Thank you for this Spaydh, a much needed conversation we as a community need to double down on. I think most of us can agree that, as you highlight, one of the core strengths of the Cosmos community is its willingness to engage in deliberation around major strategic proposals and in times of diverging opinions. On the other hand, this relative decentralisation has also, imo, left more to be desired in terms of a clear direction and structure. Coming into the ecosystem, it is hard to tell who’s working on what etc, and viewing the Hub as an organisation it is rather loosely scattered across a few teams. I believe this has altogether hurt the Hub’s strategic performance and left somewhat of a question mark hanging over its value proposition and purpose. It is also the case that governance is often technically complex and time-consuming. So the lure of free-riding can be tempting when there’s no direct reward, but perhaps even worse, also half-arsed participation; as voters are put off by the demands of active governance but remain indirectly incentivised to maintain good participation numbers, logically leading to less informed decision-making. Given what I perceive as a growing interest around the idea of subDAOs, greater recognition of DAO DAO and the possible introduction of CosmWasm on the Hub, not to mention the seemingly successful use of subDAO in Neutron governance, I will here visit the idea of why subDAOs could prove useful on the Cosmos Hub.
Strategic vs. Operational Governance
There is research, for example by Zhao et al. (2022), that suggests that it could be performatively beneficial for DAOs to make a distinction and divide responsibilities between “strategic” and “operational” governance. Said study looked at how voting task division, task allocation, reward distribution, and information provision affect platform performance in the context of MakerDAO (where strategic tasks mainly included collateral onboarding, system upgrades, and personnel changes; and operational as including adjustments of various parameters in the business process) and found that “strategic decisions arrived at through voting have a positive impact on platform operational performance … whereas operational decisions resulting from [universal] voting have a negative impact.” Put differently, when the wider community was involved in strategic decisions and operational tasks were handled by a smaller group of experts, Maker performed better. Clearly, educational information plays an important role in strategic voting and helping voters make better decisions, which can then positively feed into operational performance; while higher levels of voter-task matching can alleviate the negative effect of operational voting tasks on product quality, as there is, rather unsurprisingly, a correlation between subject specialisation and improved decision-making. Interestingly, the share of large shareholders among total voters worsened the negative impact of operational voting tasks on product quality (in this context, DAI deviation), which the authors argue may be the result of whales choosing short-term profitability over long-term sustainability. Perhaps subDAOs could serve to address this as well to some degree.
Similarly, a study by Chen, Richter and Patel (2021) found that decentralisation (the extent to which power and control in governance structures and decisions are allocated to developers and community members) has an inverted U-shaped relationship with protocols’ market capitalization, developer attention, development activity, and social media followers; concluding that “the midrange of decentralization is more likely to be associated with higher levels of market performance”. Though this might appear quite logical, as strategic governance tends to evolve slowly, whereas operations move quickly to adapt to changes in the environment. But it is also the case that operations without clear structure usually become implicit and informal with accountability suffering as a result, while governance without operations (as a looser form of social group) tends to struggle with goal setting and achievement (related to why excessive degrees of decentralisation can become slow(er) and counterproductive). Work on digital platforms shows that those who successfully apply semi-decentralised governance structures tend to (see e.g. Kyprianou, 2018; Reischauer & Mair, 2018):
- involve community members in goal-setting and decision-making.
- leverage creativity and diversity within the community to strengthen governance and performance.
- have key organisations/individuals supporting and guiding governance (deliberation) for unity and broader consensus-making.
- can achieve a relative balance of power and maintain effective value creation and fair value distribution.
Similar results are found in other studies on centralisation vs decentralisation in various governance contexts (e.g. Arcuri, & Dari-Mattiacci, 2010; Dougherty et al., 2022). Many in the crypto community will shun the idea of clustering expertise into groups like these and are, in all fairness, right to do so given how entrusted power has time and again been abused throughout history to expand arbitrary powers. To this point, it should be stated clearly that any future subDAOs on the Hub should serve as an extension of token holder governance, with the purpose of directly working towards the goals that have been set by the wider community. Focusing on broad community goals as the raison d’être moves us away from decentralisation for the sake of decentralisation, and instead towards a framework of adaptable, resilient and sustainable continuation. Blockchains are fluid and polycentric with multiple overlapping powers, very much experimental and constantly evolving. So a governance system should try to accommodate this while providing some structure which appears beneficial. The idea of depending on token holders to vote on all matters of governance is asking a lot, and its effectiveness continues to be questioned.
