Did Enforcers Make the Right Call When Clearing Google/Doubleclick?

Since its inception in 1998, Google has been on a virtual shopping spree, acquiring 256 companies,¹ including YouTube, Android, Waze, Fitbit, and DoubleClick, to mention a few. While most have gone unnoticed or at least unquestioned, this has not been the case with the 2007 acquisition of DoubleClick. Many have noted² how this acquisition has positioned Google to leverage market power into adjacent and emerging online services, creating today’s digital behemoth. Neither the Directorate General for Competition of the European Commission (DG COMP), enforcing competition law in the EU,³ nor its US counterpart, the Federal Trade Commission (FTC),⁴ opposed the transaction in 2007, later creating a perception of an enforcement failure. However, this criticism relies on hindsight⁵ and ignores how both agencies undertook substantial analysis before allowing the transaction to move forward, only clearing it after being convinced that it did not endanger competition. In the case of the FTC, the votes were split 4-1, although the argument of the one commissioner who opposed the transaction⁶ was in a sense supplemented by a second commissioner who, while concurring with the majority opinion, felt compelled to issue a separate opinion⁷ outlining how many people (probably including himself) didn’t “… fully understand the still nascent Internet advertising market”. Naturally, this does not necessarily absolve enforcers of misdoing. Still, it indicates that genuine concerns were identified in 2007, and that a serious attempt to evaluate the transaction properly was attempted across enforcers. Nevertheless, revisiting the motivations, including embedded assumptions and analysis, appears prudent, given that––as some suggest––subsequent developments might have proven these wrong.

Before examining enforcers’ (positive) review of the Google/DoubleClick transaction, the Ad Tech Industry must be presented in order to understand their conclusions better. Subsequently, this will be summarised and evaluated in more detail. For the sake of brevity only DG COMP’s decision will be analysed, as this has been published and comes with substantial internal reflections and consideration. By contrast, the publicly available FTC’s Decision is short (13 pages) and offers few substantiated analyses that can be dissected and evaluated.⁸Moreover, several concerns expressed therein pertain to privacy matters rather than antitrust.

Google acquired DoubleClick in 2007 after a bidding war in which Google agreed to pay USD 3.1 billion. It is understood that the other contenders included Microsoft and Yahoo!,⁹ making it plausible that DoubleClick would have ended up with one of these had Google not outbid them.¹⁰ Before the acquisition and its potential anti-competitive elements can be discussed, comments on the online advertising business must be offered, including how DoubleClick fits into the pending Ad Tech antitrust investigations against Google. As outlined initially, DoubleClick is often claimed to be the enabler of Google’s later (alleged) crime spree.

DoubleClick, in 2007, offered services that we today attribute to the Ad Tech Industry, involving the selling and buying of online advertising, or more specifically, the facilitation of such transactions.¹¹ Companies like DoubleClick (and Google) do not sell or buy advertising directly but function as intermediators between the advertisers and the owners of online advertising spaces, where the latter are commonly referred to as “publishers”. Using this method, the intermediator¹² operates as a two-sided platform or a network, pooling available online advertising space and/or advertisers looking for such and matching them as an alternative to direct sales between the two parties. Besides online operators like DoubleClick (and Google), traditional media agencies were also offering intermediary services in 2007.

Online ads can roughly be categorised as either searchable or non-searchable ads, where searchable ads appear next to the organic blue links yielded in response to user inquiries on an internet search engine. These are selected based on the user’s search keyword and generate payment for the owner of the search engine. Non-searchable ads, also called display ads,¹³ can appear on any webpage and be texted or enriched with pictures or videos. Often, they can be found on content pages like YouTube and Facebook or traditional content providers like online newspapers, where the owners (the publishers) monetise their real estate by making it available to advertisers.¹⁴ While searchable ads were formerly traditional revenue streams for Google and other providers of search engines, the industry is shifting to non-searchable (display) ads, making these more prominent in terms of market size and revenue.

