What’s the purpose of the roadmap paper?
When these very big antitrust investigations are being undertaken by the government, they are confidential, because the government’s asking lots of questions about the way businesses run and asking for data that’s not public and strategic thinking that’s not public. A big investigation like this could last for years. And that means that the people who are the consumers affected and the taxpayers affected don’t have any knowledge of what’s happening. They don’t have a sense of what the story is. They don’t have a sense of what the government’s theory of harm is and what markets were disrupted and why.
That’s all completely normal. That’s how these investigations work. But I think it can be healthy for discussions in society about how we want to run our country and how we want to regulate different sectors or apply the law if more people understand what’s going on. So the paper’s really an attempt to share what I think the most plausible theories of harm are with the broader public, and people particularly interested in policy.
What are the roles that Google plays in the online advertising business?
They do everything. They own the tool that the advertiser uses, they own the intermediaries that handle bidding and sending bids to networks, selecting the winning bid. Then getting that creative content to the publisher, posted on the publisher’s site, measuring what happened, and paying everybody.
That same chain of tools would be used if the advertiser were advertising on a Google site—on YouTube, for example—or on search or on one of Google’s many other properties. So Google is both the channel through which the vast majority of search advertising is run as well as the destination for a lot of it.
There are what are called walled-garden destinations like Facebook and Amazon. Facebook has a bunch of pages and they also sell the space on those pages. You go to them, you pay whatever money they ask and you get your ad on a Facebook page. That’s an enclosed space where Facebook knows who the consumer is, knows what they like, who their friends are, lots of information. Similarly, Amazon knows everything I’ve bought.
But if an advertiser wants to go to what we call the open web—that is to say the many, many thousands of websites where people might go to read about their favorite movies or how to braid their hair more effectively or grilling tips or whatever—they are not a closed, walled garden. And that’s the place where Google has these tools to place ads.
Now, one issue is that those parties don’t have any place else to go for ad placement tools. Vertically integrated sellers of display ads are not generally competing to place ads outside their properties. Further, Google now has 90% of the tools used by advertisers and publishers and more than 50% of the platforms that buy and sell. This leaves both advertisers and publishers on the open web with very little choice of competitors that can monetize at the same level as Google. Why can’t they monetize? Because Google’s conduct has excluded them to the point where they don’t have the scale and don’t have the data. Large amounts of personal data allow Google to target its ads and that means they can easily outbid the low price bid by small rivals. More competition would tend to raise those prices. So that’s one issue: the publishers have less choice when you have concentration and market power.
The second issue is that the publishers like the New York Times or the Wall Street Journal are actually competing with Google. Their sites compete with Google’s properties. Think about news or flights or shopping—any of the many things where Google has a place the consumer could go and so do other companies. If I’m Yelp and I have to use this set of tools to get advertising onto my site, I might be quite concerned that Google in its control of those tools, is disadvantaging me vis-a-vis their own properties. They’d like the best ads at the highest prices to go to their own properties and not to their rivals.
How did Google get to be so dominant?
What we describe in the paper is that Google has purchased many participants in this marketplace over the years and then has used compatibility, technological and contractual compatibility, to steer and coerce advertisers and publishers to use Google’s tools. For example, you have to use a Google demand-side platform if you wish to place an ad on YouTube. There just isn’t a way to use a competitor if you want to place your ad on YouTube, which is owned by Google.
Does this issue trickle down to us as consumers? Are we losing out in some way?
Yes, it does. When the price of ads is higher, some of that is passed through into the price of goods and services. If acquiring a customer costs me a dollar instead of 50 cents, then I, as the manufacturer of the product, am going to have a higher cost and I’m going to pass that through to consumers, at least in part.
I think almost a more important harm is when an entity is thinking about setting up an ad-funded business on the internet. Say I’m going to open up a site that reviews movies and I’m going to fund that with ads. People will come to my site and read about movies and I’ll show them ads, and that will be a business for me. If I receive a lower price for those ads than I would in a competitive market that’s a very substantial revenue impact for that business. Fewer of those businesses will start to begin with, they won’t invest, they won’t expand, because they’re not actually earning the money that consumers would like them to have, in the sense that the consumer values that content so that’s why she’s on the site looking at the ad.
Does Google’s dominance in advertising flow from its dominance in search?
It does. In the paper we start at the point where Google has dominance in search, and there’s a whole other set of issues about where that came from. But certainly Google takes properties that it has at the time, acquires more businesses in ad tech, and uses those to increase its market power.