The Possible Advantages of subDAOs in Governance
In contrast to real-life politics, we have on-chain transparency and decentralised control over both funds (community pool) as well as subDAO membership and power distributions to apply in our favour. Obviously, some of the operational components related to the overarching strategic goals set by the community will lend themselves more naturally to a subDAO—in my mind including marketing and business dev, perhaps SDK/tech maintenance, consumer chain onboarding and support—while others won’t. Indeed, subDAOs exist for different reasons, some more obvious than others: speeding up governance, addressing voter apathy and leveraging specialist knowledge. Where decisions are highly technical and niched knowledge-wise, for instance, it theoretically makes more sense to have a smaller and tighter subgroup rather than a broad community making (less informed) decisions.
Again, it should also be stated clearly as a reminder that we’re here (in part) because we’re trying to build something different, something that is fundamentally decentralised. So one cannot overstate the importance of sticking to this goal. Having internal structure and core teams clearly play an important role from a performance perspective, but so does wider community collaboration and participation, not least ideologically. The broader aim here would be to gradually work towards a semi-autonomous core of subDAOs under the control of ATOM holders that maintains the decentralised, innovative ethos of permissionless open-source, yet provides structure and (administrative) organisation that seeks to apply “extreme focus on growing its excellence in decision-making” and make “the Cosmos Hub’s governance more reliable and predictable from the outside, lowering platform risk.”
In many ways, this would resemble a network organisation. Interestingly, the literature on network organisations has highlighted their positive performance in contexts where efficiency and flexibility are central to the overall process (e.g. van Alstyne, 2009)—which one could argue fits the fast-paced innovative and experimental environment of blockchains pretty well. The same literature also emphasises transparent information sharing and leveraging expertise (for educational deliberation and organisational tasks) as key assets in network organisations: combining knowledge while enhancing the links between information flows to better manage governance complexities. Using subDAOs on-chain as a more formalised organisational structure could possibly facilitate the former while assisting the latter, and thus strengthen the alignment between contributor skills/knowledge and the needs/goals of the community. Theoretically, if subDAOs can be used to channel expertise with greater efficiency and operational quality through streamlined processes and more informed decision-making, while allowing the community to focus on oversight and directional strategy to greater effect (as more time and effort can be dedicated to this due to the alleviation of operational day-to-day burdens), that could set the stage for “desirable benefits to ‘customers’ of the Cosmos Hub’s services”.
Still, given that there already are a number of core teams working on different aspects of the Hub, one might conclude that we’re already operating some kind of informal subDAO structure. While somewhat true, if one takes an external view of the Cosmos Hub community as an organisation, these pieces have to more or less be sporadically put together as best one can, as most core work is scattered across a few entities here and there. Formalising an actual subDAO organisation on-chain could bring much-needed structure, greater clarity regarding roles, responsibilities, and work opportunities (perhaps encouraging talent retention), and most importantly, arguably better tools for the community to actually manage such delegated work—altogether contributing to “the Hub’s ability to make and enforce excellent decisions in a decentralized, efficient and reliable way”. Over time, if deemed desirable, the subDAOs could be used to decentralise operations even further, from core team to core subDAOs (though such coordination is another challenge), and in so doing allow other teams and skilled community members to contribute to the core roadmap work.
In contrast to ATOM 2.0, I am unsure about the idea of handing over strategic decision-making to an Assembly or any other concentrated group, and think it’s probably an even worse idea to have such an Assembly be made up of delegated members from each subDAO (Councils in ATOM 2.0). You want to separate different powers, not encourage their consolidation. As I’ve said, I think groups/subDAOs should be used to carry out the strategic work and direction agreed on by the community.
