Please see the “Open Source” version here.
Disclosure: DAO frameworks discussed below not referencing existing DAO communities represent theoretical frameworks for DAO formation. There is a piece by a16z covering potential legal liabilities for DAO members here.
A revolution in business organization, DAOs, or decentralized autonomous organizations, are becoming the structure of choice for many web3 organizations.
What is a DAO? According to Ethereum.org, DAOs are “member-owned communities without centralized leadership.” DAOs operate through blockchain-based governance, meaning decisions made by the community members are implemented on a blockchain. A DAO is founded by an initial group of creators/collaborators who build the smart-contract based application on the blockchain, and make initial determinations about the governance structure and rules of the organization they are creating. Versus traditional corporate entities, DAOs:
Examples of Different types of DAOs? DAOs are being used for a variety of use cases. Below is a short summary of the types of DAOs being created today:
This is by no means an exhaustive list. DAOs are constantly emerging to service communities in new and exciting ways. However, all DAOs share some fundamental traits: some of their rules are encoded in smart contracts on the blockchain, community members are able to actively participate in governance, and community members operate from a shared treasury that holds the assets of the DAO.
How are DAOs governed? DAOs are intended to literally decentralize governance across members of the organization by allowing them to vote (through various voting methods from delegation to guilds to quadratic voting) on the future of the organization. This model runs somewhat counter to traditional organizations, where votes about the direction of the entity are sometimes delegated to superiors in a hierarchy. While in some situations shareholders in traditional organizations can vote on company proposals and directly influence decision making and company direction, with DAOs the entire proposal and voting process should be decentralized and voters should be rewarded for their participation.
In a DAO, voting members might have the power to make changes or updates to a project, determine asset allocation, and more. There are two forms of DAO governance: off-chain and on-chain governance.
How do DAOs work in action? An example DAO structure is reflected in the setup of one of the original DeFi DAOs, Compound, (the DeFi protocol):
How do participants receive their tokens? Mechanisms for token distribution differ based on DAOs. In some cases, tokens are “airdropped” to initial users of a protocol, meaning users automatically receive a certain amount of tokens to reward past or current participation in a network. In DeFi, yield farming has become a primary token distribution mechanic of choice, because it allows DeFi protocols like Compound to bootstrap tremendous liquidity. What is yield farming? Yield farming describes crypto asset holders providing liquidity to networks (i.e. putting their assets into liquidity pools) in exchange for high yields on their assets as well as tokens like COMP. Users can also buy governance tokens for DAOs on DEXs (Decentralized Exchanges) and CEXs (Centralized Exchanges).
In general, DAOs and DeFi protocols give away tokens to incentivize behaviors that are good for the community. These behaviors include providing liquidity (DeFi), maintaining the protocol (DeFi), and incentivizing adoption and contribution (DAOs generally). (Thanks Zakk)
DAO structures are hailed for their ability to enable more decentralized organizations versus traditional corporate entities. In the Compound example [also used by Underscore.vc as an example here and discussed by a16z here], the Compound community shares responsibility for the ongoing management of the protocol. Because participants benefit from participating in the Compound network, community members who hold governance tokens are (in theory) highly incentivized to act in the best interest of the protocol.
However, the creators of a DAO control the initial token distribution. In Compound’s case, upon release of the new decentralized governance structure, tokens were distributed in the following manner:
Link to Sheet Here.
As per the Token Distribution Table above, Compound split governance tokens essentially 50/50; 50% among the creators and shareholders of the protocol as well as founders and current team members, and 50% reserved for future employees, users and community members. Compound’s current voting threshold is such that:
“Voting lasts for 3 days; if a majority, and at least 400,000 votes are cast for the proposal, it is queued in the Timelock, and can be implemented 2 days later. In total, any change to the protocol takes at least one week.”
In conclusion, the original holders of the governance tokens, as well as core contributors employed by Compound, still maintain a level of asymmetric power in proposing changes and generally maintaining the Compound protocol. [Later in the piece I discuss why another DAO might structure token distribution in this way].
Voting on past Compound proposals can be seen publicly on Tally here, which shows the activity of current voters.
Compound is just one example of a DAO, and while other DAOs have adopted elements of their governance structure, they are by no means a set template for DAO governance. DAO token structure and governance is a complex and contentious topic. The level of decentralization set by the creators is dependent on a variety of factors from the age of the project, number of contributors, assets under management, financing raised from traditional financial backers like venture firms, and long term goals for the product/project.
2021 has seen an explosion in the number of DAOs created, governance structures implemented, and use cases for which DAOs are created. Lisa Xu has created a lovely market map with some of the popular DAOs launched to date.
