction does not lead to any affected market and would therefore normally be suitable for treatment under the simplified procedure. Furthermore, in the current situation where the undertakings mainly compete for incremental innovation, combined low market share is likely to be indicative of the competitive position of Microsoft and Yahoo. Therefore, prima facie, it is unlikely that the proposed transaction between two players with such small market shares could cause serious doubts as to the transaction's compatibility with the common market.

C. COUNTERFACTUAL

131. In line with the Guidelines on the assessment of horizontal mergers, the Commission compares the competitive conditions that would result from the notified merger with the conditions that would have prevailed without the merger (31). In most cases the competitive conditions existing at the time of the merger constitute the relevant point of comparison for evaluating the effects of a merger. The Commission will therefore look at the current situation of the parties and the market in order to evaluate what would be the evolution absent the transaction.

132. It is important to note that as regards search advertising, Microsoft and Yahoo currently only compete in France and in the UK within the EEA and both currently have very small market shares.

133. Microsoft states that while there are plans to launch the search engine Bing in other European countries, it had no plans to launch adCenter in any more countries in Europe absent the merger. This is confirmed by Microsoft's internal documents provided in the course of the Commission's investigation.

134. Therefore, given that Microsoft does not plan to introduce adCenter in other EEA Member States within a foreseeable future, the elimination of potential competition in the rest of the EEA appears to be limited.

135. The notifying party submits that absent the transaction, neither Microsoft nor Yahoo is likely to become an effective competitor to the market leader, Google, neither in internet search nor in search advertising.

136. According to the notifying party, the weakness of Microsoft and Yahoo in the search advertising space - based on every relevant indicator - finds its main explanation in the lack of "critical mass" scale.

Yahoo's declining performance

137. According to the notifying party, Yahoo is not, and would be unlikely to become, an effective competitor in search advertising in the EEA.

138. In fact, Yahoo's competitiveness has decreased significantly and progressively in recent years. For instance, in overall market conditions characterized by significant growth, Yahoo's market shares in the UK and France (by volume as well as value) have fallen to less than half of what they were in 2006. In both countries, Yahoo has been losing market shares over the last three years, as described in the market shares section above.

139. The notifying party submits that search advertising turnover in the 19 European countries monitored by IAB/PwC (32) increased by 69% between 2006 and 2008. While Google increased its turnover by 84%, Yahoo's turnover decreased by 36%. The pattern was similar in the UK and France.

140. According to Yahoo, reversing the trend of decreasing revenues would require investments of a magnitude [...]. Yahoo claims that heavy investments in both software and hardware are vital in order to maintain a competitive position on the market. Data centres need to be upgraded, storage for web search index maintained, servers to execute advanced algorithms put in place, and content sources need to be improved to increase the amount of data crawled. Yahoo considers that it would have to invest hundreds of millions of dollars in operating expenditures and capital expenditure over each of the next two years, and make ongoing investments of similar amounts thereafter on an annual basis. In this respect, Yahoo notes that its search engine investments have only been a fraction of those made by Google; [...].

141. Microsoft [...] submit that Yahoo has been a weak innovative force in the market of organic search in the last years and [...].

142. Moreover, the decline in revenues has necessitated cost-saving measures. For instance, Yahoo announced a major cost-reduction initiative in 2008, targeting a reduction in its global workforce of over 10%. (33)

143. [...]. Therefore, according to the parties, absent the transaction, it is unlikely that Yahoo would develop into a credible rival to Google. According to the parties, this means that the removal of Yahoo as an independent competitor in search advertising will not give rise to any significant lessening of effective competition.

144. However, while the Commission acknowledges that it may be challenging for Yahoo to reverse the trend and become a strong competitor in search advertising, as well as in internet search, it must be noted that Yahoo's results in the last quarter of 2009 show that its revenues grew by 4% on a sequential basis. (This number however does not represent an important increase in absolute numbers.) The primary drivers of this performance were "improvements to our matching algorithms that yielded better than expected positive results including an 8% sequential increase in global revenue per search" (34). At the same time, internet search is an important part of Yahoo's business given the complimentary nature of internet search to other business activities of Yahoo. Therefore, the Commission considers that Yahoo has, absent the transaction, the incentives to maintain a search platform that competes with the main players on the market.

145. Moreover, with regard to the ability and incentive of Yahoo to innovate, it should be noted that the small market shares of Yahoo in Europe do not necessarily indicate a lack of ability and incentive for Yahoo to innovate because Yahoo's innovations in the US spill over in other parts of the world. However, as described in the market shares section above, although Yahoo's presence is higher in the US and worldwide than in Europe, it remains limited and falling.

146. The market investigation largely confirms that Yahoo's performance has declined over the last three years and continues to decline. While Google sees Yahoo as an important innovator, all other competitors responding to the market investigation consider that Yahoo's performance has decreased over the last three years (35). Also, advertisers and media agencies replying to the market investigation do not consider Yahoo to be an important innovator in search advertising (36) and they believe that Yahoo has weakened as a competitor in search advertising (37). Finally, they have also perceived a slowing down in Yahoo's innovation capability over the last three years (38).

147. Therefore, based on Microsoft's and Yahoo's submissions and the first phase market investigation, there is evidence that Yahoo's performance has declined over the last years. Furthermore, there is no evidence that Yahoo is likely to stop facing difficulties in competing effectively in internet search and search advertising.

Microsoft is lacking scale

148. According to the notifying party, although it has pursued the goal of becoming a credible alternative provider to Google, it would not be able to do so in the absence of the transaction.

149. Microsoft's projected market share between 2011 and 2013 absent the transaction is around [...]% in the US while it would remain stable at [...]% outside the US. However, recent US figures on search query market shares show that in January 2010 Bing had a market share greater than 11% (39).

150. In recent years, Microsoft has made very significant investments in its internet search and search advertising business, as shown in the graphs below.

R&D and CAPEX Contribution margin

[TABLE REDACTED]

151. However, in spite of these efforts, the scale and market share of its business remain modest and, financially, it is loss-making. Microsoft argues that organic growth has not been sufficient and the economics of the industry make a strategy based on organic growth inherently costly [...].

152. According to the notifying party, it is unable to compete effectively without sufficient scale. The merger would provide Microsoft with the scale necessary to become a credible competitor. This is because its current small scale hampers its ability to innovate and compete. The argument on scale will be further developed under the section relating to possible beneficial effects of the merger.

153. The first phase market investigation tends to support the view of the notifying party that scale is an important element to be an effective competitor. (40) Almost all advertisers responding to the market investigation consider that Microsoft does not have enough traffic volume to be an attractive alternative to Google (41). This will be further developed in the section below relating to possible beneficial effects of the merger.

154. Prima facie, the first phase market investigation gives the Commission reasons to believe that scale is one of the reasons why Microsoft is not able to compete effectively.

Google has a strong position on the market

155. According to the notifying party, Google is the entrenched market leader in search advertising globally and in the EEA. In each EEA Member State, the size of Google's search advertising business (whether measured by volume, value, in market shares or in absolute terms) will remain at least ten times the size of Microsoft's search business when combined with Yahoo's.