As I see it, the key to using subDAOs (beyond the key back office procedures) is to allow for autonomy, establish transparent communication procedures, and retain ultimate power in the hands of token holders; including veto optionalities and control over subDAO memberships. It’s about leveraging the specialised expertise that exists within the community while keeping the keys in the collective hands of token holders, so to speak, to preserve fundamental decentralisation and the reins over delegated centralisation, where autonomy and control co-exist to achieve better results. On this question of decentralisation vs centralisation, it really isn’t about choosing between the two, but rather seeking to balance the two together in an effective, and democratic, manner.
One idea I find interesting is the use of optimistic governance to manage the Hub - subDAO relations, which seeks to significantly lower the number of proposals token holders have to vote on while instead focusing more on active oversight. Optimistic governance essentially maintains veto rights for token holders by allowing subDAOs to pass proposals unless the community decides to interfere during a set contestation period by voting above a set rejection threshold, reasonably lower than the current approval quorum. Such a system would rely on consistent downstream deliberation (i.e., justify how and why changes are implemented and efforts carried out).
In addition, having an on-chain subDAO structure in place with systematic policies and effective administrative practices (such as clear recruitment and onboarding procedures) to support it, that remains clearly visible to the community (via a dashboard or likewise), could not only improve oversight and accountability but also go some way towards improving the budgeting and talent retention aspects as well, and with some flexibility at that. Whether a subDAO is of a more permanent or ad hoc nature, budgets could be set yearly, seasonally or in other types of sprints, paid (in part) retroactively, and/or tied to measurable goals and objectives; ideally set by the community in one way or another. Lido and Gitcoin are good examples to learn from.
Regardless of the internal talent and expertise that exists across the community, it may not be effectively utilised if not organisationally channelled. Sometimes there seems to be this assumption that networks are self-organising and autonomous, but in many cases, they must be created and fostered. SubDAOs could provide those clear working groups to channel knowledgeable community members, particularly those that do not belong to one of the core teams (and in so doing, decentralising).
Closing Thoughts
In closing, subDAOs or no subDAOs, the aim should be to do something different that keeps with the core ethos of decentralisation, autonomy, and so on, in a way that promotes high-quality decision-making, while being wary of replicating old legacy hierarchies and related inequalities. Obviously, these are more or less some thoughts written down, but if there’s interest, I with the help of my colleagues at Simply Staking would be happy to write up and/or collaborate on a more detailed plan for what an initial subDAO structure might look like.
Mentioned references
Arcuri, A., & Dari-Mattiacci, G. (2010). Centralization versus Decentralization as a Risk‐Return Trade‐Off. The Journal of Law & Economics, 53(2), 359-378.
Chen, Y., Richter, J. I., & Patel, P. C. (2021). Decentralized Governance of Digital Platforms. Journal of Management, 47(5), 1305-1337. https://doi.org/10.1177/0149206320916755
Dougherty S, Lorenzoni L, Marino A, Murtin F. The impact of decentralisation on the performance of health care systems: a non-linear relationship. Eur J Health Econ. 2022 Jun;23(4):705-715. doi: 10.1007/s10198-021-01390-1. Epub 2021 Oct 17. PMID: 34657202; PMCID: PMC8520686.
Kyprianou, C. 2018. Creating value from the outside in or the inside out: How nascent intermediaries build peer-topeer marketplaces. Academy of Management Discoveries, 4: 336-370.
Reischauer, G., & Mair, J. 2018. How organizations strategically govern online communities: Lessons from the sharing economy. Academy of Management Discoveries, 4: 220-247.
Van Alstyne, Marshall W., The State of Network Organization: A Survey in Three Frameworks (March 26, 2009). Journal of Organizational Computing, Vol. 7, No. 3, 1997, Available at SSRN: https://ssrn.com/abstract=1368930
Zhao, X., Ai, P., Lai, F., Luo, X., & Benitez, J. (2022). Task management in decentralized autonomous organization. Journal of Operations Management, 68(6-7), 649-674.