Inside Story

Current affairs & culture from Australia and beyond

2619 words

Requiem for the World Wide Web

9 January 2019

Books | Matthew Hindman offers illumination for a disillusioned age

Right:

Centralised diversity: Facebook’s Mark Zuckerberg testifying on Capitol Hill in Washington last April. Pablo Martinez Monsivais/AP Photo

Centralised diversity: Facebook’s Mark Zuckerberg testifying on Capitol Hill in Washington last April. Pablo Martinez Monsivais/AP Photo

The Internet Trap: How the Digital Economy Builds Monopolies and Undermines Democracies
By Matthew Hindman | Princeton University Press | $29.95  | 256 pages


When Nicholas Negroponte coined the term “the daily me” in his 1995 book, Being Digital, he proclaimed that “the monolithic empires of mass media are dissolving into an array of cottage industries.” The media barons of today, he said, “will be grasping to hold onto their centralised empires tomorrow.” The web’s decentralised character was foundational to its very conception; as Tim Berners-Lee had declared on one of the World Wide Web’s first pages, there would be “no top to the web.” In its first major decision on regulation of internet content in 1997, the US Supreme Court predicted the web would turn each of us into a “town crier with a voice that resonates farther than it could from any soapbox.”

And yet here we are. On the one hand, the power of the old media monopolies is as malign as ever; on the other, the new monopolies have introduced us to fresh iniquities. Twenty-eighteen saw Mark Zuckerberg appear before the US Congress to apologise for not doing enough to stop fake news, foreign interference in elections, hate speech, and the unfortunate matter of a third party accessing the private data of eighty-seven million people. Back among the old media, last year’s highlight was the ousting of Malcolm Turnbull, assisted by a sustained campaign from the Murdoch tabloids, Sky News and the Australian. Reports soon emerged that Murdoch himself, after arriving in town eleven days earlier, had declared “Malcolm has got to go,” and that the besieged prime minister had desperately phoned the mogul to plead his case. As Kevin Rudd wrote in August, Murdoch remains “the greatest cancer on Australian democracy.”

Rather than giving birth to a plethora of cottage industries, the internet has turned out to be a monopolist’s paradise. Whether it’s Facebook in social media, Google in search, Amazon in online shopping, Ebay in online auctions or Netflix in online streaming, immensely powerful incumbents dominate market niches and competition takes place on anything but a level playing field. Every month, as the Australian Competition and Consumer Commission reported in December, “approximately nineteen million Australians use Google Search, seventeen million access Facebook, seventeen million watch YouTube (which is owned by Google) and eleven million access Instagram (which is owned by Facebook).” Around 94 per cent of online searches are made through Google, and the company enjoys corresponding market power in search advertising. Google and Facebook alone earn well over half of the $8 billion spent advertising online in Australia.

Matthew Hindman’s achievement in The Internet Trap is to advance a thoroughgoing explanation of why the internet has so comprehensively disappointed the early hopes and expectations — and why those who plead that it just needs more time are misguided. The internet’s tendency to monopoly is not an aberration, Hindman argues; it is a product of the basic economic forces that are permanent features of digital life.

Hindman takes as his starting point the apparently solid basis of our erstwhile optimism: where print, radio and TV required massive investment in capital equipment, the internet reduced the cost of reproducing and distributing information to close to zero. As the American communications scholar Clay Shirky reasonably asked, “What happens when there is nothing unique about publishing anymore, because users can do it for themselves”? The blogosphere looked like the answer. In 2005 it was estimated that thirty-five million Americans were reading blogs, and that an even greater number of blogs were in existence.

To be sure, it remained costly to produce high-quality content, but the internet appeared to be breaking down even that barrier to entry. As Yochai Benkler explained in The Wealth of Networks, the web enabled widely dispersed voluntary effort to be coordinated efficiently to create content as compelling as it is popular. Wikipedia wasn’t just outcompeting Britannica; studies were finding it similarly reliable. As Clay Shirky pronounced in the title of his 2008 book, “here comes everybody.”

But this kind of thinking rested on a basic conceptual mistake. It conflated speaking with being heard. Yes, it’s cheap as chips to start a blog. But gaining an audience? That costs. Digital survival depends on what Hindman calls stickiness — a “firm’s ability to attract users, to get them to stay longer, and to make them return again and again.” Once we think about the cost of distribution in these terms it becomes apparent that it is not only really, really expensive, but it also exhibits economies of scale no less steep than those of print media. “Critically, many tactics that promote stickiness get cheaper per user as sites get bigger,” says Hindman. And much of The Internet Trap is devoted to explaining why the cost per unit decreases in cyberspace as the scale of production increases ad infinitum, inevitably resulting in monopolies.