156. The notifying party claims that Yahoo's and Microsoft's smaller audiences and market share in search advertising in Europe, and lower advertiser ROI, result in advertisers spending substantially more on Google. Therefore, neither Microsoft nor Yahoo provides a 'must have' search advertising service.

157. All the advertisers responding to the market investigation consider that Google is a 'must have' for search advertising campaigns. Also, all of the advertisers responding to the market investigation highlight the fact that Google's query volume is one of the main reasons why Google is a 'must have' (42). In addition, the market shares clearly demonstrate that Google has a much greater user reach and a much larger number of single homers that never use another search engine.

158. As discussed above, Google has a very strong position in the market and has been growing in Europe the last three years, taking market shares from Yahoo in particular. (43) Also, Google is stronger in Europe than anywhere else in the world.

Conclusion on counterfactual

159. It can be concluded from the notifying party's submission and the market investigation that, on the one hand, Google is currently a very strong player on the market and has continued to grow in the last couple of years. On the other hand, Yahoo appears to be a weak player unable to make the necessary investments in order to successfully compete and while Microsoft has invested heavily, it appears that it has not succeeded in getting sufficient scale in order to compete effectively against Google.

D. POSSIBLE BENEFICIAL EFFECTS OF THE MERGER

The notifying party's view

160. According to the notifying party, its strategic rationale in pursuing the transaction lies primarily in the notion that success in search advertising (as well as in internet search) is dependent on scale. By acquiring Yahoo, the increased scale will enable Microsoft to become a more credible alternative to Google and thus provide greater value to advertisers.

161. According to Microsoft, increased scale will have a positive effect both on users and on advertisers.

162. First, the notifying party claims that increased scale will impact on the user experience, because the trailing volume of search queries holds each party back from improving their search results. Microsoft argues that it is much easier for a large engine to spot issues related to relevance when it possesses a larger sample of queries, especially for less frequent queries ("tail" queries). With larger scale, a search engine can run tests on how to improve the algorithm and ultimately implement corrections. The notifying party submits that on average, between [...]% and [...]% of Microsoft's search traffic is currently used for experimentation while Yahoo estimates that around [...]% of its search traffic is used for experimentation. With the higher "bandwidth" that comes with increased traffic volume, experimental traffic takes up a smaller proportion of overall traffic, and it is possible to experiment more, and faster.

163. Secondly, the notifying party submits that increased scale also impacts on the advertiser experience, since a higher degree of user engagement and a better ability of the platform to show relevant ads impacts on advertiser ROI. Higher query volume in turn generates ad inventory. A larger inventory translates into more opportunities for advertisers to reach and target their intended audience. More advertisers lead to more ads for the search engine to choose from for any given query. This, in turn, improves ad relevance and the likelihood that a user will click on an ad and ultimately convert his click into a purchase. As a result, the inventory becomes more valuable to advertisers who see their return on investment increase.

164. Finally, the notifying party also suggests that the transaction will allow advertisers to manage large scale search advertising campaigns on a larger combined platform and thus reduce their fixed costs of managing multiple ad campaigns.

Studies provided by Microsoft and Yahoo on the importance of scale

165. Following the Commission's request, Microsoft and Yahoo have submitted studies examining the importance of scale with respect to the ROI for advertisers and benchmarking studies on the relevance of the main search engine for users.

166. First, concerning the relevance of search results for users, the notifying party has not submitted evidence that illustrates that scale leads to higher relevance; however, Microsoft has submitted a benchmarking study that compares its algorithmic search result relevance against that of Google and Yahoo (44). [...]. The results of this study illustrate that Microsoft and Yahoo [...]. However, for the most frequent queries, the overall relevance gap between engines is very small [...].

167. Yahoo also tracks the relevance of its search results with respect to other search platforms. The internal benchmark studies that Yahoo has submitted show [...] (45). Yahoo also measures the impact of its search ads on the relevance of the whole page. The results indicate [...] (46).

168. Therefore while the Commission notes that Google appears to perform better in terms of relevance especially for [...] queries, this does not provide evidence that scale leads to higher relevance for users, since the above studies do not take into account the technology of the different search engine which are not related to scale.

169. Second, with respect to the relevance of scale for advertisers' ROI, Microsoft has submitted empirical studies that have considered the impact of scale on the platforms' revenue per search (RPS) for a given keyword (47). These studies illustrate that RPS increases with the volume of queries. However, the studies also show that RPS increases at a decreasing rate and that highly searched keywords already achieve high revenue per search in the parties' platforms. The parties estimate that the transaction will increase RPS on adCenter by [...]% in both UK and France.

170. Also according to the notifying party, Microsoft and Yahoo have lower RPS than Google. Using US data, the parties show that the RPS in Microsoft's platform is less than [...]% of the RPS in Google.

171. These studies suggest that scale is important for the platforms to generate revenue from the queries run on their platforms. However, the studies show that the benefits increase at a decreasing return. Also, the studies do not demonstrate that advertisers will necessarily observe a higher ROI or that users will benefit from the higher scale.

172. The notifying party has also submitted a study on the advertiser feedback on adCenter ROI (48). [...]. However, the Commission does not consider that this survey alone can provide conclusive evidence that higher scale leads to higher ROI. [...].

Market investigation results

173. Competitors that responded to the market investigation indicate almost unanimously that scale is important in order to be an effective competitor in search advertising (49).

174. However, Google argues that while scale is an important and necessary ingredient of having a successful search engine, its degree of importance has been largely overstated. In particular, Google underlines that the value of incremental data decreases as the amount of data increases, something which is acknowledged by the notifying party.

175. As regards the cost of managing ad campaigns on several platforms, no elements in the market investigation contradict Microsoft's and Yahoo's view that the merger will reduce costs in this regard.

176. Overall, respondents to the first phase market investigation confirm that the merger will have pro-competitive effects and will lead to more advantages than disadvantages. The large majority of the advertisers and media agencies believe that post-merger, the new entity will be a stronger competitor to Google (50).

E. EFFECTS ON THE VARIOUS PLAYERS IN THE SEARCH ADVERTISING INDUSTRY

177. In this section the potential effects of the proposed transaction on the various market players will be assessed. This section will examine in turn the potential effects of the transaction on advertisers, users, publishers and distributors.

1. Effects on advertisers

178. As explained above, search ad platforms use keyword bidding systems. Advertisers bid on search query terms known as keywords. The auction concerns a set of several ad slots at the same time. Thus the outcome of the auction is a ranking of several ads (also known as a "position auction"). One particular feature of the bidding is the use of a second price auction with a reserve price (i.e. minimum bids below which participation in the auction is excluded) (51).

179. The price that the advertiser has to pay to the platform is charged each time an internet user clicks on the ad (cost per click or CPC). The bid itself is not the only factor that determines the slot that an ad is placed in or the price that the advertisers have to pay. The revenue of the platform is contingent on the probability of an ad being clicked and the price the advertiser pays. Advertising platforms use a "quality" score to rank the ads that reflects the expected CTR (click through rate). The keyword bidding process and the quality score algorithm determine both the CPC advertisers pay and the slots they receive.

