The Goliath versus Goliath idea was recently given some heft by a 2018 McKinsey Global Institute analysis of nearly six thousand of the world’s largest public and private companies, each with annual revenues greater than $1 billion and which together make up 65 percent of global corporate pre-tax earnings. Among this group, the top 10 percent (the “superstar” companies) take 80 percent of economic profits—defined as a company’s invested capital multiplied by its returns above the cost of that capital. The top 1 percent alone take 36 percent of the pie.32 We already know who some of the top 10 percent are—they include high-margin Big Tech companies (Facebook, Apple, Amazon, and Google), as well as a number of others who have been able to exploit the value of intangible assets such as software, data, patents, and brands (these would include not just technology firms but also a good number of financial, biotech, and pharmaceutical companies). We also know that the network effect allows such companies to capture markets quickly and at scale, giving them what’s known in the start-up world as the “first scaler advantages.”
This process is aided by the fact that we have shifted from a “tangible” economy, based on physical goods, to one based more on intangibles—namely intellectual property, ideas, and data. British academics Jonathan Haskel and Stian Westlake, who lay out the case for this in their excellent book, Capitalism Without Capital, believe that this shift upends the usual rules of economic gravity. Google and Facebook don’t need to build more factories, invest in more raw materials, or staff more assembly lines in order to capture market share, which is why they have the ability to grow much faster than corporate giants of the past. In today’s economy, the losers tend to own more things—tangible assets such as factories and equipment—whereas the winners are concerned with leveraging intangible ones.
The network effect is at the center of this shift. Whether it’s made up of Twitter users, Uber drivers, Airbnb hosts, or Instagram influencers, the network is worth far more than the value of any single node within it. The key point is that users beget users, which allows the players who can grab the most market share quickly to dominate entire industries seemingly overnight. This is not unique to Google, as we’ve seen. But it is much easier these days to grab market share if you are big, and can leverage data and intellectual property across networks. These intangible assets can scale far, far faster and further than the products and services of old.33 Networked businesses are case studies in how what goes big, goes bigger still.
But as Varian and Carl Shapiro acknowledge, there’s a dark side to the feedback loop: “Positive feedback [within platforms] also makes the weak get weaker.” In other words, even superstars of the networked era can fall, though typically to each other, rather than to upstarts.34
* * *
—
JUST LOOK AT the car industry for a sense of how profound the shift will be for traditional firms. Currently, about 90 percent of the value of an automobile lives in the hardware. But as autonomous driving and digital apps become a bigger deal, that ratio is expected to shift dramatically. Morgan Stanley predicts that in autonomous vehicles, 40 percent of the value of an automobile will come from hardware, 40 percent from software, and 20 percent from the content that streams into the vehicle.35 That would include things like games, advertisements, and news enabled by the software. This shift is partly driven—no pun intended—by the fact that millennials want their cars souped-up with all their favorite apps. But it also reflects another idea: When you are in an autonomous vehicle, brand identity disappears.
“If you take away control of the steering wheel, consumers are much less likely to care what type of car they are in,” says Nick Johnson, principal at the consultancy Applico, who has advised major automakers on the shift. Johnson is also the author of Modern Monopolies, which looks at the effects that the Silicon Valley giants are having on other firms and industries. In this world, a car is no longer something you touch and feel or “wear” like a luxurious garment, but something you use—like a phone. And if that is the case, it is the software and apps that develop around the software platform that really matter, not the plastic and metal shell they live in, as phone makers Nokia and BlackBerry can attest. Indeed, a recent survey by carmaker BMW revealed that 73 percent of people said they would exchange one brand of car for another if they could bring their digital lives into the new car.
This is the challenge facing GM, a company that was attacked in 2018 by Donald Trump and labor officials alike for laying off or buying out 14,300 autoworkers (part of that as a result of shuttering five factories in the United States and Canada). Both the U.S. president and the unions focused on arguments about sending jobs to China and Mexico. But the biggest challenge GM is facing isn’t one of labor costs, outsourcing, or steel tariffs. It is the question of whether it will be able to continue to own a large share of the economic value of the automobile industry in the networked era in which the car is becoming a smart device. It’s a question that faces any number of industries, from retail (which has already been decimated by Amazon) to healthcare (under competitive threat from both Amazon and Google), finance (which is under threat from fintech, the merging of tech platform technology and banking), manufacturing, and so on.
The businesses that have the best technology in automotive software and apps are, unsurprisingly, technology companies such as Google, Apple, and China’s Baidu, all of which are pouring money into autonomous vehicle technology and platforms to support it. Right now, motorists can mainly stream music, GPS information, and whatever other data they can access on their phone via such systems. But once the platforms are embedded more deeply in vehicles, customers will be able to tap into everything from fluid levels and engine temperatures to safety information. Those are all currently the domain of carmakers. Monetizing all that data, via new products and services, is the big prize.
