Schlagwort-Archive: Advertising

Why Google gains competitive advantage over all other competitors in the Online Ad Market

Source: https://www.wired.com/story/google-ad-market-regulated-like-stock-market/

Should Google’s Ad Market Be Regulated Like the Stock Market?

A leading antitrust scholar says yes. Congress may be listening.
A cowboy readies a lasso for a giant chrome logo.
In a new paper, Dina Srinivasan argues that “Google dominates advertising markets by engaging in conduct that lawmakers prohibit in other electronic trading markets.” Illustration: WIRED Staff; Getty Images

The days of suit-clad men shouting out orders on the bustling floors of stock exchanges are mostly gone, replaced by windowless rooms full of servers, but the stock market is still a busy place. On the 13 US stock exchanges combined, around 50 million trades happen every day. And yet there’s another digital marketplace out there that processes tens of billions of transactions daily, one whose complexity makes the NASDAQ look like a lemonade stand: online advertising.

It may sound odd to refer to advertising as a market, but that’s what it is. The industry’s own terminology provides a hint: Publishers selling ad space, and advertisers buying it, do business on so-called “ad exchanges”; one of the biggest companies involved is called the Trading Desk. Whenever you load a web page, advertisers compete in an automated process called real-time bidding to show you their ad. Multiply that by billions of internet users around the world, loading many different pages and apps per day, and you can start to appreciate the scope. As antitrust scholar Dina Srinivasan puts it in a forthcoming paper, online advertising “is likely the most sophisticated of all electronic trading markets.” And yet, despite the market’s size and complexity—and unlike other markets—online advertising is almost completely unregulated.

A former digital advertising executive, Srinivasan gained attention last year for her paper “The Antitrust Case Against Facebook,” which laid out a novel theory of why Facebook’s market dominance can be bad for users even as it offers a free product. Now she aims to do something similar for Google—specifically, for the sprawling advertising empire that accounts for the vast majority of the company’s revenue. In her new paper, which will be published in the Stanford Technology Law Review, Srinivasan takes a deep dive into the inner workings of the digital ad market. The details are astoundingly complex, but the broad argument is straightforward. When you see an ad online, the odds are very high that the advertiser used Google to buy it, the website used Google to put the space up for sale, and Google’s exchange matched them together. In other words, Google both runs the largest exchange and competes as the biggest buyer and seller on that exchange. On top of that, it also owns YouTube, one of the biggest suppliers of ad inventory, meaning it competes against publishers on its own platform. And yet there are no laws governing any of it.

That regulatory vacuum, Srinivasan argues, has allowed Google to dominate the industry by doing things that are prohibited in other parts of the economy. “In the market for electronically traded equities, we require exchanges to provide traders with fair access to data and speed, we identify and manage intermediary conflicts of interest, and we require trading disclosures to help police the market,” she writes. Her proposal flows naturally from that observation: Apply those regulatory principles to digital advertising.

The resemblance between securities and ad markets first occurred to Srinivasan back in 2014. That’s when Michael Lewis published Flash Boys, which documented the extensive mischief created by high-frequency trading and other modern tricks of the digital securities market—and which helped spur a wave of investigations, fines, and regulatory action. At the time, Srinivasan saw similar issues arising in her own industry.

“When Flash Boys came out, it was comical. That book was being passed from executive to executive,” she said in an interview. “People would laugh about how there were operatives who were arbitraging between ad exchanges too. People were just laughing at the parallels.”

Over the past year, as she researched the paper, Srinivasan realized that the resemblance went even further than she thought, sometimes uncannily so. Lewis describes high-frequency traders seeking an edge by placing their computers as physically close as possible to the stock exchange servers to shave microseconds off trade times. Srinivasan relays a similar anecdote from the world of ad tech: Last year, OpenX, one of the largest non-Google ad tech companies, announced a five-year, $110 million deal to move its exchange to Google Cloud. OpenX was open about the fact that being on Google’s servers would give it a speed edge. “You have to operate at speed, efficiency, closeness to the publisher and the demand side of Google,” one executive said. It’s almost an exact copy of high-speed traders’ tactics. The difference, Srinivasan notes, is that “in financial markets, co-location practices are tightly regulated” to make sure everyone has equal access to speed. In advertising, they aren’t.

Speed is crucial in online advertising because the auctions occur in milliseconds. If an ad buying platform submits its bid too slowly, the exchange might exclude it from the auction entirely. This gives a leg up to a platform that shares infrastructure with the exchange—in other words, to Google. Google advertises this fact. “Since Google Ads and Display & Video 360 run on servers in the same data centers as Ad Exchange, they can respond faster to Ad Exchange bid requests compared to other exchange requests,” says a Google help page. “There are no network latency or timeout issues between either Google Ads or Display & Video 360 and Ad Exchange.” When the buying platform isn’t the same as the exchange, on the other hand, latency issues “can prevent buyers from successfully submitting a bid on up to 25% of bid requests.”