Today, DAOs are viewed as a potentially viable organizational structure any time founders or a community want to:
DAOs continue to rise in popularity in conjunction with the overall increase in adoption of web3 technologies and products. DAO membership has increased by over 100% since August ‘21, with more than 1.6M entities across 164 organizations. Meanwhile, DAOs hold over $12B in assets under management.
As DAOs are constantly formed and new entrants to the world of DAOs are exposed to this new form of work and income, we are seeing the first wave of DAO infrastructure innovation. Tools are emerging to solve problems with governance (voting platforms like Tally, Snapshot), treasury management (Gnosis), participation (Layer3), identity management (ENS), payroll (Utopia Labs), collaboration (Wonderverse), token gating (Collab.land, Guild.xyz), and more.
However, the fundamental structure of a truly decentralized organization poses some problems that can not currently be solved by this wave of tools. These issues mostly fit into the following buckets:
1. Core contributors are needed to perform some key tasks that would be handled by key team members in a traditional organization
At the point of inception of a DAO, a small group of individual contributors is expending time and energy to create the DAO, program the smart contracts, and determine the rules for the future. This problem of requiring core contributors to perform certain tasks persists for today’s DAOs. DAOs have struggled with problems like onboarding, contributor retention, and internal knowledge sharing. There is also potentially a massive talent gap, due to these issues as well as the self-selection involved in participating in a DAO. For example, DAO work likely attracts entrepreneurial and self-directed builders that thrive in autonomous organizations. (Thanks to Jim for his insightful tweet thread on these topics).
Core contributors also may need to perform tasks that require a level of autonomy not afforded by a fully decentralized org. These tasks include engaging with traditional institutions as well as signing legal documents & managing ongoing coordination with legal counsel.
2. The majority of contribution is often coming from a very small percent of participants (again, thanks to Jim)
Many DAOs are being built primarily by a very small number of core contributors. This means that (1) token distribution is being concentrated among these contributors and (2) some DAOs are facing scaling issues when they struggle to increase the number of core contributors.
3.** Theoretical issues with pure “coin-voting” (on-chain, token-based) governance per Vitalik Buterin’s recent piece
As Buterin wrote in a recent essay on coin-voting based governance, (which I described previously, as on-chain governance), “pure DAOs” can solve critical problems for crypto projects like sufficiency of developer funding and “credible neutrality of funding.” However, there are unsolved issues with coin-voting, including the following issues observed by Buterin:
Today, token concentration among large holders is a particularly salient issue for DAOs pursuing greater decentralization. (Note: token concentration is also an issue facing staking protocols). And while concentration of token distribution is in part what has kept some DAOs from facing issues with pure coin-governance, in the future as these organizations become larger and more distributed they will likely face increasing pressure to adopt governance structures that go beyond coin-voting.
4. Decentralization often involves a tradeoff between speed of creation/crisis response and level of decentralization (per Compound CEO Robert Leschner’s comments in this video)
In the early days of project creation, founders may have to deal with relative tradeoffs of speed of building initial product features and bugs, with embedding decentralization into the core foundations of a product. For many DeFi products, it takes the core contributing team time to reach a point at which they feel their project is stable enough to begin decentralizing governance to the community. Then, over a certain period the core team can begin testing decentralization while working out additional kinks in the product.
Additionally, by the time a product/project is transitioned to decentralized governance, the core contributors may have already spent years as a centralized entity, which results in questions about how to fairly compensate those core contributors while giving the community power.
5. Legal uncertainty of DAOs and liabilities held by founders/investors/members
Today, a “pure DAO” is not a legally recognized entity. In order to secure the liability protections associated with a typical startup, DAOs need to register as an LLC in a state where it is possible for a DAO to do so (such as Wyoming). However, per a recent a16z piece there may be additional legal liability from participating in a DAO.
Additionally, DAOs expand the definition of work and employment. Many DAO contributors are located around the world, with various employment laws, executing on various types and duration of work.
6. Engaging with the traditional world of banks and other organizations
Lastly, DAOs are not yet recognized as full corporation alternatives by many institutions in the traditional world. DAOs looking to work with a traditional bank, for example, may face difficulties in convincing the institution of their security protocols, governance structure and risk management systems.
While the vision of a truly decentralized organization that maximizes benefit for stakeholders is quite appealing, we have not yet found many fully decentralized ways of solving the above stated issues. Jesse Walden writes that many web3 companies need to go through stages of decentralization, which he called “progressive decentralization”, where they can adjust the level of decentralization based on the stage of the product (and product/market fit), community size, engagement, and level of incentive alignment.
Thus, for those looking to create DAOs that have the potential for a more decentralized future, a current theoretical option is to adopt a hybrid model that adapts features of both a DAO and a traditional corporation. This hybrid model would blend elements of decision-making, liability, and ownership between the two structures.
