DAO Governance Tools

From P2P Foundation
Jump to navigation Jump to search


Kei Kreutler:

"At a closer look, effective DAOs start behaving much more like networks of teams, like the MONDRAGON Corporation network with 100 affiliated cooperatives, rather than the loosely coordinated swarm intelligence that they might appear as from a distance. Inspired by Rawson’s analysis, we can roughly sketch three layers of a DAO:

  • Token: Multi-organizational networks aligned by token ownership
  • Teams: Teams, guilds, and squads represented by token ownership
  • Missions: Missions, milestones, and raids financed by token ownership

From these layers, a heterarchical network emerges, meaning an organization that possesses the ability to be ranked in multiple ways.

In ecosystems that prioritize distributing ownership, tokens encourage DAO networks to be managed by their members. Tokens, teams, and missions are not confined to the quasi-institutional boundaries of a single DAO but can and, in order to meaningfully decentralize the control network, should be represented by token ownership in multiple DAOs. An ecosystem not unlike the control networks of transnational corporations emerges, but importantly, one without monocentric command and with reduced transaction costs across different levels of trust. As Rawson writes, “As long as the collective memory freely circulates within a given [DAO network], discovered solutions to problems can be reused.” When we view DAOs as multi-organizational networks aligned by token ownership, the purpose of DAO tooling becomes not only to support the operations of one team but to facilitate collaborations across many. Funded by PrimeDAO, DAO-to-DAO (D2D) collaboration mechanisms appear the most forward thinking work along these lines, which may eventually overtake traditional business-to-business (B2B) products. A new squadlike entity emerging from stealth mode, the Gnosis Guild team embraces a similar emphasis on DAO-to-DAO tools, with a new constellation of DAO tools launching in early August.

Revisiting the promise of DAOs, their potential to incorporate deeper practical knowledge in governance does not mean that decision making must involve an ever-larger number of members in every proposal, but rather that within a DAO network, teams who possess the most relevant expertise can easily share it with the ecosystem. When we view DAOs as constellations of teams, not monoliths, DAOs become networks to allow collective memory to flow freely."



Kei Kreutler:

Returning to their origin, DAOs today resemble The DAO in their emphasis on open participation and economic value creation, while their culture has shifted more toward specific niches and social connection. In the examples above, one term has been left intentionally vague: governance. Many DAOs today use the lightweight Snapshot platform for governance (13). On Snapshot, DAOs each have a space to create and vote on proposals. For example, both PleasrDAO and PartyDAO have a Snapshot space, on which they hold public votes for collective decisions. Snapshot weights votes by the amount of DAO-specific tokens an address holds, such as $PEEPS tokens in PleasrDAO.

The topic of governance has its own history within the crypto ecosystem that won’t be fully expounded upon here. Notably the initiative MolochDAO, which uses the classic font Papyrus and heavily references gaming guilds, rekindled the fire of decentralized governance after The DAO hack. MolochDAO went on to inspire a legion of new DAOs, many direct forks, in its wake.

This essay’s history of DAOs is far from complete, as other projects like Aragon, Colony, DAOhaus, and DAOstack continue to develop their platforms for DAOs and modular initiatives like Block Science and Commons Stack arise. These projects offer DAO tools supporting many governance mechanisms. However, a prehistory of DAOs is also incomplete without its less frequently mentioned relation to Platform Cooperativism."



DAO's as Institutionally Innovative

Overview of research directions: [1]

Organizational science

Editors: Rolf Hoefer, Ellie Rennie

Contributors: Rolf Hoefer (Cultur3), Seth Frey (UC Davis, Metagov), Sarah Hubbard (Harvard), Tony Douglas (Stanford, Convex Labs, DAO Research Collective), Scott Moore (Gitcoin, Metagov), Michael Muthukrishna (LSE), Mason Youngblood (Stony Brook University), Michael Price (Santa Fe Institute), Ellie Rennie (RMIT University), Alexia Maddox (La Trobe University), Anna Weichselbraun (University of Vienna, Berggruen Institute), Kelsie Nabben (RMIT University), Primavera de Filippi (CNRS, Harvard University), Tara Merk (CNRS, Metagov)

DAOs constitute a fascinating, rapidly emerging organizational form in the wild. In a prototypical DAO, on-chain software is central to the organizing process and every member is able to directly influence the execution of key organizational decisions. Blockchain infrastructure also offers systematically rich, granular, and longitudinal data.

DAOs represent one of the most exciting empirical phenomena in organizational science. As of 2022, DAO participants in over 4000 active DAOs collectively managed over $20 billion USD worth of assets in their treasuries. DAOs are organizations operating in all kinds of industries, organizing around common goals ranging from everything in finance to gaming, politics, culture, arts, and civil society. One illustration of the potential of DAOs and other Web3 developments comes from reflecting on Web2’s impact on organizational scholarship. The Internet’s second wave made new, unprecedented organizational forms possible. For example, Wikipedia demonstrated the scalability of flat organizations and the ability of well-designed technological mediation to organize the incremental contributions of millions of people (Puranam et al., 2014). Poorly explained by existing organizational thought, Wikipedia and other examples motivated the development of Benkler’s theory of peer production, which provided an intellectual foundation for a new generation of organizational scholars (Benkler et al., 2015).

