By Marco Reuter
Source: a16z
Compilation: The Way of DeFi
If you want to create a “new Facebook” or a “new Google”, should you use a regular company or take advantage of a decentralized implementation via blockchain? The answer seems simple: cryptocurrency enthusiasts will answer resoundingly “yes,” while skeptics will shake their heads. But the decision has less to do with belief than with practical market design considerations.
The fundamental question is, exactly, how decentralization via blockchain brings value to your business. In this article, I guide you in analyzing this question, rooted in economic theory, and offer some insights to help you make a decision.
Like all models in economics, mine is built on several simplifying assumptions that may not apply in all situations. But the simplified model still helps reveal key sources of value for blockchains, and helps describe the context of when decentralization may or may not have value.
At the heart of this analysis is the “lock-in effect”, which means that it is often difficult for users to leave the network after joining it. Lock-in effects arise for different reasons. Users may incur switching costs – the annoyance cost of setting up an account in another network. Or, after spending time and effort training the network’s algorithms to suit their needs, users may find it less appealing to leave.
Think about how Google learns from your past searches and personalizes search results to match your interests. In addition, network effects can exacerbate lock-in effects, giving large networks an advantage over competitors. Lock-in effects are inherent in many businesses, but their magnitude varies from situation to situation.
For business owners, lock-in comes with the temptation to leverage locked-in users to increase profits. Say you have a company with millions or even billions of users, and market research shows that you can increase monetization—perhaps through advertising or charging higher platform fees—without losing customers. what will you do? Game theory analysis (and, frankly, common sense) predicts that you should increase monetization.
But the obvious danger is that users may know they’ll be locked out and don’t want to be exploited. If you can’t convincingly commit to not exploiting them (and you usually can’t), then in my model rational users will demand upfront “compensation”. This compensation may be in monetary form, but it can also take the form of no or minimal advertising during the growth phase of the network. Alternatively, users may refuse to join the network entirely for fear of being locked out.
However, the key insight when considering blockchain is that it allows you to generate trusted commitments through network design. That is, you give control over monetization decisions to users, enabling them to make decisions through decentralized governance. This means users can safely join the network because they don’t have to worry about being exploited later, even if they do get locked out.
This leads to the main result of my analysis: using blockchain to decentralize your business makes sense when the lock-in effect is strong enough.
Let’s explore this conclusion in more depth.
Model: Game Theoretic Analysis
The following are the assumptions supporting my model. Let’s say you’re an entrepreneur building a network and are considering whether to use a centralized company to build the network, or to decentralize the network through a blockchain. Additionally, users interact with the network many times over a long period of time. Every day there are potential users who learn about your network and can decide whether to join or not. Existing users can stay or leave, but existing users are locked in, so leaving the network for another option has become less attractive.
If you decide to implement the network through a formal company, you can change the intensity of monetization on a daily basis as needed. If you decentralize the network through the blockchain, users will vote on-chain to determine the strength of monetization.
In this model, as an entrepreneur, you are interested in maximizing profits (another simplifying assumption), and profits are determined as a function of the strength of network monetization, the number of users monetized, and price, e.g., the network The number of ads displayed on the website and the price advertisers are willing to pay for each ad and user.
In this model, users are interested in three aspects of the network. First, they derive utility from using the network—in economists’ parlance, using the network that gives them some form of value. Second, they don’t like monetization (like being forced to watch ads). Third, they enjoy whatever platform monetization revenue is shared with them.
Now we can solve the model by backward induction. That is, we look at game-theoretic predictions of centralized and decentralized governance, which will indicate which governance model you should choose to maximize profits at the outset of your network.
Choose Centralized Governance
Centralized governance allows entrepreneurs to decide the level of monetization for each period. Intuitively, two levels of monetization seem reasonable: first, a low level of monetization that makes new users willing to join; to let existing, locked-out users go.
A more profitable monetization strategy critically depends on the growth prospects of the network. When future growth is strong enough, you’re incentivized not to take advantage of the lock-in effect because you want to keep the network growing. But when growth slows sufficiently, it makes more sense to forego network growth in favor of leveraging existing user lock-in through increased monetization.
In balance, however, users — who have been at the mercy of the Web2 platform’s whims for years — have grown smart and foresee future exploitation. So their insistence on being “compensated” upfront (compensation could be money in my model, but could also take the form of little or no advertising during the growth phase of the network, for example) is expensive for you. Therefore, the control over monetization that you retain in centralized governance is both an advantage and a disadvantage of centralized governance. You are free to increase monetization to increase profits. However, you cannot commit to future monetization choices, but instead make the best choice at each given moment.