Theory of harm

180. The proposed transaction's main effect would be the creation of a single search advertising platform where advertisers who want to have their ads placed on Microsoft's and on Yahoo's owned and operated (O&O) sites and any of their syndication partners' sites will have to place their bids. As a consequence, advertisers which previously participated in auctions on Yahoo's platform, Panama, will be migrated to Microsoft's adCenter platform and Panama will no longer exist.

181. Post-transaction advertisers will no longer be able to switch from adCenter (nor from AdWords) to Panama and thus Panama will no longer represent a competitive constraint on either Microsoft or Google. Even if the price mechanism takes place through an auction, Microsoft may be able to influence the prices (and therefore to exercise market power). This is because the ad platform completely controls the auction mechanism and has an obvious interest in making sure that this mechanism's characteristics are profit maximising for the platform. Theoretically, absent the competitive constraint exercised by Panama, the combined platform may influence the final price by altering either the quality score or the reserve price.

182. Moreover, the pooling of advertisers that results from the migration of advertisers from Panama to adCenter (and/or to adWords if advertisers eventually choose to advertise on Google instead) may in theory lead to an increase of the average CPC for each query compared to the CPCs for these queries pre-transaction. Following the elimination of one advertising platform some advertisers will shift their bids to the new platform which may increase the bidder density for a particular keyword.

The notifying party's arguments

183. The notifying party argues that in the EEA they compete to a very limited extent with each other given their low market shares and therefore the transaction will not lead to anticompetitive effects with regard to advertisers.

184. Furthermore, according to Microsoft, the transaction will generate efficiencies through greater scale of the combined platforms and post-transaction the combined entity will be in a position to challenge Google more effectively, offering advertisers a higher traffic volume and a better quality of matches. Therefore, Microsoft stresses that the proposed transaction is pro-competitive and will be beneficial for advertisers because what really matters to the advertisers is not the absolute CPC but ROI.

185. Moreover, Microsoft claims that the transaction will allow advertisers to manage large-scale search advertising campaigns on a larger combined platform which will reduce their fixed costs of managing ads on multiple platforms.

Results of the market investigation and Commission's analysis

(i) The possible price increase is likely to be limited due to the current market structure

186. Most advertisers replied to the market investigation that they consider it unlikely that the CPC will increase (52). Moreover, the market investigation indicated that advertisers are inclined to believe that, should such an increase occur, the effects will be limited. Indeed, they unanimously replied that they would not consider this a problem (53).

187. With regard to the theoretical elimination of the competitive constraint that Panama might exercise on Microsoft's market power and particularly on its ability to alter the auction mechanism, it should be noted that the parties' activities overlap only in two national markets, the UK and France, where both parties are currently offering their own independent advertising platforms. Moreover, in those countries the parties have very low market shares.

188. Finally, as described in more detail in paragraphs 137 to 147, and as confirmed by the market investigation, Yahoo's performance in the search market has declined in the last three years. The competitive constraint exercised by Panama on adCenter has therefore been decreasing in the past years.

189. Many advertisers have indicated in their replies to the market investigation that they consider Yahoo and Microsoft to be close competitors (54). Nevertheless, the vast majority of the respondents consider them to lag behind Google (55), the latter being the main competitive constraint that they face.

190. With regard to the possible price increase due specifically to the increase in bid density, the latter depends on the degree to which advertisers pre-transaction bid for the same ads/keywords-combinations on both Panama and adCenter. In theory, posttransaction, advertisers that are budget constrained (56) or that were not previously bidding on both platforms for the same ads/keywords-combinations, will reallocate their budget to adCenter for certain ads/keywords-combinations, thus increasing the bid density, and average CPC, for those ads/keywords-combinations.

191. However, the possible impact of this theoretical effect is limited because many advertisers are already running their campaigns on both adCenter and Panama. Indeed, the vast majority of the respondents of the Advertisers and Media agencies questionnaires indicated that they multi-home on all three platforms (57).

192. Finally, it must be stressed that this is a static analysis, that is to say this effect is based on the assumption that nothing in adCenter changes apart from the number of bidders and their budgets. However, it can be expected that the mere fact that adCenter and Bing will have to deal with much more traffic, and thus will have much more data available for experimentation, will tend to increase the quality of their product, i.e. it might lead to better ad matching and thus higher conversion rates. Furthermore, the new platform might be perceived as an effective competitor to Google, thereby dynamically also decreasing the bidder density, and the CPC, in adWords.

(ii) A higher return on investment seems to matter more than a possible increase in price

193. The market investigation has also indicated that although the CPC is an important consideration for an advertiser when choosing the platform on which to place its ads - and when deciding on how to allocate its budget between different platform - advertisers' decisions are primarily driven by their ROI. ROI in addition to CPC, depends also on the CTR, and the conversion rate (the share of clicks that lead to a purchase of the advertised good).

194. The market investigation showed that several advertisers consider that on average Google's CPC for the same keywords are higher than Yahoo's (58) and the majority consider Google's CPC to be higher than Microsoft's (59) even though almost all of them perceive the ROI offered by Google to be higher - or much higher - than Yahoo's (60) and Microsoft's (61).

195. Advertisers seem therefore to be willing to pay a higher CPC as long as their ROI is also higher. Higher ROI could result from a better matching of the ad to the query. The more relevant the ad is, the more likely it is that if the user clicks on the ad he will proceed to purchase the relevant product (higher conversion rates).

(iii) Advertisers expect to benefit from the transaction

196. The market investigation shows that the large majority of advertisers expect the proposed transaction to be positive for a number of reasons and none of the respondents has raised concerns.

197. As extensively mentioned in the previous paragraphs, advertisers have almost unanimously indicated in response to the first phase market investigation that scale is an important factor to effectively compete in search advertising (62). Moreover, the large majority of the respondents to the Advertisers and Media Agencies questionnaires believes that post-merger, the adCenter entity will be a stronger competitor to adWords that will offer them more advantages than disadvantages (63).

198. Also, almost all respondents to the Advertisers and Media Agencies questionnaires welcome the transaction that may lead to an effective competitor to Google (64). A more relevant and credible alternative advertising platform may act as a countervailing force against Google currently considered almost unanimously by advertisers as a "must have" (65).

(iv) The advertisers will benefit from the reduction of fixed costs

199. Finally, there are no elements in the market investigation that contradict Microsoft's claim on the positive effect on advertisers through the reduction of fixed costs for the monitoring of ads on different platforms.

Conclusion on the effects on advertisers

200. It follows from the above analysis that it is unlikely that the proposed transaction will increase the CPC at all or to a significant extent. The impact on advertisers will not be significant, if any. Moreover, it is likely that as a result of the transaction adCenter will be able to compete more effectively with Google. On this basis, it can be concluded that the transaction does not give rise to serious doubts as to its compatibility with the common market as far as its effects on advertisers are concerned.

Agencies questionnaire.