How should companies think about competing in this world? The case study on what not to do comes from Nokia. Remember the once-mighty Finnish phone maker? You may have used one of its brick-like handsets to type some of your first text messages back in the 1990s. Then Apple’s iPhone came along, followed by Google’s Android operating system. Both companies offered not just snazzy products but successful platforms for developers. Ecosystems of apps grew around them, while Nokia’s Symbian operating system became, by comparison, hopelessly passé. By 2011, Nokia was in free fall, never to recover.36
As Symbian head Jo Harlow told the Financial Times at the time, the company simply hadn’t been quick enough to make the shift from being “device-led to software-led.”37 True enough. But the larger problem was that, in a deeper way, Nokia, like many before and after, failed to see that not only would the vast majority of value migrate from hardware to software, but specifically to the platform on which that software would operate. The network effect of developers and users around those platforms is what would create value—much more so than the product itself.
Carmakers are not standing still. GM’s chief executive Mary Barra has for years referred to GM as a technology company, one increasingly dependent more on data than on steel or manpower. In 2018, she explicitly tied the big GM job cuts to a shift in resources toward electric cars and autonomous vehicle development. There are some nascent industry partnerships, such as Ford’s SmartDeviceLink Consortium, an open source community working via a standard set of protocols to connect smartphone applications to vehicles. But no big carmaker has shown itself able or willing to create the platform ecosystems that big technology companies create. This is a problem, because the network effect really kicks in when a company controls 30 or 40 percent of a given market, which means that the major auto companies of the world would need to team up in order to achieve such a share. Certainly, it would be a big shift for a company to think about its most aggressive competitors as collaborators. Yet it may be the only choice they have. Developing an ecosystem and owning the software and data within it will be the key t
o success not just in the car business, but in many industries.
Neoliberalism on Steroids
As powerful as the network effect is, to understand the seemingly unstoppable growth of the platform companies like Google or Facebook, you also have to look at how much the politics of Silicon Valley changed between the era of hippie idealism represented by Steve Jobs, and the libertarian epoch of Peter Thiel and his ilk. “It was a titanic shift,” says Roger McNamee, who has worked in tech for more than forty years. “While the rank and file in Silicon Valley is liberal, the top people at the top firms tend to believe that greed is good.”
How could they not? Ever since the 1980s, most of American business has been subscribing to the trickle-down “markets know best” doctrine popularized by the so-called Chicago School of economics. The Internet platforms in particular have benefited enormously from the Chicago School’s antitrust philosophy, which maintains that as long as products are cheap or free, there’s no monopoly issue. As McNamee outlines in his own book, Zucked, “Google leveraged its dominant market position in search to build giant businesses in email, photos, maps, videos, productivity applications, and a variety of other apps. In most cases, Google was able to transfer the benefits of monopoly power from an existing business to a nascent one.” When you go back to the economic history books, this should come as no surprise—monopoly power was a central feature, even an aim, of those who invented the economics of information technology.
How the Big Get Bigger
The neoliberal politics of Silicon Valley are reflected in the work of the economist Hal Varian, who joined Google as a consultant in 2002. Eric Schmidt had run into Varian back in 2001 at the Aspen Institute, one of those places where tech titans and their admirers gather to discuss Big Ideas. Schmidt informed Varian that the company had an auction model that “might make a little money” and asked if he would come and help the company perfect it.38 Varian, who’d been dean of the Berkeley School of Information, was one of the top economists studying data markets at the time. He had cowritten an influential book entitled Information Rules: A Strategic Guide to the Network Economy, and would eventually have a rule named after himself—a sort of trickle-down theory for the digital age. The Varian rule posited, incorrectly, that everything rich people had today, the middle classes and eventually the working classes would have tomorrow, thanks to the price-crunching effects of technology. (Big Tech critic Evgeny Morozov later rephrased it in perhaps a more factually accurate way: “Luxury is already here, it’s just not very evenly distributed.”)
Right around that time, he gave a series of lectures and wrote a number of papers that laid out some of the key ideas emerging from the burgeoning field of data economics, ideas that make it hard to believe that the people at the top of today’s platform technology firms didn’t understand the far-reaching and potentially disturbing effects their innovations might have on our economy, our politics, and our society.
Varian, like most economists, believed in Chicago School theory, and layered ideas about the network effect and the power of big data on top of that intellectual framework. He understood that companies who can harness the network effect “have significant market power,” particularly since the data they acquire “allows for fine-grained observation and analysis of consumer behavior.”39 As such companies build scale, they acquire a kind of monopoly power based on that relationship. As Varian puts it, “An extended relationship allows the seller to understand ‘their’ consumers’ purchasing habits and needs better than potential competitors. Amazon’s personalized recommendation service works well for me, since I have bought books there in the past. A new seller would not have this extensive experience with my purchase history, and would therefore offer me inferior service.”40 (Particularly if that seller can’t get a leg up on the dominant platform, as was the case in the Google-Foundem conflict.)