Some of the advantages of scale in the digital economy are already generally appreciated. Like the telephone, a social network becomes more valuable the more widely it is adopted. The best feature of Facebook is that everybody is there, which means that any would-be competitor has to do more than simply come up with a better product. It has to persuade potential customers to join a network where they risk not finding the people they want to connect to.

Other economies of scale documented by Hindman are less widely understood. He tells how, in early 2000, Google conducted one of its first live online experiments to find out the optimal number of results users preferred in response to search queries. Users were randomly assigned to receive twenty, twenty-five or thirty results. What Google discovered was that traffic declined precipitously among the groups with more results. “Searches in the thirty-result group had fallen by more than 20 per cent, and tens of thousands of users had abandoned Google altogether…” The ultimate diagnosis was that more results slowed response time — only by half a second, mind you, but enough for users to get frustrated and go elsewhere. These findings, widely corroborated, underscore how speed is one of the fundamental building blocks of stickiness.

While speed derives from a very clever algorithm, it also depends on capital-intensive industrial plant. Hindman points to Google’s mega data centre in Oregon, and fifteen like it running worldwide, which together “represent a jaw-droppingly massive investment, larger than the GDP of more than 100 countries.” It is here that we have to dispense with the mythology that the internet is a post-industrial technology. “Google’s facility is exactly what it looks like: an industrial mill, a digital smelter refining not ore but information,” writes Hindman. “Google’s data factories are just as critical for it today as broadcast equipment was for NBC in an earlier era.” This massive data processing capacity gives Google a critical speed advantage over competitors at a lower cost per user. And the energy efficiency of the massive servers run by firms like Google and Facebook dwarf the industry standard.

“Debates about the internet still often start with talk about the medium’s ‘openness,’ about how the ‘peer to peer’ architecture the Internet treats all websites equally,” says Hindman. But when the bandwidth of a corporate behemoth like Google equates to all the bandwidth between Europe and the United States, such talk is increasingly obsolete. As Hindman explains, “content delivery networks and paid peering mean that big sites take a shortcut to users.” In other words, the internet is less a web than an information superhighway after all.

Google’s search-results experiment illustrates yet another benefit of scale. Large sites have the necessary engineering expertise, infrastructure and user base to continuously run large-scale online experiments. When even small design changes can have substantial effects on users’ propensity to stay and return, large firms leave nothing to chance, using comparison testing to definitively determine the optimal approach. These experiments increase audience stickiness and substantially reduce the risk of a new product or feature failing.

Compounding other advantages enjoyed by big sites, the power of brands and our preference for the familiar are as pronounced online as elsewhere. We gravitate to sites we have learned how to use. And the digital heavyweights not only achieve audience growth at a lower cost than smaller competitors, they can sell their audiences to advertisers at much higher prices — because scale enables them to offer more precise targeting and more effective measurement tools.

“With so many strong economies of scale, of so many different types, in so many different areas of digital media,” Hindman concludes, “it is time to stop pretending that the internet is a level playing field.”


To illustrate the point, Hindman tells the story of Campus Network, a student-created social networking site that took Columbia University by storm in 2004. Within a month of its launch, “most students on campus were posting photos and blogs and polls, sharing music and upcoming events, and commenting on their friends’ activities. After conversations with Silicon Valley luminaries, such as Napster creator Sean Parker, the site expanded to many other campuses. The founder dropped out of school to work on the site full time. Soon the nascent social network had hundreds of thousands of users.”

Campus Network ended up folding in 2006. Today, around 2.3 billion of earth’s 7.7 billion inhabitants use another social network created in 2004, Facebook. (Of the 5.4 billion non-Facebook users, 3.5 billion are not even connected to the internet and well over a billion live in China, where it is blocked.) What explains the very different fates of the two companies? Campus Network was actually better than Facebook in some respects, says Hindman. From the start, it had features like blogging and cross-profile discussion, which Facebook didn’t offer. On the flipside, Facebook looked nicer and was easier to sign up to, and we could also speculate that its brand name afforded an easier journey when it expanded beyond its university origins.

But when we scrutinise the myriad possible explanations for Facebook’s ultimate and total victory, says Hindman, it’s easy to miss the woods for the trees. Facebook’s real advantage over Campus Network was that it got its nose in front. It vigorously pursued venture capital and advertising revenue in a way Campus Network didn’t, enabling Zuckerberg’s outfit to hire and expand faster. Once that happened, Facebook began enjoying economies of scale that its rivals didn’t have. Before long, the leading firm in the market was the only firm.