2. Effects on internet search users

201. Effective competition brings benefits to consumers, such as low prices, high quality products, a wide selection of goods and services, and innovation. Through its control of mergers, the Commission prevents mergers that would be likely to deprive customers of these benefits by significantly increasing the market power of firms. By 'increased market power' is meant the ability of one or more firms to profitably increase prices, reduce output, choice or quality of goods and services, diminish innovation, or otherwise influence parameters of competition. (66)

Theory of harm

202. The Commission has analysed plausible theories of harm that may arise following the transaction and that would primarily affect the users in the internet search market. The Commission has analysed whether the transaction will post-merger influence the incentive of the players to innovate, to lower the quality of organic search (i.e. degrading the relevance of the results) and whether users will be harmed by a loss of variety. This has to be contrasted against the situation absent the merger (the counterfactual explained in detail above) where Yahoo is currently not perceived as a strong innovator and Microsoft is lacking scale. The starting point for this assessment is the fact that the transaction will remove one provider of an algorithmic search engine from the market and, accordingly, a separate engineering team.

203. Concerning innovation, post-transaction Yahoo may lose the ability to provide important innovations since such innovations are dependent on having its own search index, engineering team and independent search business. Also, it must be noted that economic theory is ambiguous regarding the incentives to innovate depending on the level of concentration in a given market. The Commission therefore has to assess the facts of the case and to analyse the incentives and capabilities of the merging parties and its impact on other players in the market and contrast it with the situation pre-merger as explained in the counterfactual section.

204. Also, the Commission has assessed any possible anticompetitive effects of the transaction on the quality of the organic search as compared to the counterfactual explained above. Theoretically, the rationale for possibly degrading the organic search stems from the trade off that search platforms appear to face between the incentive to provide relevant organic and paid results. The trade off arises because when a platform tries to attract more users through greater relevance on the organic search it runs the risk of losing revenues on the advertising side (i.e. less clicks on ads) due to users clicking predominantly on the organic side (especially if both types of clicks would bring the user to the same kind of information). The degradation of organic search could possibly be achieved through various means. For example, platforms might have an incentive to dedicate a smaller part of the result page to organic results in favour of search advertising links (also called sponsored links) thereby providing proportionally more advertising links. Alternatively, the platforms may rank the sponsored and organic search results in a way that firms offering competing products to the sponsored links are ranked, from the user's perspective, on the organic side lower than optimally. Given that platforms provide, at least to some degree, different organic results and that the platforms are the gatekeepers of the information available on the web, it is inherently difficult for the user to assess whether the platform engages in this behaviour. Alternative search platforms may offer "checks" against such incentives of the search platforms. As a result of the transaction, one alternative source of such checks, Yahoo's independent search index and algorithm, will be eliminated and therefore to the extent that Yahoo's organic search is exercising such a constraint the transaction could possibly harm users.

205. The Commission also considered whether the transaction may lead to a decrease in variety available to users of internet search. The underlying rationale is that platforms provide differentiated services (based on inherently different algorithm and access to different index) to users who then self-select the platform that is likely to provide organic and/or sponsored links closer to their preferences. The loss of Yahoo as a competing platform may harm its current users as long as a close substitute is not present in the market, i.e. as long as Google or Microsoft does not offer an organic/sponsored search similar to what is offered today by Yahoo. The relevant benchmark to assess the possible effect of this theory is the counterfactual established above.

The notifying party's views

206. Microsoft submits that post-transaction the incentives to innovate and provide relevance would not be reduced. It argues that Microsoft and Yahoo will strongly compete for users and innovate in terms of content, features and products on their websites to attract users. A greater number of users will, in turn, make the website attractive to advertisers that value the ability to reach them.

207. According to Microsoft, Microsoft's and Yahoo's separate publishing businesses will compete, post-merger, to offer personalisation, new features, applications and complementary products. In fact, as users become more accustomed to the different features and product offerings of the various search experiences, they begin to customise and expand the search experience to their own personal preferences. Examples of personalisation include customising the number of results displayed for each query, security settings, etc.

208. Microsoft claims that the structure of the transaction provides a large incentive to innovate on the search experience because, according to the definitive Agreements, the publishing businesses of Microsoft and Yahoo will remain separate, and the revenue attributable to either Yahoo or Microsoft depends on the source of the search query. Specifically, Yahoo will only earn revenue from ads that appear on its own websites (and those of its publisher affiliates). Therefore, Microsoft argues that Microsoft and Yahoo will have a strong incentive to attract users to their own entry points.

209. Microsoft [...] submit that Yahoo has been a weak innovative force in the market of organic search in the last years and [...]. The notifying party argues that, following the transaction, Yahoo will be in a better position to compete for users since Yahoo's search-related profits are a direct function of maximising user-engagement on its O&O properties.

210. Furthermore, Microsoft submits that in terms of performance and feature offerings, Google is the market leader which tends to set de facto industry benchmarks and standards through innovation. Accordingly, search engines such as Yahoo and Bing have had to react continuously to new developments on Google's platform and they have focused on competing with Google rather than against each other.

211. The notifying party also submits that the transaction does not diminish but instead increases its incentives to invest in relevance (quality considerations). Microsoft argues that its incentives to maximise algorithmic relevance are aligned with its incentives to maximise paid advertisement revenue. The overriding objective of the combined platform, according to the notifying party, is to increase user engagement. Its incentive to compete for users will increase because any higher RPS resulting from the transaction will render each new user more valuable to Microsoft.

212. Furthermore, users, according to the notifying party, notice inferior (i.e. less relevant) search results and given the strong presence of Google in the market they would direct their queries back to Google lowering user engagement in Microsoft's platform. Microsoft also states that the search industry is closely monitored by industry participants and opinion leaders. Their view as well as mainstream media influences users' perceptions of search engines. Microsoft therefore claims that it cannot run the risk of even small degradations in quality, as users will become aware and switch to Google.

213. Finally the notifying party submits that following the transaction the existing matching algorithms will perform better at higher scale which as explained above will lead to higher advertisers and users engagement.

Results of market investigation and the Commission's analysis

214. The notifying party's view has been to a large extent confirmed during the market investigation and by the Commission's analysis of the file. However, given the limitations of a first phase market investigation, the Commission has not been able to confirm some of the arguments.

215. The Commission considers that Yahoo will still have the incentive to compete for users and thereby innovate post-transaction. This conclusion is supported by the structure of the transaction that leaves Yahoo a large degree of independence on the publishing side. In fact, the ability of Yahoo to complement its internet search services with other services traditionally offered to users on its portal-like web site will be independently continued post-transaction (67).

216. However, the Commission considers that this arises to the extent that users value the overall user experience highly relative to the results of the search query. Microsoft and Yahoo, however, have not provided strong evidence to confirm this.

217. In any case, Yahoo's users are not likely to be negatively affected by the transaction since Microsoft's innovations in internet search will be shared with Yahoo. Against the counterfactual, confirmed in the market investigation, that Yahoo, over the last three years, has been a weak innovative force (68) and that its market position is trailing that of the two other competing platforms, this sharing of innovation between Microsoft and Yahoo may ultimately benefit Yahoo's users.

218. Also, the incentives of Yahoo to enter into this agreement seem to confirm the view that Microsoft will have a strong incentive to innovate post-transaction. The overall profitability of Yahoo depends also on the performance of the internet search in its portal since a low quality of that search will jeopardise Yahoo's revenues on other complementary products offered through its portal.