Varian eventually became chief economist at Google, where he quickly built his reputation as a practical diviner of this new economics of information. He hired an entire team of “econometricians,” who combined neoliberal theory, mathematical ideas, and data to help Google make as much money as possible, then went to work helping Page, Brin, and Schmidt develop more efficient auction algorithms and build the auction model that became such a gold mine for Google.
One of his tasks was to analyze the signal in the noise of all the data that Google was gathering. He brought the new efficiencies of data economics to resource allocation at the company itself, developing an auction model that calculated and allocated internal computing power as sharply as any Wall Street trading scheme. (The paper that came out of that experience was titled “Using a Market Economy to Provision Compute Resources Across Planet-wide Clusters.”)41 Predictably, his theories of the new data economics tended to favor his employer. As Shoshana Zuboff has written, in the sort of surveillance capitalism practiced by Google and other Big Tech firms, “Contract and the rule of law are supplanted by the rewards and punishments of a new kind of invisible hand,”42 the algorithmic hand of Silicon Valley.
Varian and his team were unique, and foreshadowed an era in which most big companies would hire data scientists and data economists in great numbers. The existing laws that governed commerce were, like most laws in the view of Big Tech, made to be broken.
Stewards of Trust?
To be fair, pioneers like Varian have acknowledged a number of downsides of this new networked business model being pursued by Google and numerous other Silicon Valley giants, even the big one: privacy. Amazingly, in 2011, he admitted that, as a user, he would not want the platforms to share personal information with third parties without his consent.43 He concludes, however, that this problem isn’t much of a risk; the sale of information to third parties without consumer consent wouldn’t be economically efficient, since it would breach trust.
It was a position only slightly more self-aware than that of his boss, Eric Schmidt, when asked, in a 2009 CNBC documentary, Inside the Mind of Google, about whether people should trust Google with their most personal secrets. Schmidt replied, “I think judgment matters. If you have something that you don’t want anyone to know, maybe you shouldn’t be doing it in the first place.” Translation: Your privacy isn’t our problem.
According to Nobel Prize–winning economist Paul Romer, much of our willingness to trade our right to privacy for the sleekest new iPhone model has to do with the fact that “there are tremendous information asymmetries in these markets. Do both parties understand enough to know whether the transaction taking place is in their mutual interest?” he asked, rhetorically. Romer (like me) would argue that they do not; he believes that the complexity of today’s data markets “means that notions like ‘consent’ [to long and complex disclosures about how your data might be used by platform companies] have become meaningless.” The differences in what either party knows simply undermine the fair functioning of the market itself. “I’ve been in these discussions with people like Hal Varian, and I get more and more frustrated,” he told me in 2018, shortly after winning the Nobel. “There’s a dishonesty about giving people something that’s eighteen thousand words long and expecting them to read and understand it,” Romer says.
But what’s the solution? For starters, says Romer, we should stop using the word privacy. “It doesn’t really exist anymore,” he says. We should focus more on transparency and clarity. “If nobody—let’s call that fewer than five percent of users—can get an even partial understanding” of the terms of a transaction, then Romer says companies simply shouldn’t be doing them. What’s more, “we should put the burden of proof on the companies themselves,” rather than allowing them to circumvent responsibility via “phony disclosures.”44
Some companies, such as Apple and even IBM, which is still very much a key player in the technology world, are finally waking up to the idea that protecting user privacy is a competitive advantage.45 Apple, for example, has rolled o
ut a new website to better showcase privacy features that it believes differentiate the company from its competitors, including an algorithmic function whereby search data is stored within individual devices rather than in the “cloud,” thus giving users more control over what the company can see.
Apple is also touting a technique known as “differential privacy,” which allows the company to gain insights into what users are doing, while preserving a certain amount of privacy by encrypting the data before it leaves a user’s device, in such a way that Apple can’t associate the data it receives with any particular user. The data is used to improve the devices and services that are sold within the Apple ecosystem, rather than to send customers hyper-targeted ads from other businesses that they had no idea were getting their data to begin with. That is, again, quite a departure from the Google/Facebook approach. Does this solve all the problems? No. But on the other hand, Apple’s business model doesn’t lend itself to influencing an election like Russia attempted to do in the United States via Facebook. It’s also refreshing to hear Tim Cook say that he believes “privacy is a fundamental human right.”
Ginni Rometty, IBM’s chief executive, has also announced a new set of principles and practices around data aimed at increasing trust in Big Tech. These include a pledge not to keep any proprietary data in their servers for more than a specified contract period, never to turn over client data to any government surveillance program in any country, and a promise that clients will own not only the rights to their end data, but to any algorithmic “learning” from it.
Don't Be Evil Page 17