Srinivasan also explores the way Google benefits from unequal access to information. Modern digital advertising is all about being able to target users with the most precision. When someone arrives on a website using Google’s DoubleClick ad server, Google’s exchange “hashes” the ID, passing a different one along to the ad buying platforms. Those buyers then must match their ID with the hashed one to make sure they’re targeting the right person—a process called “cookie syncing.” But cookie syncing, Srinivasan writes, “is inherently inefficient.” Some percent of the time, the platform will fail to match the user. In those situations, she writes, advertisers aren’t willing to pay as much, or anything, because they aren’t guaranteed to reach the right audience.

Google doesn’t have this problem, because it allows its own exchange, and its own ad buying platform, to see the DoubleClick ID. That means it automatically knows who the user is. Google says it shares the DoubleClick ID only with its own platforms to protect user privacy. But another result is to put a thumb on the scale of Google’s own properties: If you want to make sure you’re targeting the right user, you have an extra incentive to buy ads using Google. Google advertises this advantage as well.

While securities law has its share of problems, it does broadly curtail the kind of flagrant information and speed imbalances that Srinivasan describes in the ad market. Indeed, the contrast between the digital advertising regulatory vacuum and the world of financial markets is striking.

“It’s a highly, highly regulated system,” said Kevin Haeberle, a professor at William & Mary Law School who specializes in securities law. Only registered brokers are allowed to execute trades, and those brokers must register with the Securities and Exchange Commission. “You’ve got to take tests, you’ve got to be registered, you have to be supervised in certain ways, you’ve got to pay into various insurance mechanisms to make sure the trades actually do settle.” He added, “There’s this whole regulatory regime, it’s very complex, and it applies to regulating these exchanges that run this important market for our society. In the ad market, we don’t have that.”

Why does that matter? At the broadest level, when one entity is allowed to both run a market and participate in it, and when there are no rules requiring it to let everyone else participate on equal terms, there’s nothing stopping it from enriching itself at the expense of the other buyers and sellers. In digital advertising, that means Google could be inflating prices advertisers pay, or depressing the amount of money publishers receive, or both. Google, of course, denies this characterization. It says its ad tools benefit both advertisers and publishers, no regulation necessary. To Srinivasan, believing that claim would be like trusting J.P.Morgan to run the New York Stock Exchange.

Srinivasan is particularly worried about the publishers who rely on digital advertising for revenue. “From a very big-picture perspective, we are a democracy and we want a healthy and robust economy of news,” Srinivasan said. “We want the news business as a sector in our economy, we want to make sure that it works. And so we should make sure that the market is not rigged for the middleman, so that entrepreneurs are encouraged to enter the business of news.” (In the paper, she discloses that she is “advising and consulting on antitrust matters, including for news publishers whose interests are in conflict with Google’s.”)

Her argument may be catching on. At the tech CEO hearing held by the House antitrust subcommittee in July, Pramila Jayapal, a Democratic congresswoman from Washington state, cited Srinivasan’s paper directly as part of her questioning of Google CEO Sundar Pichai.

“The problem is that Google controls all of these entities,” she said. “So it’s running the marketplace, it’s acting on the buy side, and it’s acting on the sell side at the same time, which is a major conflict of interest. It allows you to set rates very low as a buyer of ad space for newspapers, depriving them of their ad revenue, and then also to sell high to small businesses who are very dependent on advertising on your platform. It sounds a bit like the stock market. Except, unlike the stock market, there’s no regulation on your ad exchange market.”

In an interview after the hearing, Jayapal said she was looking into developing legislation that would address that regulation gap. She suggested that the underlying principles of any regulation would be straightforward. “It seems to me that the simplest thing to do is say, you can’t control the market and engage as a buyer and seller. Those two things have to be separated. And then, if you’re buying and selling, then you’re regulated by insider trading rules.” She added, “I think it’s just an unregulated marketplace that should be relatively easy to do something about.”

Lets Get Rid of the “Nothing to Hide, Nothing to Fear” Mentality

With Zuckerberg testifying to the US Congress over Facebook’s data privacy and the implementation of GDPR fast approaching, the debate around data ownership has suddenly burst into the public psyche. Collecting user data to serve targeted advertising in a free platform is one thing, harvesting the social graphs of people interacting with apps and using it to sway an election is somewhat worse.

Suffice to say that neither of the above compare to the indiscriminate collection of ordinary civilians’ data on behalf of governments every day.

In 2013, Edward Snowden blew the whistle on the systematic US spy program he helped to architect. Perhaps the largest revelation to come out of the trove of documents he released were the details of PRISM, an NSA program that collects internet communications data from US telecommunications companies like Microsoft, Yahoo, Google, Facebook and Apple. The data collected included audio and video chat logs, photographs, emails, documents and connection logs of anyone using the services of 9 leading US internet companies. PRISM benefited from changes to FISA that allowed warrantless domestic surveillance of any target without the need for probable cause. Bill Binney, former US intelligence official, explains how, for instances where corporate control wasn’t achievable, the NSA enticed third party countries to clandestinely tap internet communication lines on the internet backbone via the RAMPART-A program.What this means is that the NSA was able to assemble near complete dossiers of all web activity carried out by anyone using the internet.

But this is just in the US right?, policies like this wouldn’t be implemented in Europe.