219. Additionally, the market investigation has revealed that currently Google enjoys a large competitive advantage compared to other search engines (69) and is perceived as a "must-have" for users. Therefore, it is possible that if the transaction, through the scale effects, leads to a stronger competitor more able to innovate, Google will also have an incentive to keep, or even accelerate, its innovation efforts in the market.

220. Concerning the possible post-merger effects of the transaction on the level of relevance, the respondents to the market investigation stated that algorithmic search engines' quality and relevance are the most important factors to attract users to a particular search engine (70). Therefore, users are considered to be responsive with respect to variations to the relevance provided by a search engine.

221. The very limited share of user multi-homing between Microsoft and Yahoo shows that users rarely run checks between these two platforms (71).

222. Additionally, it should be noted that currently Ask provides "independent" organic search results across a number of European countries. Although Ask partially has an agreement with Google to have access to its index, its organic search results may exercise a constraint on the main competing platforms not to degrade their organic search relevance. Furthermore, the existence of vertical search platforms also may further limit the impact of the transaction on quality as well as on the variety of choice at least for the verticals under consideration.

223. Furthermore, as submitted by the notifying party and as analysed above, the effects of scale are likely to allow the merged entity to run more tests and experiments on the algorithm in order to improve its relevance.

224. Concerning the possible reduction of variety of choice, it should be noted that the transaction does not significantly affect the Yahoo users that value more the overall user experience rather than the results of the organic search since Yahoo will post-merger still offer its independent complementary services. Also, the [...] share of users that single-home on Yahoo indicates that the potential loss of variety will affect a limited number of users. Furthermore, the very small degree of multi-homing of users between Yahoo and Microsoft shows that Yahoo users' second choice for a search platform is provided by Google and not Microsoft.

225. To the extent that the transaction will allow the merging parties to provide greater relevance through greater scale, the effect on variety is likely to be limited. Also it is plausible that the merged entity through innovation and through its access to a larger index will be able to provide personalised search results better aligned to users' preferences.

226. In conclusion, compared to the counterfactual as described above where Yahoo is a weak innovative force, the available evidence to the Commission does not confirm that the transaction may raise risks on the market for internet search through its effects on innovation, relevance and variety. Furthermore, it is possible that due to the transaction some benefits will materialise due to larger scale of the merged entity. On this basis, the transaction does not give rise to serious doubts as to its compatibility with the common market as far as its effects on search engine users are concerned.

3. Effects on publishers

Publishers' business

227. The business of online publishers consists of making available web content to internet users, for example various forms of journalism, literature, entertainment in general. The types of content include pure text, graphics, audio, video and combinations thereof.

228. Publishers follow different business models. A few publishers earn all their revenues from charging their users directly for their use of the content, i.e. through subscriptions or ad-hoc payments for the consumption of content. (72) Other publishers offer their content to users for free but sell parts of the space on their web pages to interested advertisers. Many publishers combine both approaches by, for example, offering some free content, monetising the attracted users through advertising and offering additional content for a fee.

229. For this reason publishers are interested in simple and efficient ways of monetising their online inventory (the totality of the space on web pages that can be dedicated to advertisements). There are alternative ways to monetise through advertisements. Publishers can sell their inventory directly to interested advertisers or to media agencies that pool advertisers' demand for inventory. Publishers can also sell their inventory indirectly by making it available to an intermediary (such as an ad exchange or an ad network) which earns a fee for selling the inventory to advertisers.

230. Large publishers normally combine both direct and indirect monetisation because not all parts of their inventories have the same characteristics. For example, the top section of a publishers "home page" (i.e. the "portal" to its offerings) would be especially interesting for advertisers who want to reach a large audience and do not care for very precise targeting of specific user groups. This type of inventory will often be sold directly. In contrast, white space next to results of a search function for a web site's content or next to many newspaper articles of a more general nature would be more suitable for indirect selling.

231. Publishers are interested in the profit they can earn from their inventory. When comparing different monetisation options they therefore compare the expected profit relative to the amount of inventory at stake. For example, when comparing two direct offers from advertisers for the same inventory, a publisher would compare how much they would be willing to pay for each impression of their ads and select the higher bidder (73).

232. The Microsoft and Yahoo search businesses both have a search advertisement platform pooling advertisers' demand for web space. In addition to selling their own web space to these advertisers, both Microsoft and Yahoo partner with publishers to also make available these publishers' inventory to their advertisers (74).

Theory of harm and analysis

233. There are currently only three relevant independent technical platforms for intermediation of search ads (or closely related text-based contextual ads (75)), which operate on an international level: Microsoft, the Yahoo search business and Google. The proposed transaction technically removes one advertising platform. Even though the combining platforms currently service a very low number of publishers in the EEA, this could theoretically harm publishers because it may result in reduced competition between the remaining advertising platforms and consequently lower prices paid for publishers' web space.

234. However, it is unlikely that publishers' monetisation opportunities will be negatively impacted by the proposed combination of Microsoft and the Yahoo search business. First, there is no indication that Yahoo will not continue to act as an intermediary. According to the Agreements, Yahoo can both continue its current syndication agreements and enter into new such agreements. While Yahoo will source both algorithmic search results and paid listings from the combined entity's technical platform, it will nevertheless be able to offer independent pricing to its publishers. (76) Indeed, publishers expect Yahoo's competitiveness vis-à-vis Google to benefit from the proposed transaction. (77) Second, Microsoft currently has no syndication agreements in the EEA and that Yahoo's share of the syndication of search ads in the EEA does not exceed its overall share in paid search. (78) It is likely that the transaction will give both Microsoft and Yahoo a chance to increase their syndication businesses in the EEA. (79) In this sense the transaction has the potential to increase the competition for publishers. Third, it can be expected that competition between Microsoft and Yahoo for syndication agreements will increase post-transaction because both will have access to a combined pool of advertisers and thus to potentially higher revenues from every syndication agreement. (80)

235. In response to the market investigation, publishers also identified benefits of the transaction in relation to Google. They consider Google to be a partner that is difficult to avoid but that also comes with certain problems. For example, certain respondents pointed out that there is a lack of transparency for publishers when dealing with Google because the exact ways in which revenues are shared are not known to publishers. Some publishers expect the transaction to create real competition to Google that would make it more likely for Google to address concerns of its partners such as this one because the existence of stronger competitors to Google will give publishers a better bargaining position.

236. The same line of reasoning can be applied to the financial aspect of Google's offerings. Currently, Google can afford to offer lower TAC percentages to publishers than its competitors because it has higher revenue per search (RPS). (81) It can set its TAC percentage in a way that ensures that in absolute terms its offer to publishers is only very slightly higher than what its competitors can offer. Following the transaction, the combined platform with access to a larger pool of advertisers will likely be able to generate a higher RPS than today. Consequently, Google would have to increase the TAC percentage it offers to publishers if it wants to continue outbidding its competitors vis-à-vis publishers. This would be beneficial to publishers.

237. In addition to the analysis of this theory of harm it is also important to consider, from the perspective of publishers, the situation in which the proposed transaction would not have taken place. In such a situation it would take Microsoft much longer to build up a presence in the various national markets within the EEA than it would posttransaction, leaving Yahoo in the role of sole competitor against Google in many such markets. For publishers such a situation would be less beneficial because the transaction would increase the attractiveness of both Microsoft's and Yahoo's offerings.

