In 2010, Facebook introduced the “Open Graph”. In Mark Zuckerberg’s own words:
“Today the web exists mostly as a series of unstructured links between pages. This has been a powerful model, but it’s really just the start. The Open Graph puts people at the center of the web. It means that the web can become a set of personally and semantically meaningful connections between people and things.”
The “unstructured links” Zuck refers to are a clear jab at Google. While Google has built its business on publicly available data, organized through links, Facebook’s model is closed. This is a huge advantage for Facebook, since exclusive data can be a tremendous competitive edge. The downside, however, is that closed doors prevent valuable data from getting in. Open Graph changed this by offering slices of Facebook’s data in exchange for data from the outside. This offer of open innovation allowed Facebook to build greater integration with external services, which, at the time, were more likely to provide Facebook with a fair trade. Then, in 2017, it was discovered that Cambridge Analytica had used Open Graph to harvest more than 50 million user profiles, for the purposes of election manipulation.
Facebook started shutting down parts of its Open Graph in 2015, but not after significant damage had already been dealt. Regulation in the form of GDPR in Europe has already forcibly locked down data sharing. This issue comes with hard trade-offs. Decreased data sharing improves privacy but entrenches incumbents. Instagram was launched off the back of Twitter’s exportable social graph. Later, Instagram’s growth was further accelerated by piggybacking off Facebook’s social graph. Now, amusingly, Twitter has been slaughtered by Facebook, which acquired Instagram in 2012. This rapid growth, and surprising takeover, would not have been possible without open networks. Open networks can compromise monopolies, and now all social networks hold their user data under lock and key.
We can observe a general trend from open innovation to closed innovation. When Windows launched, it was the ultimate open platform (not in terms of how it was built, but by how it was used). Developers could build anything on top of the Windows operating system and sell directly to users. Windows was a business that only captured a minority of the total pie. Apple started to shake this model. Their App Store, for example, captures a tax on all app sales. By facilitating the relationship between users and app providers, Apple can claim a majority of the pie. Now, Facebook has taken this trend to the extreme. By commoditizing content creation and completely controlling consumption, Facebook can own the entire pie. Furthermore, if we look at the cross section of these companies individually, we can infer that open innovation is powerful when you’re the underdog, but less valuable when it’s time to build your moat and collect maximum rent.
Is this trend harmful? I would argue not necessarily, and certainly not for the incumbent. By closing innovation, Facebook can better control the user experience. We’re shifting focus from distribution, which hinges on exclusive distributor-supplier relationships, to user experience, which Facebook can curate exclusively. Uber followed a similar model by commoditizing dispatch vehicles so that it could focus on hailing, pricing, and dispatch. The powerful network effects of Facebook and Uber’s businesses mean that they can sustain massive growth despite offering suboptimal service. As it turns out, you don’t need the best service (one driven by open competition), you just need the service that everyone uses.
It’s worth noting that the tradeoffs of open innovation are not as simple as I represent above. I’d like to hear some counterexamples from my classmates, but here are a few. Google, despite having first mover advantage and the best engineers in the world, is getting trounced by AWS. Amazon’s technology stack was architected from the ground up to be service-oriented. Jeff Bezos made this a requirement for all projects, whether they were externally accessible or not. This made the leap from a closed system (services for Amazon) to an open one (services for everyone via AWS) incredibly seamless. By planning for open innovation, they were able to transform their infrastructure into a generalized computing platform. It’s not clear when, if ever, open innovation will cease to be effective for Amazon’s business model. Another interesting twist on open innovation is the blockchain. By tying the value of the applications to the underlying protocol, we can hoard value in the underlying platform even if (and especially if) the platform is open. Typically, monetary value is concentrated at the application layer (e.g. the founders of Google are much richer than the inventors of tcp, http etc.), but blockchain flips this on its head. Since applications make the underlying tokens more valuable, and the value of the tokens is greater than the combination of all the applications built on top (due to speculation), we can encourage open innovation while still concentrating value in the platform.