Wrong unfortunately.

GCHQ, the UK’s intelligence agency allegedly collects considerably more metadata than the NSA. Under Tempora, GCHQ can intercept all internet communications from submarine fibre optic cables and store the information for 30 days at the Bude facility in Cornwall. This includes complete web histories, the contents of all emails and facebook entires and given that more than 25% of all internet communications flow through these cables, the implications are astronomical. Elsewhere, JTRIG, a unit of GCHQ have intercepted private facebook pictures, changed the results of online polls and spoofed websites in real time. A lot of these techniques have been made possible by the 2016 Investigatory Powers Act which Snowden describes as the most “extreme surveillance in the history of western democracy”.

But despite all this, the age old reprise; “if you’ve got nothing to hide, you’ve got nothing to fear” often rings out in debates over privacy.

Indeed, the idea is so pervasive that politicians often lean on the phrase to justify ever more draconian methods of surveillance. Yes, they draw upon the selfsame rhetoric of Joseph Goebbels, propaganda minister for the Nazi regime.

In drafting legislation for the the Investigatory Powers Act, May said that such extremes were necessary to ensure “no area of cyberspace becomes a haven for those who seek to harm us, to plot, poison minds and peddle hatred under the radar”.

When levelled against the fear of terrorism and death, its easy to see how people passively accept ever greater levels of surveillance. Indeed, Naomi Klein writes extensively in Shock Doctrine how the fear of external threats can be used as a smokescreen to implement ever more invasive policy. But indiscriminate mass surveillance should never be blindly accepted, privacy should and always will be a social norm, despite what Mark Zuckerberg said in 2010. Although I’m sure he may have a different answer now.

So you just read emails and look at cat memes online, why would you care about privacy?

In the same way we’re able to close our living room curtains and be alone and unmonitored, we should be able to explore our identities online un-impinged. Its a well rehearsed idea that nowadays we’re more honest to our web browsers than we are to each other but what happens when you become cognisant that everything you do online is intercepted and catalogued? As with CCTV, when we know we’re being watched, we alter our behaviour in line with whats expected.

As soon as this happens online, the liberating quality provided by the anonymity of the internet is lost. Your thinking aligns with the status quo and we lose the boundless ability of the internet to search and develop our identities. No progress can be made when everyone thinks the same way. Difference of opinion fuels innovation.

This draws obvious comparisons with Bentham’s Panopticon, a prison blueprint for enforcing control from within. The basic setup is as follows; there is a central guard tower surrounded by cells. In the cells are prisoners. The tower shines bright light so that the watchman can see each inmate silhouetted in their cell but the prisoners cannot see the watchman. The prisoners must assume they could be observed at any point and therefore act accordingly. In literature, the common comparison is Orwell’s 1984 where omnipresent government surveillance enforces control and distorts reality. With revelations about surveillance states, the relevance of these metaphors are plain to see.

In reality, theres actually a lot more at stake here.

With the Panopticon certain individuals are watched, in 1984 everyone is watched. On the modern internet, every person, irrespective of the threat they pose, is not only watched but their information is stored and archived for analysis.

Kafka’s The Trial, in which a bureaucracy uses citizens information to make decisions about them, but denies them the ability to participate in how their information is used, therefore seems a more apt comparison. The issue here is that corporations, more so, states have been allowed to comb our data and make decisions that affect us without our consent.

Maybe, as a member of a western democracy, you don’t think this matters. But what if you’re a member of a minority group in an oppressive regime? What if you’re arrested because a computer algorithm cant separate humour from intent to harm?

On the other hand, maybe you trust the intentions of your government, but how much faith do you have in them to keep your data private? The recent hack of the SEC shows that even government systems aren’t safe from attackers. When a business database is breached, maybe your credit card details become public, when a government database that has aggregated millions of data points on every aspect of your online life is hacked, you’ve lost all control of your ability to selectively reveal yourself to the world. Just as Lyndon Johnson sought to control physical clouds, he who controls the modern cloud, will rule the world.

Perhaps you think that even this doesn’t matter, if it allows the government to protect us from those that intend to cause harm then its worth the loss of privacy. The trouble with indiscriminate surveillance is that with so much data you see everything but paradoxically, still know nothing.

Intelligence is the strategic collection of pertinent facts, bulk data collection cannot therefore be intelligent. As Bill Binney puts it “bulk data kills people” because technicians are so overwhelmed that they cant isolate whats useful. Data collection as it is can only focus on retribution rather than reduction.

Granted, GDPR is a big step forward for individual consent but will it stop corporations handing over your data to the government? Depending on how cynical you are, you might think that GDPR is just a tool to clean up and create more reliable deterministic data anyway. The nothing to hide, nothing to fear mentality renders us passive supplicants in the removal of our civil liberties. We should be thinking about how we relate to one another and to our Governments and how much power we want to have in that relationship.

To paraphrase Edward Snowden, saying you don’t care about privacy because you’ve got nothing to hide is analogous to saying you don’t care about freedom of speech because you have nothing to say.

http://behindthebrowser.space/index.php/2018/04/22/nothing-to-fear-nothing-to-hide/