Intermediators, connecting advertisers and publishers, offer several partly overlapping and complementary services when an advertiser wants to contract with publishers to run a campaign. Traditionally, this would be done manually and based on meetings and discussions, and the publishers would be newspapers and the owners of public billboards. However, the proliferation of online advertising has fundamentally changed this industry, including how to buy and sell advertising. Not only have newspapers and billboards been replaced by online options, but the whole process has been automated, such that increasingly, advertisers’ and publishers’ bids are cleared over an exchange akin to a traditional marketplace or stock exchange. How this is done depends on many factors, including whether one is a large or small advertiser; and direct sales remain an alternative to relying on an intermediator, particularly for premium space and by large publishers.¹⁵

DoubleClick was and is an intermediator, but its activities originally had a more limited and targeted focus. After acquiring DoubleClick, Google changed its functionalities, and DoubleClick today serves publishers and advertisers more broadly than in 2007. However, at that time, two separate services had already emerged, involving a) the transaction directed at matching advertisers and publishers and b) the subsequent placement of the advertising. The former consists of buying and selling online advertising space, whereas the latter involves the physical placement, or management, of the advertising on the acquired online space. Intermediators, therefore, often bundle their services, managing both the sales process and the subsequent placement of the advertising, but, e.g., publishers relying on direct sales will source only the management part.¹⁶ As explained below, in 2007 DoubleClick’s activities were confined to managing the placement of advertising space once acquired, with a minimal stake in buying and selling advertising space.¹⁷ Google, in contrast, offered both as a bundle in 2007, though it focused on facilitating the sale, with the placement being more of an ancillary service. This explains why, e.g., DG COMP did not identify Google and DoubleClick as direct competitors during the review, as outlined below.

As already indicated, Google serves both advertisers and publishers when it comes to non-searchable display ads. As YouTube’s owner, Google is probably also one of the most significant publishers, and Google even owns the most prominent exchange (AdX) where bids are cleared. This creates an inherent conflict of interest in which Google could take advantage of its position and show partiality to its own services, to the detriment of competitors and customers (i.e. the advertisers and publishers).¹⁸ If pursuing a foreclosure strategy, the victims could also include competing exchanges. The risk is not abstract. In 2021, the French Competition Authority concluded¹⁹ that Google had failed to secure the best bids, at the expense of publishers and advertisers. Several enforcers have followed suit and opened investigations, including DG COMP²⁰ (2023) and the US Department of Justice²¹ (2023), but only the French case had been decided by the end of 2024. Presumably, the investigations vary across jurisdictions to take account of local factors. Still, at the core, they are directed at the embedded conflict of interest associated with serving several parties with opposing interests within the Ad Tech Industry. Moreover, the services Google acquired through DoubleClick play a pivotal role. With the benefit of hindsight, it’s tempting to conclude that DG COMP and the FTC got it wrong when they did not oppose the acquisition in 2007.

Google acquired DoubleClick in 2007 for USD 3.1 billion, but as DoubleClick’s turnover did not meet the required thresholds (> EUR 250 million in the EU), the transaction was not initially noticeable under the EU’s Merger Regulation.²² However, upon a request from Google, the case was referred to DG COM in order to avoid parallel and overlapping reviews in several EU member states. The EU’s Merger Regulation provided for this on condition that no member state would object. After opening a phase II investigation and some lengthy and complex evaluations, DG COMP cleared the transaction unconditionally in March 2008.²³ DG COMP’s Decision and its assumptions are outlined below, including comments. Section 3 will offer more substantial comments, including where DG COMP might have gone wrong and why.

When reviewing DG COMP’s move not to oppose the transaction, we must first understand that online marketing in 2007 was still nascent, accounting for less than 6 % of global advertising (2006),²⁴ and with searchable slightly larger than non-searchable online advertising. It had excellent growth potential, with projections for 2007 of 28.2 %.²⁵ Google probably understood this earlier than others—including traditional advertisers and publishers, who still relied on newspapers, magazines, and billboards. This is obvious today, where online marketing, both searchable and non-searchable advertising, is booming while conventional channels have stagnated, but in 2007, things looked very different. Across DG COMP’s Decision, the nascent and immature nature of the online advertising market is apparent, complicating data collecting and verification and often limiting the indicated market data to some broad intervals.

