The Network Effect
February 11, 2010
The Economist recently published a special report on social networking, which attributed Facebook’s extraordinary growth to ‘the network effect’. This is more or less true but the writer is conflating two different phenomena: network effects, a concept drawn from economics, and positive feedback loops, from the world of engineering. It is their combination that sometimes produces exponential growth and allows one winner to take all.
Network effects arise because the bigger a network is, the more valuable it becomes to its members. This is true whether it is a physical network like a telephone system or a virtual network like the buyers and sellers on eBay and the subscribers to dating agencies: they have more people to speak to, to trade with, to fall in love with. Bob Metcalf, the creator of Ethernet and of the ‘law’ that bears his name, suggested that the value of a network is proportional to the square of the number of its members.
But networks are not all alike and even the most dynamic only grow exponentially for a few years at most – in their early days, growth is very slow and eventually, as they mature, it tails off. What happens in between is that positive feedback loops kick in and growth becomes self-reinforcing: small changes are amplified to produce a bandwagon effect. Partly because of self-fulfilling expectations, success breeds success, and failure, failure. The importance of feedback has long been recognised in physical systems, but it is only recently that its role in social networks has been understood, and even now not with any precision. Although he didn’t use the terms, when Malcolm Gladwell wrote about social epidemics in The Tipping Point, he was essentially talking about the combination of network effects and feedback loops. (All these terms are metaphors, but mostly useful ones.)
Positive feedback accelerates the market growth of all successful new products and services, but it is most dramatic in the case of networks like Facebook, which really has been a social epidemic. The first Internet business to benefit as spectacularly was eBay, but at the time, the founders and most investors were only dimly aware of what was happening. eBay only became a business in 1996, and then only part-time, but it grew so explosively, without any active stimulation that it was soon impossible for competitors to attract traders to their online auctions. With 300,000 members in 1997, eBay offered sellers more buyers, and buyers more sellers, than anyone other site, and very few traders strayed away for long. When Yahoo and Amazon entered the market in 1999 with superficially more attractive services – eBay’s customer service has never been remotely as good as Amazon’s – it had 10 million members and 2.3 million listings against its rival’s 55,000. eBay’s feedback loops have since run out of steam, but it still enjoys powerful network effects. Despite having alienated many of its traders, the size of its network alone keeps most of them there.
The Internet has been an immensely fertile breeding ground for networked businesses. AOL and Netscape both enjoyed combinations of network magnetism and self-generating growth comparable to eBay’s, though in their cases they didn’t last. Google has been another major beneficiary: exponential growth in its search audience made its proposition to advertisers ever-more compelling. The quality of the search results is the product of another virtuous circle, its extraordinarily scalable network of computers that as it got bigger, got better. Perhaps the most adept network strategists were the founders of YouTube, which Google acquired for $1.65 billion in 2006. It is no accident that Chad Hurley and Steve Chen had both previously worked at PayPal, whose understanding of network dynamics had enabled it to outmaneuver eBay. PayPay duly won the race to be the leading online payments system, and eBay was obliged to pay $1.5 billion to acquire it in 2002.
These phenomena did not start with the Internet – American Express’s traveller’s check in 1891 was an early example. Fax grew explosively in the early 1980s, yet it had been technically possible for a long time before that. When cheap machines came onto the market it spread like wildfire, largely by word-of-mouth. Fax was soon eclipsed when e-mail became contagious and even more ubiquitous. The original email systems were closed, proprietary ones, used mainly for internal communications by a few large organisations. When the Internet emerged as the common standard and big companies switched over, there were suddenly hundreds of thousands of email users on the Net. Very quickly, every business and most individuals could see a reason for joining the Internet, which more than doubled in size every year in the 1990s.
Nobody owned these customers and proprietary online systems often lost theirs. AOL though convinced many people, notably Time Warner, that ‘owning’ so many eyeballs made it the most valuable company in cyberspace. This particular network proved to be a perishable asset and Time Warner paid dearly for its naivety.
Probably the biggest beneficiary from network effects and feedback loops has been Microsoft, which achieved its domination of the desktop long before the Internet became a mass medium. The key step was holding onto the rights to the DOS operating system it obtained for IBM’s first PC. But it was far from inevitable, even in the late 1980s, that DOS’s successor, Windows, would become the universal standard. IBM had a new operating system for business customers and Apple’s Mac was easier for consumers to use. However when it became clear that Windows 3.0 would have by far the largest market share a bandwagon started: most businesses and consumers wanted the same system that everybody else was going for, and Microsoft sold 10 million licenses between 1990 and 1992.
By now it was extending its dominion to other desktop markets, edging out Lotus and WordPerfect, the previous leaders in spreadsheets and word-processing. The new vogue for connecting computers over local area networks encouraged standardisation in these programs as well as operating systems. Once people started exchanging files, it was infinitely easier if they all used the same software. Naturally they standardised on the products that had emerged as the new market leaders, and which worked so conveniently well with the new standard in PC operating systems. By 1994 Microsoft had 70% of the markets for spreadsheets and word-processing and more than 90 percent in operating systems.
Like most cases where positive feedback kicks in, this was not entirely planned. Bill Gates acknowledged that he had been slow to spot the significance of computer networking in general and the Internet in particular, but he benefited enormously. Microsoft went on to pulverise Netscape and to extend its hegemony to the Internet browser market. It really did look for a while as if one winner was going to scoop up all the prizes, but it didn’t quite work out that way. However, its networked assets have continued to yield it rents of billions of dollars every year since.
Shakespeare grasped the basic principle 400 years ago:
There is a tide in the affairs of men
Which, taken at the flood, leads on to fortune;
Omitted, all the voyage of their life
Is bound in shallows and in miseries.