Conclusion

238. On this basis, the transaction does not give rise to serious doubts as to its compatibility with the common market as far as its effects on publishers in both the market for online advertising and the potential market for search intermediation are concerned.

4. Effects on distributors

Distributors' business

239. As explained in section V.A.3., there are a number of different entry points ("search technology") and a wide range of different groups of distributors for these entry points, including, in particular, OEMs and ISVs. Competition for distribution deals is generally driven by the returns which distributors expect to achieve by directing new users to the search platform's site. Distributors receive a payment for deploying toolbars, default settings or any other entry point to their customers. Thus, according to Microsoft, they have an incentive to conclude a deal with the search platform which will guarantee the highest payment.

240. Distribution deals follow different payment models, which may also be combined. These include minimum revenue guarantees, placement fees (for example per PC sold), activation fees (triggered by the first search performed by the customer through the relevant entry point) and revenue sharing arrangements. The less control a distributor has over the event triggering the payment (for example for activation fees on a toolbar), the riskier a deal can be for it. The larger a search platform's market share and search advertising revenue, the more attractive revenue sharing arrangements with this platform can become. The lower the search platform's market share, the more it may have to guarantee a minimum revenue (possibly as an up front payment) to obtain a distribution agreement. Distributors attempting to maximize their revenue from the distribution of entry points will therefore not necessarily agree on the same pricing model with any given search platform. Distribution deals for entry points are at times part of larger deals also covering the distribution of non-search products.

241. According to Microsoft, the lower revenue per search vis-à-vis Google creates a disadvantage for both Microsoft and Yahoo on the search component of distribution deals.

Theory of harm and analysis

242. The proposed transaction technically removes one advertisement platform. Theoretically this could harm distributors for the following reasons. First, there may be reduced competition between search platforms for distribution agreements as a result of the proposed transaction, making it more difficult for distributors to secure attractive deals as the combined entity might have a stronger bargaining position.

243. Secondly, in a conglomerate context, the proposed transaction may increase Microsoft's ability to leverage its market power in areas other than online advertising (for example client PC operating systems and personal productivity applications) when negotiating distribution agreements for its search technology (for example by bundling products). This may ultimately affect users.

244. However, according to the market investigation, future harm appears to be unlikely to be significant. Indeed, the market investigation shows that the large majority of distributors having responded expect the proposed transaction to positively influence their opportunities to monetize the distribution of search technology. No distributor has come forward to the Commission to oppose the transaction.

245. Regarding the first concern, the market investigation shows that today Microsoft and Yahoo currently only compete weakly for distributors of their respective search technology in the EEA (82). Out of twenty-five distribution contracts on which detailed data was provided by nine respondents, Microsoft and Yahoo were among the successful bidder and the runner up (i.e. the top two bidders) only in three instances. (83) The market investigation also shows that the closest competitor of both companies respectively is Google. (84) Furthermore, under the proposed transaction, Yahoo will have no fewer incentives to compete for distribution deals than before the transaction, as it will only generate search revenue from users that Yahoo itself attracts to its own websites. As far as Yahoo's ability to compete for distribution deals is concerned, distributors having responded to the market investigation do not expect Yahoo's ability to compete with Google for distribution deals in the EEA to be negatively affected by the proposed transaction (85). The majority of the respondents to the Distributors questionnaire expects Microsoft's ability to compete with Google for distribution deals to increase under the proposed transaction (86).

246. As mentioned before, in line with these findings, the market investigation also shows that the large majority of the respondents to the Distributors questionnaire expect the proposed transaction not to negatively influence their opportunities to monetize the distribution of search technology, but actually to increase these opportunities (87). To the extent the respondents give reasons for their assessment, they expect Microsoft and Yahoo to be better competitors to Google than before the proposed transaction. Those respondents consider that the transaction would increase Microsoft's and Yahoo's ability to pay for search distribution and thus lead to better outcomes (from the distributors' perspective) of competitive bidding situations for search distribution. The major concern for distributors seems to be Google's bargaining power which could possibly be constrained to the benefit of distributors due to the expected strengthening of its competitors (Microsoft in particular) post-transaction.

247. This analysis of the competitive situation appears plausible given Google's apparent lead over Microsoft and Yahoo in RPS, as explained in section VI.D., which cannot be sustainably compensated for by minimum revenue guarantees if these are not matched by comparable search revenues of the guarantor.

248. Regarding the second concern, the agreement does not seem to contain any elements that would suggest that the transaction would increase Microsoft's ability to leverage its strong market position in non-search areas onto any of the relevant markets as discussed in section V. The proposed transaction neither foresees Microsoft entering a new market susceptible to leveraging nor distributing search technology via different types of distributors than before the transaction. Its ability to tie or bundle products, for example its client PC operating system Windows or its personal productivity application Office, with search technology does therefore not appear to be affected by the proposed transaction. Nor is the transaction likely to change Microsoft's leveraging incentives given the limited change in the ratio between the revenues generated by Microsoft in the former business segments with those generated in the latter business segment post-transaction. (88) The market investigation confirms this analysis. (89)

Conclusion

249. On this basis, the transaction does not give rise to serious doubts as to its compatibility with the common market as far as its effects on distributors in a potential market for the distribution of entry points is concerned.

F. OVERALL RESULTS OF THE INVESTIGATION

250. The first phase investigation conducted by the Commission has confirmed to a large extent the submissions of the notifying party.

251. None of the respondents to the market investigation expects the transaction to have negative effects on competition.

252. Microsoft's and Yahoo's activities in internet search and search advertising in the EEA are very limited. The parties have a combined market share below 10%, while Google has a market share in excess of 90%. The transaction benefits thus from a presumption of compatibility under the Horizontal Merger Guidelines.

253. Furthermore, both Yahoo and Microsoft have encountered difficulties in competing with Google. During the last three years, Yahoo has been losing market share essentially to Google.

254. The market investigation has also shown that Yahoo finds it difficult to compete with Google in particular because the required costs are too high. Microsoft finds it difficult to compete with Google because amongst other reasons it does not have the necessary scale.

255. Moreover, a detailed analysis of the effects on advertisers, users, publishers and distributors has shown that (i) in several areas Microsoft and Yahoo were not competing to a significant degree with each other before the transaction (for example in search advertising, adCenter competes with Panama only in the UK and France) and (ii) in other areas Microsoft and Yahoo will continue to compete at least to some extent posttransaction (for example on the user experience).

256. Against that background a vast majority of the respondents to the market investigation stated that the transaction will be pro-competitive allowing the parties to more effectively compete with Google.

VII. CONCLUSION

257. For the above reasons, the Commission has decided not to oppose the notified operation and to declare it compatible with the common market and with the EEA Agreement. This decision is adopted in application of Article 6(1)(b) of Council Regulation (EC) No 139-2004.

Notes:

1 OJ L 24, 29.1.2004, p. 1.

2 Syndication properties are advertising space on websites that do not belong to the advertising platforms themselves but to third parties with whom they have concluded a syndication partnership. For example, if a newspaper has a syndication partnership with Yahoo, the web pages on which it lets Yahoo put ads are part of Yahoo's syndication properties.