The market power in search—an almost virtual monopoly in search—can be used to require advertisers to employ other Google tools. For example, if you want to have a display ad campaign and a search ad campaign that are coordinated, Google can require use of its tools to design on a search campaign, and then that same tool also plans the display campaign. Or if an advertiser wants to look at the data—who clicked, who bought, try to analyze rates of return and so on—that data and that analysis can only be done inside the Google Data Hub and if you want to combine with all the other data that come out of your advertising campaign, then you have to use the Google Hub. Once you’re in the Google Hub with that data, you then are only able to do certain things, like your analytics and planning the next campaign, if you do it with Google tools using Google services.
You mentioned in the paper that Google is going to be blocking third-party cookies in its Chrome browser, which is the biggest browser in the world. This has been presented as a victory for privacy, but you suggest actually it’s another way for Google to shut out other people who want to do digital advertising.
If your definition of privacy is that you don’t want to see targeted ads, then elimination of the third-party cookie is not going to do that. What the elimination of third-party cookie means is that Google is the only one who can develop a super-profile on everybody, based on everything you do on the web. They can see your activities if you’re logged into Chrome; they can read your Gmail; they can look at what vacations you’re shopping for on Google Flights. If you’ve got an Android phone, they have location information, pressure, how fast you are traveling, what kind of conveyance you are likely in, and so forth.
They have tracked, collected, and bought an enormous profile about each person on the web and that profile, if it belongs to a monopolist, means that digital advertising is only effective if done through the monopolist who holds all the data.
We need to explain that the information about you can be passed among 25 different ad tech companies or it can be passed among 25 parts of Google. Which one of those is better or safer? That’s hard to know. I know which one of them is cheaper—the one where there’s competition. And it’s also better for consumers in the sense of stimulating the content that they want to consume on the internet.
Suppose that I start up a dog-walking service and I want to advertise it. If I know who owns a dog, then I can advertise just to those people. I am willing to pay more and I have to send fewer ads, and they’re useful ads to the people who receive them. The party that holds the information about who has a dog is able to extract all of that surplus. So there’s an intermediary with a complete knowledge of the information that is the bottleneck and can extract all the profit, which reduces the ability of nascent businesses to start and grow.
So the question then is, what do consumers want? Do they want no targeted advertising? Well, then you can’t have an ad tech industry. Do they want targeted advertising that benefits the business that’s sending it, like the dog walker or the New York Times? That would require competition in the ad tech market.
The federal governments and state governments are apparently working on an antitrust case against Google. According to the New York Times, they’re focusing primarily on the digital advertising business, but they’re also looking at search. What is the argument that Google is violating the law when it comes to its monopoly in search?
The search case was first pursued by the European Commission, so we have lots of information about that. When it first started down the path of monetizing its search engine, Google pursued quite a few exclusive contracts. To be the exclusive search engine on the iPhone, for example, on portals like AOL and on properties like News Corporation.
Google also, in the move to mobile, developed Android and included a lot of defaults and bundling in the Android contract so that their browser and search engine were on the front of the Android phone in a prominent place, and anything else would have to be installed by the user in a less prominent location. Google also had a contract called the anti-forking contract which prevented anybody who was currently selling an Android handset from making their own version of Android. Let’s say Samsung had wanted to make Samsung Android using the royalty free, open source Android base code. If they had done that with their own operating system, their own store, their own set of apps and so on, if they had sold a Samsung Android phone, they would have been out of compliance with the Google Android contract. They would have had to stop selling Google phones. That’s a very big business risk to say, “Oh, I’m going to start up one of my own, on my own operating system. But if that new product doesn’t work I will have no revenue because I cannot keep selling the Google phones.”
That contract likely blocked entry by those who understood how operating systems work and prevented competition in the operating system, which in turn was then the barrier to search and the browser.
If the federal government is in fact in the beginning stages of antitrust action against Google, what should we expect to see publicly?
They’ve said that they’re going to file a complaint in the summer, which is really quick. So that would be obviously something to watch. And then the states are going to presumably file a complaint also at some juncture, probably after that. So then we will see what those complaints say, what kind of conduct they allege as being anti-competitive. Or perhaps the federal government doesn’t file a complaint, but they settle. They say, “Well, here are the six things we didn’t like, but you’ve changed your behavior, and so therefore that’s remedied the lost competition and we’re not going to bring a case.”