Light on the Dark Side of Network Effects

Many Internet services have network effects — the platform that takes a critical mass of users would take most. While this “winner-take-all” pattern creates a fair reward for first movers, if abused, it could result in a monopoly that harms users. It locks in users and suppresses competition. However, the current Internet ecosystem fails to prevent the problem because users are bound to service providers. If users are able to fully control their identities and relationships out of any service provider, we will see competition that is healthier and more equitable. This is similar to the WLNP protecting consumers’ right to keep their phone numbers when switching to a different carrier. A consumer doesn’t have to tolerate a carrier only because he/she doesn’t want to change his/her phone number which is known by all his/her contacts (banks, apps, etc.). Unfortunately, the current Internet ecosystem lacks this type of protection. We believe decentralization can help build this protection and provide independence for users, even though some current blockchain-based projects take an opposite position and repeat the old story.


Network Effect: Why is GitHub Worth $75 billion?

(See what is GitHub if you are not familiar with it.)
On the day when Microsoft announced its acquisition of GitHub, a whopping number of repositories (~100,000) were ported to GitLab, a major rival of GitHub. However, GitHub still has over 100x more repositories than GitLab. The answer to what prevents users from choosing another service provider is why GitHub is worth the acquisition price. Interestingly, the answer can be found on GitLab’s webpage for importing from GitHub:

Import your project from GitHub to GitLab

The information on the top half of the page sounds really pleasant. Most of the features most users care about are portable, including pull requests, which are “the greatest thing” that GitHub invented. It is implied that those features alone are actually not the most valuable asset of GitHub. However, the information on the bottom half of the page reveals what is the “grain of sand in your shoe” — you’ll have “wild pointers” to your peer developers and collaborative repositories, and you’ll miss them after moving to the new platform.

It isn’t the mountain ahead that wears you out — it’s the grain of sand in your shoe.
 — Robert W. Service

Collaboration among developers is the essence of GitHub or GitLab. Any developer can fork a public repository, work on the fork, and create a pull request to contribute code back to the original repository. For example, the popular JavaScript framework, vue.js, has over 14K forks. If vue.js suddenly switched to GitLab, the workflow would be broken. Another form of collaboration is developers’ discussing issues of a project. For instance, the most commented issue of TensorFlow involves 145 developers. Furthermore, in-depth research has shown that the social and technical activities of developers play a key role in their collaboration (e.g., contribution acceptance). All those collaborations go void when a project moves to another platform. That is the network effect: the value of a service increases with the number of users because more users build more valuable relationships.

The network of relationships among users and repositories is the most valuable asset of GitHub. Next, we will see significant reasons to make this network independent from any service provider.


The Good, the Bad, and the Ugly of Monopolies

Monopolies make successful businesses, and many of them are guarded by network effects — thanks to your friends on Facebook or your collaborators on GitHub, you would find it hard to say goodbye to those platforms.

Monopoly is the condition of every successful business.
 — Peter Thiel

Consequently, the “Frightful Five”, Amazon, Apple, Facebook, Microsoft, and Google, have got us. It doesn’t sound too bad, because they are “creative monopolies”. They offer the best services, so they have gained the most users. While that is true in a certain time period, it isn’t true forever. Network effects give current winners unbeatable power, but they make it tough for late comers to stand out even though they are really outstanding. A super network-effect power is liable to incubate evil or mistakes.

Take Amazon for example. On the light side, Amazon has revolutionized the publishing industry and it makes books more affordable and easier to access. On the dark side, Amazon has been reported to bully small publishers by threatening them with moving their books to full price, taking them off the recommendation engine, or promoting competitors’ titles. An Amazon employee was even fired because he questioned the ethics of revisiting a contract with Oxford University Press.

Similarly, on the light side, Facebook helps people connect with each other, share information, and discover new things. On the dark side, Facebook is sitting on its huge proprietary data network and controls a dominant channel of information flow among people. This type of monopoly attracts stealers, cheaters, and manipulators. After the Facebook–Cambridge Analytica data scandal (a data breach exploited to influence the 2016 U.S. presidential election), we know it is not alarmism. Our privacy, views, and even democracy are at risk.


Network Effects Made Independent

We need to continue the technology revolution. But, the direction is not to reallocate billions of dollars from content creators to monopoly platforms, such as Amazon.

Neither is it to create the filter bubbles through a black-box recommendation algorithm trained or controlled by a single dominant service provider, like Facebook, who exclusively owns massive amounts of users’ data.

Nor is it to limit the opportunities to innovate and to deploy innovations because users are locked by a proprietary platform like GitHub even though the users themselves are innovative developers.

We seem no longer to control our own destinies.
 — Joseph Stiglitz, “America Has a Monopoly Problem — and It’s Huge”.

Network effects at the cost of users’ independence are dangerous. We are alert to predatory dumping and monopoly pricing, but not alert enough to the dark side of network effects in the current Internet era.

We believe the right direction is to decouple users’ data and networks from platforms and service providers through the power of decentralization. In a decentralized world, users manage their data and relationships in an open format on a blockchain, which is a public facility and independent from any service provider. Users still need traditional service providers, such as an “Amazon” to find suppliers, a “Facebook” to collect News Feed, or a “GitHub” to discuss issues and manage pull requests. The fundamental difference is that individual users have the freedom to switch to a different service provider without losing their functional data and relationships, just like they can switch to a different cellphone carrier without changing their phone numbers. This freedom can reduce the monopoly rents a company like Amazon strives to exploit. It can give people a choice of their own views of the world beyond the single Facebook. It can also enable developers to optimize their workflows in a variation of GitHub.

Decentralization is the return of users’ rights to control their data, identities and relationships. It makes network effects independent from any individual service provider.

Is the vision technically viable?

We currently don’t have solutions to all of the platforms mentioned above, but we are developing drepo, a decentralized GitHub. We articulate its system architecture and design rationale in this previous article. Note that feature-level portability between multiple service providers is not a big issue as “Import your project from GitHub to GitLab” indicates, although we haven’t supported multiple front ends yet.

Drepo is in a very initial stage. Your input is welcome!