3 The Commission relies in this regard on the relevant provisions of the Consolidated Jurisdictional Notice under Council Regulation (EC) No 139-2004 on the control of concentrations between undertakings (OJ C 95, 16.4.2008) ("the CJN").

4 Paragraph 72 of the CJN.

5 Paragraph 77 of the CJN.

6 Turnover calculated in accordance with Article 5(1) of the EC Merger Regulation.

7 Financial year ending on 30 June 2009.

8 Financial year ending on 31 December 2008.

9 Microsoft: EUR [...] million in France, EUR [...] million in Germany and EUR [...] million in the UK; the Yahoo search business: EUR [...] million in France, [...] million in Germany and EUR [...] million in the UK.

10 A web crawler is a software application that downloads web pages. Typically web crawlers have the ability to determine the addresses of more web pages to download by extracting the links contained on already downloaded web pages. Web crawlers are used by search engines to build their indexes of web content. For this purpose the content of the downloaded pages is further analysed and indexed (for example, the language of the page and when it was last changed is determined). The resulting index is a database that is searched when the user enters a query into the search box. Web crawlers have to work continuously in order to keep the index "fresh", i.e. to ensure that the indexed content to the largest possible extent equals the actual content of the web pages.

11 An auto search is activated when a user enters a search term into the address bar of its browser instead of entering a URL. The browser will automatically redirect the user to a pre-determined search engine, which will display results that are responsive to the search terms used. A toolbar is a software add-on to a web browser. It appears as a row or column of on-screen buttons within the browser window and can among other functionalities include a search box. Users are able to enter search terms into the toolbar at any time, regardless of the web page they are viewing. Toolbars can be affiliated with a particular search engine. Unlike a toolbar, a browser search box is a component of the web browser itself and so is automatically supplied by the browser provider without the need for installation. Most major web browsers, including Internet Explorer, Firefox, Safari and Opera offer a search box functionality. The use of the browser search box is similar to the use of the search box in a toolbar. A web portal is a website that functions as a gateway through which users can access a range of online services. It operates by aggregating a large number of internet users or subscribers around specific types of services. Web portals generate revenue for their publisher in a number of ways, including through advertising, subscriptions for paid-for content, and commissions on e-commerce transactions. The alternatives through which users access search platforms are developing rapidly with new entry points emerging all the time. In particular, many users now use smart phones to search the internet, or access search functions via interactive subscription television. There also non-browser based search functionalities accessed via software such as media players, instant messengers, desktop widgets and file-sharing tools.

12 Case COMP/M.4731 Google/DoubleClick, 11 March 2008, paragraphs 44 to 47 and paragraph 56.

13 Case COMP/M.4731 Google/DoubleClick, 11 March 2008, paragraphs 48 to 56.

14 32 out of 48 responses to question 1 of all questionnaires (competitors, advertisers, publishers, distributors, consumer associations, media agencies).

15 16 out of 42 responses to question 3 of all questionnaires.

16 Please refer to paragraph 43.

17 See answers to question 2 of all questionnaires.

18 31 out of 45 responses to question 3 of all questionnaires.

19 Case COMP/M.4731 Google/DoubleClick, 11 March 2008, paragraph 68.

20 Case COMP/M.4731 Google/DoubleClick, 11 March 2008, paragraphs 70 to 73.

21 Case COMP/M.4731 Google/DoubleClick, 11 March 2008, paragraphs 82 to 84.

22 31 out of 37 responses to question 20 of the Advertisers and Media Agencies questionnaires, to question 17 of the Publishers questionnaire, to question 18 of the Competitors questionnaire and to question 8 of the Distributors questionnaire.

23 9 out of 11 responses to questions 21 and 22 of the Advertisers and Media Agencies questionnaires.

24 Data provided by the notifying party, source IAB Europe/PWC.

25 End user multi-homing in European Countries, Keystone strategy.

26 May 2008 data.

27 Paid search is a search that fulfils the following two criteria: 1. it provides general web search and 2. it shows paid keyword ads on the search result pages.

28 Note that figures provided concern paid search query data and not simply search query.

29 Please note that for some countries numbers may exceed 100 due to rounding.

30 See paragraph 18 of the Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings.

31 See paragraph 9 of the Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings.

32 AdEx 2008. European online advertising expenditure. IAB Europe.

33 Yahoo Annual Report 2008.

34 http://www.123jump.com/earnings-calls/Yahoo!-Q4-Earnings-Call-Transcript/36294/21

35 9 out of 12 respondents to question 66 of the Competitors questionnaire.

36 14 out of 15 respondents to question 89 of the Advertisers questionnaire and question 90 of the Media Agencies questionnaire.

37 10 out of 15 respondents to question 81 of the Advertisers questionnaire and question 82 of the Media Agencies questionnaire.

38 11 out of 15 respondents to question 83 of the Advertisers questionnaire and question 84 of the Media Agencies questionnaire.

39 http://www.eweek.com/c/a/Search-Engines/Microsoft-Bing-Grabs-Search-Share-From-Google-Yahoo- 875016/

40 See answers to question 86 of the Advertisers questionnaire and question 70 of the Competitors questionnaire.

41 See answers to question 78.1 of the Advertisers questionnaire and question 79.1 of the Media Agencies questionnaire.

42 See in particular answers to question 71.1 of the Advertisers questionnaire.

43 Data provided by the notifying party, source IAB Europe/PWC.

44 Microsoft Algo Relevance Benchmarking, Keystone Strategy.

45 Yahoo internal documents.

46 Yahoo internal documents.

47 Form CO, Annex 63 Scale, Innovation and Competition in Search (14 October 2009).

48 Annex 48, Form CO

49 10 out 11 answers to question 70 of the Competitors questionnaire.

50 14 out of 15 respondents to questions 98, 99 and 100 of the Advertisers and Media Agencies questionnaires expect the combined Microsoft / Yahoo search business entity to be a stronger competitor to Google that will give them more advantages than disadvantages.

51 In a standard (non position) second price auction the highest bidder wins, but the price paid is the secondhighest bid.

52 9 out of 17 respondents to question 95 of the Advertisers questionnaire and question 96 of the Media Agencies questionnaire (out of the 17 only 4 replied yes).

53 See replies to questions 96 of the Advertisers questionnaire and question 97 of the Media Agencies questionnaire.

54 11 out of 15 respondents to question 80 of the Advertisers questionnaire and question 81 of the Media Agencies questionnaire.

55 14 out of 16 respondents to question 78 of the Advertisers questionnaire and question 79 of the Media Agencies questionnaire with regard to Microsoft. And 13 out of 16 respondents to question 79 of the Advertisers questionnaire and question 80 of the Media Agencies questionnaire with regard to Yahoo.

56 The market investigation indicated that many advertisers are to some extent budget constrained (12 out of 16 respondents to question 56 of the Advertisers questionnaire and question 57 of the Media Agencies questionnaire). However it is not easy to interpret these responses in the context of this theory because the respondents did not give a precise indication of the degree of their constraint and of how often their constraint is exhausted.

