The quadrant below compares web vs. crypto service models across two dimensions: the production model (from centralized to decentralized) and the data model (from custodial to non-custodial). The more decentralized and non-custodial a service, the more distributed its cost structure. This is important because markets tend to allocate value along the line of costs. So the more we decentralize the cost structure of a service, the more broadly we distribute its value.
The web model is a combination of centralized production and a “custodial” data model. This means companies like Google or Facebook manage the entire product lifecycle of their services, and stay in control of all the data contributed and generated by their users. They bear all the costs of production (e.g. hiring employees or building data centers) plus all of the liability of centralized data, which is an “underwriting risk” kind of cost.
Cryptoservices travel in the opposite direction, combining decentralized production with non-custodial data models. Cryptonetworks distribute the costs of production across a much larger group of people by incentivizing networks of independent nodes to peer-produce the service according to a shared cryptoeconomic protocol. Think of it as a digital franchise agreement, of the style that allowed McDonalds to achieve phenomenal global scale, but with a co-op type governance philosophy that exists only in digital form.
As for the data, every node has a copy so nobody has monopoly control over it. However, the user’s private data is encrypted with their own private key, so only they can operate on it. Custody is no longer about who has the data, but who has the power to use it – who has the keys. By default, the user owns their keys, but they can choose how to share or delegate control to third parties and applications to act as their “user agent”. It’s a non-custodial model in the sense that, while the service nodes in the network have collective “physical” custody of the data, they lack “effective” custody or control, so they are not exposed to the associated risks. This transfer of liability (and its many costs) from the service to the user is a powerful cost savings mechanism for the service providers.
How costs are distributed matters because value accrues to wherever the costs are. To understand why, think of each cost as an investment. The investors in an outcome are those who assume the cost of realizing it. You don’t make an investment without an interest in an equal or greater return, and you don’t assume a cost without some expected benefit. So value has to accrue to the investors, or the investment (and the outcome) doesn’t happen in the first place. In the modern startup playbook, that looks like a small group of entrepreneurs and investors assuming the bulk of the cost and the risk in exchange for the majority share of future value, which leads to value concentrating around the equity.
Now think of distributed cost structures as distributed investment structures. Cryptonetworks push as many costs to the edge and coordinates it all by means of tokens. Open source contributors and networks of nodes take on the costs of development and production and users take on the responsibility (and liability) over their piece of the data. The use of a token adds a third dimension to decentralized production which is decentralized capitalization. Every user who buys, holds or uses a cryptonetwork’s token is continuously capitalizing the network and bearing a share of the service’s cost of capital, and is thereby exposed to the increasing value of the token as the network grows over time.
Every participant who assumes a cost and a risk – by working, providing liquidity, injecting capital or taking care of their own security – earns a stake in the network, and is entitled to a piece of its future value as it manifests in the appreciation of the network’s capital. The market will naturally select cryptonetworks which allocate value proportionally to the community. Those which don’t will fail to attract the resources needed to thrive long-term. And therein lies the brilliance of the crypto model when compared to the web’s: if we’re right that self-sovereign cryptonetworks will win, then the path to success involves optimizing for cost distribution, which has the effect of distributing value more broadly. A total shift from how web services allocate value today.