Thesis #9

Winners take all in digital markets

In analog markets, there’s generally room for competition. Coke isn’t going to wipe out Pepsi. Toyota hasn’t wiped out Honda or Ford. But digital marketplaces appear to be different: they tend to produce winner-takes-all outcomes: Google in search, Facebook in social networking, Amazon in online retailing.

Why is this? It’s partly down to the fact that digital technology is distinctively different from older ones (see Thesis 1). Zero marginal costs mean that if you make the necessary early investment to develop a digital product, then it costs very little to roll out a million copies. The power of network effects mean that there’s a need to “get big fast” — and the quickest way of getting to the point where network effects really matter is giving away the product for free. And once you’ve got network power working for you then a positive feedback loop kicks in.

As an example, consider the history of Google. When it first appeared, its founders had invented an original way of ranking Web pages that were returned as a result of a search query. This was elegantly implemented in a fast and responsive way, and once users had discovered it they rarely went back to its earlier competitors. So usage of the search engine increased very rapidly, so Google used its early investment funding to build inexpensive, highly-optimised server farms to ensure that the service stayed responsive. All of this meant that the company was effectively training millions of users to search more often and more efficiently. And as it chose a business model — surveillance capitalism — to generate revenues, it began to collect data on all its users. This data made it possible to return search results that were more intelligently ‘personalised’. And the more people used the service the better it got. Which is how we wound up with a company that has an apparently unchallengeable monopoly of search. Google was the winner that took all.

Much the same story can be told about Google. And Amazon. And Facebook. The formula seems to be:

(mastery of digital technology) + (network effects) + (user data) = market dominance.

Winner-take-all outcomes are an indication that Power Law distributions are at work — even in non-commercial areas like blogging. Clay Shirky wrote a famous article on this way back in 2003 in which he showed that winner-take-all outcomes happen in the blogosphere too.

Further reading

Bernado Huberman, The Laws of the Web: Patterns in the Ecology of Information, MIT Press, 2003. Amazon UK:

Clay Shirky, “Power Laws, Weblogs, and Inequality”, 8 February, 2003

Om Malik, “In Silicon Valley Now, It’s Almost Always Winner Takes All”, The New Yorker, 30 December, 2015.

Back to List