57 11 out of 14 respondents to question 38 of the Advertisers questionnaire and question 39 of the Media Agencies questionnaire.

58 6 out of 12 respondents to question 72 of the Advertisers questionnaire and question 73 of the Media Agencies questionnaire.

59 8 out of 12 respondents to question 73 of the Advertisers questionnaire and question 74 of the Media Agencies questionnaire.

60 11 out of 14 respondents to question 75 of the Advertisers questionnaire and question 76 of the Media Agencies questionnaire. 3 respondents perceive the ROI offered by Google comparable to Yahoo's.

61 11 out of 15 respondents to question 76 of the Advertisers questionnaire and question 77 of the Media Agencies questionnaire. 2 respondents perceive the ROI offered by Google comparable to Microsoft and 2 respondents perceive the ROI offered by Google lower to Microsoft.

62 13 out of 14 respondents to question 86 of the Advertisers questionnaire.

63 14 out of 15 respondents to questions 98 and 99 of the Advertisers questionnaire and questions 99 and 100 of the Media Agencies questionnaire.

64 Only one advertiser replied negatively to question 98 to the Advertisers questionnaire. See also replies to question 99 of the Media Agencies questionnaire.

65 Only one negative reply to question 70 of the Advertisers questionnaire and question 71 of the Media

66 Paragraph 8 of the Horizontal Merger Guidelines.

67 The above considerations are especially relevant for portal entry, which is very significant for Yahoo, but less so for other entry points such as syndication or browser search box, since users value the complementary services of the portal where they receive a user experience.

68 8 out of 12 respondents to question 89 of the Advertisers questionnaire.

69 8 out of 8 respondents to question 78 of the Competitors questionnaire. Note that Google has not been counted in the respondents of the Competitors questionnaire.

70 See answers received to question 34 of the Consumer Associations questionnaire.

71 However, this need not be indicative of no exercise of competitive constraint. The mere presence of an alternative check may suffice to induce the search engines to enhance the relevance of their organic search.

72 This model is not very common for online publishers because as long as the advertisement is not intrusive it can be combined with non-free offerings for additional profit.

73 This reasoning assumes that the publisher's non-monetary considerations with regard to the type and content of advertising in question are satisfied, i.e. some advertisers likely would not agree to carry offensive ads even if they would bring a higher short-term profit.

74 When publishers deal with advertisers directly they are usually paid a fixed sum per unit of inventory that they have made available. As explained above, when dealing with intermediaries that provide syndicate ads for which advertisers pay only when they are clicked, such as Google or Yahoo, publishers are usually remunerated with a percentage of the total revenues generated by clicks on ads shown on their web sites. From the intermediaries perspective such payments to publishers can be characterized as a TAC.

75 Usually advertisers use the same ads for both purposes.

76 For pure search (and related contextual) advertisement partnerships a theoretical argument could be made that Microsoft will always be able to outbid Yahoo since Yahoo will only be able to share 88% of the concerned revenues with a publisher whereas Microsoft is not constrained in this way. However, it is significant that Yahoo also has a display advertisement offering that may be covered in some agreements, making a direct price comparison with Microsoft much more difficult for publishers. Moreover, any bidding war between Microsoft and Yahoo would be beneficial to publishers.

77 In responding to question 24 of the market investigation questionnaire sent to publishers, 6 out of 10 respondents indicated that they expect Yahoo will more effectively compete with Google as a result of the transaction (1 said Yahoo's competitiveness would be unchanged while 3 said they did not know).

78 For example, in the UK, where most of Yahoo's largest syndication partners operate, Google's two largest syndication partner AOL and Ask.com have a higher market share than Yahoo has for syndication and its owned and operated sites combined. It is also significant that Yahoo does not offer self-service syndication comparable to Google's AdSense offering. Consequently Yahoo has almost no agreements with very small publishers of which Google has a very large number.

79 Publishers expect that post-transaction Microsoft will be a more effective competitor to Google for deals with publishers. Indeed, in responding to question 23 of the market investigation questionnaire sent to publishers 9 out of 10 respondents indicated that they expect Microsoft will more effectively compete with Google as a result of the transaction (1 said it did not know). Responses to question 90 of the questionnaire sent to competitors confirm this result. 6 out of 7 respondents said that Microsoft will be able to more effectively compete with Google whereas 1 said that Microsoft's competitiveness will be unchanged.

80 In the market investigation, publishers confirmed that Microsoft and Yahoo currently compete only moderately for such deals. See the responses to question 22 of the market investigation questionnaire sent to publishers where only 2 out of 10 respondents said that Microsoft and Yahoo currently compete "strongly" whereas 6 said they compete either weakly, moderately or not at all (2 said they did not know). Responses to question 89 of the questionnaire sent to competitors confirm this result. 2 out of 7 respondents each said that Microsoft and Yahoo currently compete either weakly, moderately or not at all (1 said it did not know). This argument assumes that currently there is a group of advertisers who do not multi-home on both Yahoo and Microsoft, i.e. advertisers who bid for search ads on Yahoo or on Microsoft but not on both. The larger this group is, the more attractive will both platforms become. (This is related to the argument about the bidder density in an auction system; see above the section related to the effects from the perspective of advertisers.)

81 Note that the payment to the publisher is determined by multiplying the TAC percentage, the RPS and the number of searches effected on the publisher's site. The first two factors are directly affected if a publisher switches to another search intermediary. The last one is constant, at least in the short run (in the long run worse, or different, search results may make users less willing to actually search on the publisher's site).

82 7 out of 9 respondents expressing an opinion on question 13 of the Distributors questionnaire.

83 These three instances are attributable to two distributors which both consider current competition between Microsoft and Yahoo for search distributors of their respective search technology in the EEA to be weak.

84 10 out of 10 clear responses to question 12 of the Distributors questionnaire for Microsoft and 7 thereof for Yahoo.

85 All respondents expressing an opinion on question 16 of the Distributors questionnaire.

88 The market investigation confirms this analysis.89

86 7 out of 8 respondents expressing an expectation in response to question 15 of the Distributors questionnaire.

87 10 out of 12 respondents expressing an expectation in response to question 19 of the Distributors questionnaire.

88 Microsoft's worldwide revenues in its financial year 2009 (ending 30 June 2009) in the business segment "Client" (comprising its Windows offerings) amounted to EUR 10 504 million (EEA: EUR [...] million) and to EUR 13 774 million (EEA: [...] million) in the business segment "Microsoft Business Division" (comprising its Office offerings) compared to only EUR [...] million generated with search advertising in the calendar year 2008 (EEA: EUR [...] million). Yahoo's revenue generated worldwide with search advertising amounted to EUR [...] million in its 2008 financial year (ending 31 December 2008) (EEA: EUR [...] million). In view of the 88%/12% revenue sharing arrangement between Microsoft and Yahoo under the proposed transaction, only a fraction of this revenue can be used to estimate Microsoft's search advertising revenue post-transaction.

89 6 out of 9 respondents expressing an expectation in response to question 17 of the Distributors questionnaire and 9 out of 12 respondents expressing an expectation in response to question 87 of the Competitors questionnaire expect Microsoft ability to remain unchanged.