Decentralization using blockchain
When you choose to use blockchain decentralization, the first question to answer is how to share governance, that is, how many tokens are kept by yourself and how many tokens are distributed to users. If you want to reliably decentralize your business, you have to relinquish control, which means handing over most of your tokens to a third party. Otherwise, you can easily beat the user in any poll.
The trade-off in this is obvious. If users have control over monetization, then you can’t take advantage of user lock-in, and they won’t take advantage of their own lock-in. Thus, users do not need to be wary of future exploitation, nor do they need to be “compensated” beforehand. As a result, the network will be able to continue growing instead of going through a growth phase followed by an exploitation phase.
However, relinquishing control and giving up governance tokens comes with serious costs. First, you cannot choose the monetization of the network. Second, you have to share revenue with users through governance tokens. This is necessary to align the incentives between you and your users. If you don’t share any revenue, users won’t want any monetization of the network.
Should it be decentralized from the start?
It might be clearer considering some edge cases.
First, consider the case where there is no lock-in effect. To recap, the downside of a centralized implementation is the lack of commitment to exploit the lock-in effect in the future, which can lead to dissatisfied users, but they are still locked. But if there is no lock-in effect, there is no lack of commitment. This removes the disadvantages of centralization and leaves only the advantages. Therefore, if there is no lock-in effect or very small lock-in effect, centralization is usually preferred.
If the lock-in effect is very large, it may not even be possible to attract users to the network. In particular, the threat of future exploitation may be too great to overcome, even if you do offer monetary compensation or not monetize the network during the growth phase. In this case, decentralization is clearly preferable because only then can users be attracted to the network.
best choice
So, for a range of lock-in effects between the two extremes, what is the best option for governance? A graph can help visualize the dynamics. The x-axis is the magnitude of the lock-in effect, and the y-axis is your profit. (Blue line – centralized governance, red line – decentralized governance)
The revenue under decentralized governance does not depend on the size of the lock-in effect (users do not use their own lock-in effect), so it can be drawn as a horizontal line. When there is no lock-in effect, centralized governance is more profitable than decentralized governance. For lock-in effects above some sufficiently large threshold, we gain zero profit through centralized governance.
As the lock-in effect increases, the red line delineating the profits of decentralized governance does not change. But as the lock-in effect increases, the profit “blue line” of centralized governance slopes downward. This may seem counterintuitive, since greater lock-in effects allow entrepreneurs to achieve higher profits during the development phase. But in general, the cost of enticing users to join the network outweighs the profits generated by exploiting them in the future. As a result, perhaps surprisingly, centralized governance will end up being less profitable than decentralized governance can achieve.
This logic leads us to our initial conclusion: if the size of the lock-in effect is large enough, then it is optimal to choose decentralized governance. How “big enough” is has to be determined on a case-by-case basis and depends on the specifics of the network, such as the strength of network effects, user aversion to monetization, the potential profit of monetization, and user growth. (About how to measure the size of the lock-in effect, some academic and work may be of directional help.)
Airdrops: When Should You Decentralize?
The analysis so far assumes that you have to decide whether to decentralize at the beginning of the network. In practice, it is possible and common to start with a centralized network and gradually decentralize it over time. But decentralization cannot be dragged on for too long, because the decision to decentralize is subject to commitment issues, just like monetization decisions.
For example, suppose you have promised your network users that decentralized governance will be implemented later, but have been delaying decentralization of the network until you are reasonably satisfied with the results. But by then, the network has grown so large that it is more profitable to exploit users than to hand over control of the network through decentralization. If you delay decentralization for too long, users will expect that you will not be able to complete your plan.
That said, it may be beneficial to delay decentralization of the network for a period of time. In this model, it is profitable to delay decentralization until the promise of not exploiting users becomes sufficiently valuable. For practical purposes, it may also be beneficial to retain a greater degree of control over the network as it is built—for example, to troubleshoot problems, fix bugs, and improve the network in other ways (what some call “progressive decentralization”). ).
But again, the network must be decentralized before the temptation to maintain control of the network to increase profits becomes too great. In practice, this often means temporarily controlling the network, then airdropping governance tokens and switching to decentralized governance. It may also be beneficial to distribute tokens over time to gradually reinforce the required commitment, rather than distributing them all at once. Decentralization and the timing of token issuance are also often influenced by applicable regulatory considerations.
At last
In this article, I try to convince you that decentralization can create value for your business in one way (there may be many ways): by committing to not exploiting the value of your users in the future. I take the strength of the lock-in effect as the key variable. If user lock-in is strong enough—and thus your commitment problem is severe enough—decentralization via blockchain is the more profitable option. If the lock-in effect is small—if users can come and go freely and easily—it may be more beneficial to implement the network through a centralized company.