So what might a hybrid model for a DAO look like? There are a few theoretical ways to find middle ground in DAO structure today:
1. Follow a Compound-like, more centralized, token structure
Compound’s more centralized token structure allows for incentivization of future employees and core contributors, so that they can perform crucial tasks for the company. As per the cap table earlier in this piece, Compound has armed itself with the power to issue a significant percent of tokens to these employees. In theory, these employees can perform the tasks that a community contributor might not, including dedicating all of their productive working hours to projects like onboarding guides, internal knowledge documentation and sharing, and employee culture/happiness/retention. Compound can also use tokens to solve near-term talent shortages as the community continues to grow in size and engagement.
While following Compound’s exact token distribution might not be best for every organization, considering a more centralized token distribution that can then shift to a more decentralized distribution over time could be an option for a DAO.
2. Adopt governance structures beyond pure coin-voting
Per Buterin’s piece, I agree that coin-governance will not be the only governance option adopted by successful decentralized organizations. Vitalik proposes a few intriguing hybrid models that solve some of the issues presented by coin-governed DAOs today, particularly around off-chain versus on-chain governance and the concentration of token distribution among “whales” in a DAO ecosystem.
One key idea is the proposal of “limited governance to fixed parameter choices,” an idea adopted by Uniswap where governance only impacts specific parameters; in Uniswap’s case the parameters are token distribution and a 0.05% fee in the Uniswap exchange.
Another proposed framework is reputation-based voting, wherein a participant votes with a token that can’t be transferred to another holder. Participants gain more reputation if their decisions in voting processes lead to desired results, and lose reputation if their decisions lead to undesired results. Reputation-based voting is a hybrid structure that attempts to reduce the ability to engage in vote-buying behavior, for DAOs where community members are more frequently making decisions and voting on proposals.
Another is “loosely coupled (advisory) coin votes,” where a coin vote does not directly implement change, it just exists to make its outcome public to build legitimacy for off-chain governance to implement that change. According to Buterin, advisory coin votes provide the benefits of coin votes with fewer risks, as the legitimacy of a coin vote drops off automatically if evidence emerges that the coin vote was bribed or otherwise manipulated.
3. Create a DAO with a more centralized model, but then allow subDAOs to operate from a shared treasury with more decentralized structures
What is a SubDao? According to YieldGuildGames, a subDAO is “akin to a specialized, miniature economy that interacts with a larger, all-inclusive economy, which is the DAO itself.” Mechanically, SubDAOs are operated with their own governance rules but can borrow assets from the ParentDAO. SubDAOs are an innovative development in decentralized governance that open up the possibilities for experimenting with governance models.
A DAO could operate a hybrid or mixed governance model by operating a ParentDAO with more centralized ownership & governance than its SubDAOs. SubDAOs could leverage the benefits of centralization of risk management and security, while experimenting with decentralized governance models in their communities. SubDAOs also enable community members to contribute to and work in the communities that best align with their values and interests.
While SubDAO structures are in the early phases of experimentation, they present an exciting potential future model for a hybrid DAO structure.
4. Create discrete pods of agents that can engage with the traditional world and perform key tasks with greater autonomy, when necessary
As the BarnBridge DAO discusses, there are certain situations where greater autonomy of a certain group of community members is needed to complete a key task. For example, BarnBridge posits a situation where a team member needs to enter into an NDA with a traditional finance firm that has to comply with a different regulatory environment, for a partnership that can’t be shared with the community. The compromise they proposed to their community is as follows:
BarnBridge’s solution offers one potential idea for situations where core team members must be able to exercise greater autonomy in the long-term interest of the DAO and its community members.
There are other situations where a separate set of “agents” with their own multi-sig might be necessary. In particular, the management of legal matters. As Gottlieb, Isaacs and Wang note in their piece, a DAO could use decentralized voting to decide which counsel to use. However, the DAO “must be able to effectively and nimbly direct its counsel on a day-to-day basis, receive reports from legal counsel and preserve the confidentiality and privilege of attorney-client communications.” The proposed solution here is also delegation, and potentially this delegation could follow a similar structure to BarnBridge DAO’s solution for engaging with Traditional Finance institutions. However, as per Gottlieb/Isaacs/Wang, DAOs would need to find ways of communicating outputs of this delegation while ensuring certain discussions/communications are kept confidential.
While today there are some important issues faced by those building and participating in fully decentralized organizations, I am excited for a future where these issues are solved either through hybrid orgs or through other mechanisms that enable organizations to transition over time to pure decentralized governance.
If you are part of a DAO and have observed any of these tactics (and seen them work/fail), have great examples of some of these tactics used, or I am missing strategies you have seen work, please comment and share. This is intended to be an open-sourced document that will receive input from the community and improve over time based on that input.
Looking forward to hearing what you think & have seen!