Real estate, alternative real assets and other diversions

Even online, there’s no such thing as a free audience How the internet economy works and why it matters

Technology

The internet was supposed to fragment audiences and make media monopolies impossible. Instead, behemoths like Google and Facebook now dominate the time we spend online — and grab all the profits from the attention economy. A timely new book explains the stunning rise of the digital giants and the online struggles of nearly everyone else.

How did online audiences and digital revenue get so concentrated? Is online oligopoly inevitable or is there a way out of the “internet trap”? These are the questions Matthew Hindman sets out to answer in his new book The Internet Trap: How the Digital Economy Builds Monopolies and Undermines Democracy.

Many think of the internet as a cheap, sometimes even free, place to operate. Hindman shows that is a common misconception. The internet has not reduced the cost of reaching audiences — it has merely shifted who pays and how.

The digital distribution costs — server farms, software platforms, site designs, mobile app development, content production and search costs — are expensive. They are also necessary for audience-building. No one likes a slow loading page. Stickiness — what makes a person click and stay on a page – depends heavily on distribution investments.

To put the costs into perspective: the largest cities in the world pay roughly an eighth or a tenth of the overhead electricity costs of Google’s or Facebook’s data centres. But money isn’t everything. On the internet, data is king.

What the internet giants have in common is being the first to use personalised recommender systems, one of the most powerful tools available for sites to keep and grow their traffic. Such systems are challenging to get right and to be successful, size matters, both in terms of resources but also in terms of personal data. You need to know, or at least be able to make a pretty good guess, what someone prefers in order to recommend something to them. The big firms started collecting personal data very early on and grew bigger the more data they collected.

Recommender systems are powerful because they make sites sticky. The example of Google News illustrates that. Originally, Google News recommended entertainment news to everyone because it was generally popular, it didn’t matter that you had never been interested in showbiz news. By changing to a new content-based method which instead looks at what you have looked at in the past, they can offer a truly personalised feed based on your interests.

This is, of course, convenient and more likely to keep you on that site. Your cost of spending effort finding interesting content on the web is dramatically reduced, while the cost of switching app or site is increased. Why switch to Bing when Google already knows what you are interested in? Companies like Netflix and Amazon depend on content recommendation systems for a large part of their revenue and an even bigger chunk of their profits.

The rise of recommender systems was a tectonic shift in the media landscape. They also favour firms with lots of content, better hardware and more data. Therefore, they promote audience concentration. As Hindman points out, economies of scope, such as Google and Facebook where the efficiencies come from providing a broad mix of different products, generate concentration just as surely as economies of scale.

That’s how the monopolies on the internet were formed, and that is what Hindman calls the “internet trap”, which is incredibly difficult for smaller organisations to escape. To get big, you need the resources and data, but you only get that if you’re already big. Growth is a function of how large a site’s audience already is. Being the first to use recommender systems is how Google, Facebook and Amazon got their initial advantage, and helps explain the continuing inequalities in digital participation.

These inequalities matter because as long as they are there, smaller organisations are made to depend on the big internet firms to attract audiences. Newspapers, for instance, rely heavily on Google and Facebook in order to increase their web traffic. In January 2018, after Facebook changed its algorithms, many online news sites saw a dramatic fall in traffic to their pages. The analytics site Parse.ly noted that overall Facebook traffic referrals to publishers were sharply down, to about 26 per cent from 40 per cent over the year, while Google’s was growing.

Online news illustrate just what power the big firms have. Many might think that people still read news online – they just don’t pay for it, but that is another misconception, according to Hindman. While the introduction of tablets and smartphones certainly seemed to promise greater online readership – the iPad came to life in 2010, by 2017 around 55 per cent of US adults owned a tablet device, the sad truth is that news only consist of three per cent of web traffic. Newspapers can’t monetise an audience they don’t have.

To gain more traffic, newspapers need their sites to get stickier to be able to compete with Google and Facebook for our attention. They need to implement better models of web traffic and personalised content recommendation systems (not just a ‘most read’ list).

Newspapers also have great difficulties attracting and retaining technical staff. That is mostly down to a culture of not seeing tech workers as vital to the paper’s core function. Talented software engineers and web designers who could help increase web traffic immensely receive six-figure salaries at tech firms where they are seen as integral to the organisation’s mission, and not just someone on hand to fix the printer. Without smart design and constant testing, the digital audience is lost and newspapers must acknowledge that technical staff is as important to distribution as their printing presses.

Our Partners