Solving the Internet Monopolies Problem - Facebook, Google et al

FEBRUARY 24, 2018

The good news is that an increasing number of people seem to agree that:

  1. Facebook, Google etc are monopolies
  2. That is a problem

Agreeing we have a problem is always a crucial first step. But to go further we need to:

  1. Correctly diagnose the disease – in particular, avoid confusing the symptoms with the root cause
  2. Identify a cure that actually works

On point one, the answer is that he root cause is costless copying (plus platform effects) combined with monopoly rights. Costless copying (and platform effects) would always lead to some kind of standardization – witness the Internet. But that is not a problem – in fact it is advantage to have a single standard. The problem arises when the standard platform is owned by one entity and becomes a monopoly as we have today with Google in search, Facebook in social networking etc.

The solution flows from the diagnosis: make these platforms open and have an open-compatible, market-oriented way to pay innovators (i.e. have remuneration rights). By make a platform open I mean make its protocols, algorithms, software and know-how open, free to all to use, build on and share.

More details on all of this in my upcoming book:

An example of a mis-diagnosis: their control of our personal data

One prevalent misdiagnosis is that the issue with Facebook and Google and the source of (much) of their monopoly power is to do with their control of our personal data. See, for example, the Economist’s cover in April 2017 showing Internet monopolies as oil rigs mining personal data – an allusion to the common assertion that “personal data is the new oil” (see more examples added in the appendix below).

From this mis-diagnosis flows a proposed solution: limit Facebook and Google’s access to our personal data and/or ensure others have access to that personal data on equal terms (“data portability”).1

In fact, personal data is a practical irrelevancies to the monopoly issue. Focusing on it serves only to distract us from the real solutions.

Limiting Facebook’s and Google’s access to our personal data or making it more portable would make very little difference to their monopoly power, or reduce the deleterious effects of that power on innovation and freedom – the key freedoms of enterprise, choice and thought.

It make little difference because their monopoly just doesn’t arise from their access to our personal data. Instead it comes from massive economies of scale (costless copying) plus platform effects. If you removed Google’s and Facebook’s ability to use personal data to target ads tomorrow it would make very little difference to their top or bottom lines because their monopoly on our attention would be little changed and their ad targeting would be little diminished – in Google’s case the fact you type in a specific search from a particular location is already enough to target effectively and similar Facebook’s knowledge of your broad demographic characteristics would be enough given the lock-hold they have on our online attention.

What is needed in Google’s case is openness of the platform and in Facebook’s openness combined with guaranteed interoperability (“data portability” means little if everyone is one Facebook!).

Worse, focusing on privacy actually reinforces their monopoly position. It does so because privacy concerns:

  • Increase compliance costs which burden less wealthy competitors disproportionately. In particular, increased compliance costs make it harder for new firms to enter the market. A classic example is the “right to be forgotten” which actually makes it harder for alternative search firms to compete with Google.
  • Make it harder to get (permitted) access to user data on the platform and it is precisely (user-permitted) read/write access to a platform’s data that is the best chance for competition. In fact, it now gives monopolists the perfect excuse to deny such access: Facebook can now deny other competing firms (user-permitted) access to user data citing “privacy concerns”.

An example of a misguided solution: build a new open-source decentralized social network

Similarly, the idea sometimes put forward that we just need another open-source decentralized social network is completely implausible (even if run by Tim Berners-Lee).2

Platforms/networks like Facebook tend to standardize: witness phone networks, postal networks, electricity networks and even the Internet. We don’t want lots of incompatible social networks. We want one open one – just like we have one open Internet.

In addition, the idea that some open-source decentralized effort is going to take on an entrenched highly resourced monopoly on its own is ludicrous (the only hope would be if there was serious state assistance and regulation – just in the way that China got its own social networks by effectively excluding Facebook).

Instead, in the case of Facebook we need to address the monopoly at root: networks like this will always tend to standardization. The solution is ensure that we get an open rather than closed, proprietary global social network – just like we got with the open Internet.

Right now that would mean enforcing equal access rights to facebook API for competitors or, enforcing full open sourcing of key parts of the software and tech stack plus getting guarantees ongoing non-discriminatory API access.

Even more importantly we need to prevent these kind of monopolies in future – we want to stop shutting the door after the horse has bolted! This means systematic funding of open protocols and platforms. By open i mean the software, algorithms and non-personal data are open. And we need to fund the innovators who create and develop these and the way to do that is replacing patents/copyright with [remuneration rights][].

Fake news: confusing which symptoms are related to which disease

We must also be careful not to confuse which symptoms are related to which disease. For example, fake news is a problem but it is only tangentially related to the disease of monopoly. The causes of fake news are many and various and much more complex than simple monopoly. In fact, one could argue that more diversity in media actually makes echo chambers worse. Reducing monopoly at Facebook and Google level may bring some improvement but it is probably secondary. For more, see

Appendix - More Examples of Erroneous Belief that Data is the Asset

Added: 24 March 2018

From Guardian piece from Cambridge Analytica “scandal”

“Corporations like Google, Facebook, Amazon, all of these large companies, are making tens or hundreds of billions of dollars off of monetising people’s data,” Kaiser says. “I’ve been telling companies and governments for years that data is probably your most valuable asset. Individuals should be able to monetise their own data – that’s their own human value – not to be exploited.”

This is a perfect illustration of the current misunderstanding. Google and Facebook are making money from their monopoly of your attention not your data. Even if they zero data on you they would stil make lots of money and have a dominant position (just as dominant TV networks made lot of money without any data-driven ad-targeting in the old days). Amazon are slightly different: they are making their money from a monopoly marketplace at huge scale (and it’s not yet clear how much money they make from that). Once gain this has nothing to do with having your data.

Facebook / Google / Amazon’s monopoly come in turn from two key aspects of the information age – costless copying and platform effects – combined with exclusive ownership and control based in intellectual property monopoly rights (plus control of access to the platform).

This misunderstanding matters. The initial flaw in understanding the sources of these firms power leads then to the proposal of a wrong and misguided solution: allowing people to “monetise their own data” would make no difference to these monopolies (even if was clear how people could monetise their own data).

Instead what we need to do is open up these platforms: making their software and protocols open, and, where relevant, providing access to their platforms on fees and terms that are regulated, reasonable, universal and equitable.

  1. Data portability means almost nothing in a world where you have a dominant network. So what if I can get my data out of Facebook if no other network has a critical mass of participants. What is needed is that Facebook has a live, open read/write API that allows other platforms to connect if authorized by the user. [return]
  2. See as an example of just this kind of proposal [return]