This is Part 2 of a series of posts. Part 1 can be found here.

We often see the term “network effect” in technology discussions. In common usage, the term can be misused to imply a natural tendency for networks to add value to all users as they grow. But this is not always the case.

Firstly, there are many different types of network effects, each with different characteristics and differing strengths. One thing that is often overlooked is that network effects can sometimes be negative and, in fact, sometimes both positive and negative effects can co-exist in the same network. An effect is positive when more people using the network gives everyone access to more value. But a broadband or mobile network may have negative network effects, where more users may lead to more congestion. We see something similar in a social network, where noisy content feeds caused by user growth can make it harder to find quality content. In a marketplace more sellers result in more competition for other sellers (negative), but more products for buyers (positive), which in turn leads to more buyers which is good for all sellers (positive). This specific example is called an indirect network effect. The NFX venture fund has an amazing collection of essays on network effects here: https://www.nfx.com/essays.

In this post, we’ll show why network effects are so important in Cosmos. The main reason is that Cosmos is not just a network: it’s a network of networks. Each sub-network has its own network effects which will interact with each other. When Ebay started they could focus on one network: buyers and sellers in a marketplace. Likewise for Uber: drivers and commuters. The Cosmos Hub has at least three types of network effects at launch.

NFX defines two-sided platform network effects as follows:

“... 2-Sided Platform nfx … the supply side actually engineers products that are only available on the platform. The supply side has to do work to integrate to the platform. The products created and sold by the suppliers are a function of the platform, not independent of it.”

At first glance, it may seem that Cosmos SDK fits into this classification. Developers build apps as the supply side, with app users as the demand side. And these apps are “not independent of the platform”. So this looks like a typical developer platform such as iOS, Windows, or Xbox. Here we can see that Cosmos is a platform.

But... NFX also define protocol network effects that “arise when a communications or computational standard is declared and all nodes and node creators can plug into the network using that protocol”. When we look at Cosmos this way our focus is on its ability to become a global standard for inter-blockchain communication, much like TCP/IP powers the internet or VHS became the video standard. So Cosmos is also a protocol.

Another perspective on Cosmos focuses on resource provision. We can look at Cosmos as having two-sided marketplace network effects, with validators providing computational resources on one side and app developers paying for these resources (either directly or by passing the costs onto their end users) on the other side. In fact, we can also see a second two-sided marketplace, with delegators as resource providers providing capital and validators as the demand-side, providing security and a safe return on that capital. So, from this perspective, Cosmos consists of two back-to-back two-sided marketplaces.

Fig.1: The Cosmos Hub as two back-to-back two-sided marketplaces

For Cosmos to succeed, each of these three classes of network effects (platform, protocol, and marketplace) need to strongly reinforce each other. They will each come into play at different phases in the growth of Cosmos.

As a platform, the focus needs to be on growing the developer community. The Cosmos community will need to build out the best tooling for developers to build, deploy and support applications. It will need to be cost-effective for developers to get their apps into production and in use. The community needs to create the best forums for the community of developers, sharing sample code, supporting each other, updating documentation, etc.

But it’s also important that the validators and delegators help to build developer and user confidence in the network, by securing the infrastructure, avoiding network downtime and improving overall resiliency. Cosmos validators will need to work together to ensure that as the first apps take off, the user experience is on par with centralized services like AWS and other centralized networks. The network must also remain cost-competitive so that developers and users are not put off by high fees, while maintaining high enough rewards to support world-class infrastructure providers and to provide competitive risk-adjusted yields for delegators, especially in light of competing investment opportunities in the decentralized finance (DeFi) space. This is why the effectiveness of the marketplace mechanisms are so important.

But the Cosmos vision of multi-token services and the interoperability of chains will become increasingly important over time. This is where protocol network effects come into play. While each effect can be built up over time, all three are mutually self-reinforcing. An early win with a multi-token service, even one with low transaction volumes, could have a meaningful impact on the long-term chances of the Cosmos Inter-Blockchain Communication (IBC) Protocol becoming the de facto standard for token exchange. But if there were weak rewards for validators or a high number of slashing events (causing weak delegator returns), this could start to negatively impact the quality of the network infrastructure thus damaging the long-term potential of IBC and weakening the attractiveness of Cosmos as a developer platform.

What is especially interesting is the relative strength of each network effect. Platform effects are weaker than protocol network effects. This is because developers love nothing more than trying out new tools and platforms, as is evidenced by the huge number of developer libraries and frameworks. Right now this is a strength for Cosmos, as Ethereum developers are more likely to experiment with a new platform. But even though platform effects are weaker, they can still provide an advantage, so it’s important for Cosmos to build an effective community of developers helping each other, contributing to tools, videos, Q&A forums etc. This will serve to strengthen the overall proposition, giving itself more time to build out the stronger protocol effects.

In summary, Cosmos is a network of network effects. If these network effects can reinforce each other, this could make Cosmos more powerful than any protocol, platform or marketplace that has previously existed. But it requires a delicate balance. It will be an exciting experiment!

Stay tuned for more in this Internet of Blockchain series, where we dive into staking economics, value capture, governance and more.

Follow us on Twitter and join our Telegram and mailing list to find out more.