At one level, we only have to look to see that companies like Facebook and Google are gargantuan monopolies. And yet the internet of our imaginations — where attention flows freely and competition is only a click away — can prevent us from seeing what is right in front of our noses. Hindman’s clear theoretical exposition equips us to confront the empirical reality: the internet giants are natural monopolies. Following from that, governments need to start treating them as such — by much more aggressively prosecuting anti-competitive behaviour, for instance.

And yet, as disillusioned about the internet as we might be, the techno-optimists of the nineties and noughties got something right. As Nicholas Negroponte foresaw, we do indeed curate our daily news from an extraordinary variety of sources, limited by little more than our imaginations. In keeping with the future envisaged in the 1990s by the US Supreme Court, we can all be microbloggers publishing to an audience of potentially global reach (even if the publishing platforms we use are owned by remarkably few people).

In other words, monopoly internet platforms are perfectly compatible with radical content diversity. In fact, as Hindman shows, content personalisation and recommendation systems are crucial ingredients of audience stickiness that large firms can do better, cheaper. This is not just because personalisation requires massive investments in both hardware and software; it’s also because the capacity of an algorithm to accurately match content to a user’s preferences grows as more data from a larger user base is fed into it. It’s one more factor stacking the odds against would-be competitors to the Facebooks and Twitters.

Mark Zuckerberg’s monopoly is problematic in many ways but — by its nature — it is host to an immense diversity of independently published content. So why hasn’t that diminished our concern that powerful proprietors of old, like Rupert Murdoch can set the agenda of public discourse and exert excessive leverage over decision-makers? If we can all see and say anything we like on new media platforms, why hasn’t Rupert Murdoch’s voice become just another cry in the agora?

This question is further urged by the commercial fortunes of news publishers. Today, we worry less that they are monopolies than that they are going broke because they have lost so much revenue to Facebook, Google and online classifieds. If the news proprietors of old are so weakened, surely their hold on democratic life should have been loosened too?

If such a transformation is occurring then it is very hard to make out. As the editor of the Sun, Tony Gallagher, texted moments after Britain voted to leave the European Union and following months of ceaseless campaigning by his and other British tabloids: “so much for the waning power of the print media.” Indeed, between Brexit and the Iraq war, the ousting of Malcolm Turnbull and the defeat of Neil Kinnock (“It’s the Sun Wot Won It”), more continuity than change is apparent, notwithstanding the surrounding disruption of the intervening years.

How can this be? In this respect, Hindman could have more extensively delineated between the fate of platforms and publishers, but his basic arguments still go a long way. We might all have an equal capacity to publish on Facebook, but what matters is our ability to recruit and maintain an audience. That is distributed unequally because it requires intensive capital investment and hugely favours scale. Big publishers thus enjoy a decisive advantage over small ones, just as small publishers enjoy an advantage over readers. In print or online, “survival of the fattest” is still the law of the newspaper jungle. This unequal capacity to command an audience — to be heard — accounts for the enduring chokehold of the media barons over political life.

By adding Hindman’s analysis to what we already know, the major economic forces facing the news industry can be summarised: high capital requirements and powerful economies of scale together with diminishing total revenue and an increased proportion of that revenue coming from subscriptions (rather than advertising). Most likely to prosper are large news organisations with a global reach that appeal to a specific shade of opinion: think the Guardian, the Daily Mail and the New York Times. It’s a news ecosystem that can produce a hundred different flavours on Trump’s latest tweet but is less likely than ever to uncover corruption at the local council. What incentive is there to break news for a micro-audience that can’t offer the advantages of scale (especially when Google and Facebook are better at targeting advertising to local and hyperlocal audiences)?

One public policy takeaway is that Australia likely needs more extensive subsidies for local news and investigative journalism than the $60 million the federal government is delivering over the course of 2018 and 2019. It is a good thing that the ACCC’s Digital Platform Inquiry is contemplating such a move. The Canadian government’s recent announcement of a $600 million package of support for its media sector is one example we might look to, as well as the longstanding tradition of public support for journalism in much of Europe. The deeper lesson from Matthew Hindman’s illuminating book is that an egalitarian information environment depends as much on access to an audience as the capacity to publish; on the economics of attention as much as the tools of communication. •

Read next

1181 words

Yes, the Coalition does have a woman problem

The polls show a clear trend in how voters are responding

Right:

Frozen in the 1990s: the Coalition benches last month. Mick Tsikas/AAP Image

Frozen in the 1990s: the Coalition benches last month. Mick Tsikas/AAP Image