Both Google and DoubleClick were active in online advertising, and in its decision DG COMP defined this as a separate market from offline marketing. DG COMP contemplated²⁶ further distinctions, including searchable vs. non-search advertising, large vs. small publishers, and direct sales vs. the use of intermediates. However, no final decision was required, as it did not influence the transaction assessment. By contrast,²⁷ it was relevant to identify separate markets for a) publishers and b) advertisers, because they required different functionalities, and between c) buying and selling space, called intermediation services and d) the subsequent management of these, including placing the ads, called ad-serving services. As indicated earlier, many intermediators offered both as a bundle, but separate services could be identified. Not only were they clearly distinctive and separate services, but some providers, such as DoubleClick, had specialized their business case, only offering to manage the placement of advertising space once acquired, in contrast to Google, which, in principle, offered both services.²⁸

DG COMP, therefore, defined three relevant product markets that the transaction could potentially influence, encompassing the market for:²⁹

While these markets are not directly comparable to the markets defined in the recent French Ad Tech³⁰ case (2021) and DG COMP’s Google AdSense³¹ (2019), they nevertheless correlate with them, indicating that this part of the Decision has aged well. The only thing that differs materially is how Google AdSense distinguishes searchable from non-searchable advertising and sees a difference between relying on intermediates and direct sales. However, this probably stems from the nascent nature of the services in 2007 and the way in which the market subsequently has matured.³²

To evaluate a transaction, reliable market data must be collected, usually by the parties, but DG COMP must be able to evaluate them, including their quality, and all steps in this chain proved challenging in 2007. DG COMP appeared uneasy about the submitted market data on several occasions,³³ as these could not be verified or came with broad margins. Eventually, a baseline was established,³⁴ including the finding that Google was the leading provider of online advertising, particularly concerning searchable advertising, with a 50-70 % EU market share. By contrast, Google had only a minimal stake in non-searchable (display) advertising (< 5 %), while DoubleClick had 30-50 % here, making it the EU market leader,³⁵ with aQuantive/Atlas (recently acquired by Microsoft) as a strong contender. However, as 60-75 % of non-searchable (display) advertising was traded directly³⁶ between advertisers and publishers without involving an intermediator, it was evident that the market was nascent and very immature in 2007. Moreover, Google and DoubleClick’s high market shares did not account for direct or internal captive sales, indicating that they were inflated.³⁷

While both Google and DoubleClick were seen as leading suppliers in their respective sectors, with high market shares,³⁸ they had specialized their operations and were not identified as direct competitors by DG COMP.³⁹ With the benefit of hindsight, this might be incorrect,⁴⁰ but it did not influence the evaluation, as DG COMP considered whether DoubleClick’s unbundled services potentially restrained Google’s bundled offerings,⁴¹ indirectly vectoring the possibility into the analysis. DG COMP also considered whether Google could mature its nascent services without acquiring DoubleClick but found this to be implausible, as Google lacked the experience.⁴² After establishing the parties’ market position, DG COMP could assess the transaction’s possible impact on competition under the available theories of harm.

1. DoubleClick might have restricted Google, making its elimination as an independent force problematic. DG COMP⁴³ rebutted this based upon Google and DoubleClick not being close substitutes, the existence of high switching, and the availability of strong alternatives to DoubleClick that could replace it as a competitive counterforce.
2. DoubleClick could be a potential competitor if it matured its exchange and moved into intermediation services, making its elimination problematic. DG COMP⁴⁴ rebutted this. Besides having strong alternatives, including Microsoft, AOL, and Yahoo!, publishers multi-homed, relying on several providers.
3. The acquisition of DoubleClick could eliminate Google’s interest in maturing new ad-serving services, including unbundling its current offerings. DG COMP⁴⁵ rebutted this, as Google lacked the experience to develop current offerings and would be confronted with strong alternatives anyway.

DG COMP also evaluated whether the transaction could create non-horizontal concerns, originating in Google’s offering intermediator and ad-serving services as a bundle. By contrast, DoubleClick offered the latter on a stand-alone basis, and by denying or downgrading these thereafter, Google could potentially⁴⁶ leverage its position into vertical markets. Alternatively, it could gain portfolio power that could be used to leverage the position into adjacent markets, most significantly between search- and non-search online advertisements. However, to make these concerns plausible, both ability and inventiveness would have to be present, and eliminating a customer would, impede sales and might not be profitable. This was not lost on DG COMP,⁴⁷ making it a potential mitigating factor when reviewing the vertical aspect of transactions. DG COMP’s concerns, therefore, involved whether:

1. Google would leverage DoubleClick’s position in ad-serving, i.e., by degrading the quality when used by competitors. However,⁴⁸confronted with strong alternatives and users’ multi-homing, Google lacked ability and incentive. It wasn’t even evident that any such policy would have material effects, due to low entry barriers and the broad utilisation of internal ad-serving services.
2. By bundling search and non-search advertising, Google could leverage its position in the former into the latter. However, DG COMP⁴⁹rejected this for lack of incentive, due to low margins and alternatives available to customers. Strong competitors also made it implausible that it would produce anti-competitive effects.
3. Combining Google and DoubleClick’s data would create a non-replicable position that would marginalise the competitors progressively through superior offerings. However,⁵⁰ DG COMP found this implausible, as DoubleClick’s customer contract did not allow data consolidation, and attempts at modifying this thereafter would probably be met with pushback.

As the transaction yielded neither horizontal nor non-horizontal concerns, DG COMP granted an unconditional clearance in March 2008. The conclusion was that the parties operated in two separate parts of the Ad Tech sector, with only limited and indirect overlaps, and that Google would be met with strong competition from, e.g., Microsoft. With the benefit of hindsight, there are two potential flaws in DG COMP justification. Firstly, contrary to DG COMP expectations, Microsoft could not successfully integrate its own ad-tech acquisitions (aQuantive/Atlas and AdECN), otherwise considered equal to DoubleClick in terms of market positions.⁵¹ The strong competition from, e.g., Microsoft that DG COMP expected, never materialized itself, but it is doubtful that DG COMP could have predicted this in 2007. Secondly, the transaction might have vested Google with portfolio power, but this remains somewhat speculative. DoubleClick is only a minor element in Google’s larger ad tech ecosystem, and while it is evident that Google subsequently has been able to leverage its market power into new services, this probably speaks more to Google’s strong business case than to any specific contribution from DoubleClick. Regardless, DG COMP’s potential blind eye when it came to portfolio power merits comment as the critique of DG COMP (and other enforcers) primarily reached back to this. Before contemplating if DG COMP should have invested more time into the issue of portfolio power, a few considerations on portfolio power must be made.

Combining complementary services, as seen in large digital ecosystems, can create strong lock-in effects, where customers are unwilling, or unable, to shift to an alternative on a product-by-product basis.⁵² While competition is only a click away in theory, reality differs. Moreover, the owner of large digital ecosystems can leverage market power between the different services. This can be done either by relying on a classic bundling strategy or, more subtly, through micro nudging, where consumers are slowly pushed in the direction of other offerings of the ecosystem’s owner.

Today, enforcers understand these risks and strategies, but that cannot be said for the situation in 2007. Compounding the matter further, DG COMP had tried to move on a portfolio theory of harm in 2001 when it first blocked Tetra Laval’s acquisition of Sidel⁵³ and, later, in the same year, GE’s acquisition of Honeywell.⁵⁴ Both cases concerned a possible leverage risk associated with conglomerate mergers, but neither of the decisions survived juridical appeals. In 2003, the first was overturned⁵⁵ in its entirety, and in 2005, the same fate befell the latter⁵⁶ concerning the leverage risk. Granted, the portfolio effects outlined in these vintage cases are much weaker compared to today’s digital ecosystem, but once bitten, twice shy, explaining why DG COMP might have had second thoughts about trying its luck in 2007.

Secondly, DG COMP based its conclusions on several assumptions, for example that Microsoft and Yahoo! would represent mitigating factors and alternatives. While this later proved wrong, the assumption appeared reasonable in 2007 and governed DG COMP’s Decision (and probably also the FTC’s). After losing the bidding war over DoubleClick, Microsoft had acquired aQuantive/Atlas and, before this, AdECN. Both were seen as direct (and strong) competitors to DoubleClick in the DG COMP Decision,⁵⁷ and while it remains unclear why this competition did not materialise, apparently Microsoft failed to integrate its acquisitions into its business.

On balance, the risk of portfolio power might have been underestimated. Google has effectively integrated DoubleClick into a more extensive advertising ecosystem, leveraging market power between offerings––initially, between search advertising and non-search advertising, now dominating both, but potentially also within other sectors of the Ad Tech Industry.⁵⁸ In DG COMP’s defense, this understanding mostly follows from what happened later, including the way in which Google developed its ecosystem. The latter is particularly notable, as this means Google’s current ability to exercise market power did not necessarily follow from its acquisition of DoubleClick but rather was the compounded effect of separate but complementary offerings. It should also be recalled that the Courts had overturned DG COMP’s attempt to rein in portfolio power in the early 2000s, making it understandable why DG COMP might have had a limited appetite for a rematch.