What new organizational forms will DAOs permit, and what flaws will they reveal in how we conceptualize organizational possibilities? New mechanistic corrections for the shortcomings of decentralized organization will further advance the frontier that was opened by Web2. New forms of collective ownership will blur the line between ownership and management regardless of whether they embrace decentralization. Instantaneous algorithmic design and incorporation of ephemeral businesses will blur the lines of Williamson’s transaction cost rationale for firm versus market exchange (Williamson, 1975). For their potential to both build on and extend a century of organizational thought, DAOs merit close attention from organizational scholars.

In the following we neither claim nor attempt to depict the breadth and depth of overlap between organizational science and DAOs. Instead, we pick a few topics we believe are interesting and fruitful explorations for those interested in DAOs and organizational science.

Organizational imprinting

Summary: We still do not fully understand why and how emerging organizations come to adopt their particular social structures and strategies. DAOs, which generate a significant amount of granular, longitudinal data through on-chain processes, can provide valuable insights into the evolution of social structures and practices within organizations, offering opportunities to explore dynamics, histories, and the impact of elements like smart contracts, incentive schemes, and forks.

A seminal piece in organizational theory is Stinchcombe’s 1965 chapter on Social Structure and Organizations (Stinchcombe, 2000). Over time, this chapter has come to be known for highlighting an observation unique to organizational theory: the phenomenon of organizational imprinting. Organizational imprinting describes the observation that organizations founded at one time tend to have a different social structure from those founded at another time. Once an organization forms its social structure, its social structure tends to persist for an extended period.

While Stinchcombe’s observation is widely taken-for-granted, why and how emerging organizations come to adopt their social structures and strategies has been underexplored (Hannan et al., 1996; Johnson, 2007, 2009). In this context, DAOs are interesting because they represent a new, fledgling organizational form. As a novel type of organization, DAOs are in the process of forming and adopting the social structures and strategies that will, according to past organizational imprinting research, persist for a long period of time. This suggests that DAOs represent a fertile ground for researchers to understand the formation phase of organizational imprinting, including when, why, and how an organization does or does not become imprinted.

Evolutionary social science

Summary: Evolutionary social science applies ideas from biology and evolution to questions in the social domain. DAOs represent a fertile new domain for developing and testing some of these ideas due to the availability of data, their use of technical infrastructure that can be precisely characterized, and the speed at which they evolve.

Historically, much research on organizational imprinting has drawn inspiration from biology and evolution (Lorenz, 1935) with concepts such as an organization’s “DNA”; organizations evolving through adaptation and selection pressures (Zhong et al., 2022); and complex patterns of mutualism (TeBlunthuis & Hill, 2022). More broadly, the field of evolutionary social science has a long history of applying evolutionary ideas to questions of social evolution, cultural evolution, and organizational evolution. DAOs represent a fertile new domain for developing and testing some of these ideas.

A research program applying evolutionary ideas to DAOs would need to include at least three components: conceptual mapping, data and tooling development, and comparative case studies.

Conceptual mapping. To apply evolutionary methods to a field, we need to define a set of evolutionary primitives, in particular mechanisms for variation, transmission, and variation reductive (which, if adaptive, should be selective). What is the mechanism by which variation is created within DAOs? Variation is produced sometimes in the act of transmission—is that true for DAOs? For any evolutionary system you want high fidelity in transmission, but how much actual transmission is there? Finally, what are the mechanisms for selection? Are there informal mechanisms, ways that communities can change, or fork to change? Note: having “genes” or “building blocks”—e.g. granular privacy primitives—can be useful but it is not essential for this modeling.

Data and tooling development. How does the infrastructure of a DAO allow a working evolutionary social scientist to engage with either individual DAOs or the broader ecosystem? Given the issues with off-chain data, what affordances and tooling exist to allow researchers to understand and parse what is going on within DAOs? What particular aspects of DAO data are evolutionary social scientists particularly interested in?

Comparative case studies. We want to compare DAO evolution to existing examples within social evolution, for example the evolution of churches (Finke & Stark, 2005), of constitutions (Rockmore et al., 2018), within coding competitions (Miu et al., 2018), and of other online communities such as subreddits (C. Tan, 2018).

Within the scope of such a research program, there are a number of different open questions:

Forks. Forks are phenomena where an organization splits from another organization yet retains the exact same history as the previous organization (Berg & Berg, 2017). Due to its public nature, the data and contracts that define a DAO can be easily forked as a technical matter. Indeed forks have happened, e.g. SushiSwap from Uniswap or during the attempted takeover of Steemit, and the threat of forks represent a significant check on bad behaviors by founders and other powerful members of a DAO or blockchain.

Infrastructure for merging. The US federal government is a way for states to merge. The EU is another model for merging. Australia is a very different model. In the infrastructure of a DAO (or a blockchain), what are the mechanisms that would allow you to explore new possible configurations and to possibly incentivize merging of communities?

Cultural evolution. How does the new technical infrastructure of a DAO enable and constrain the variation, transmission, and variation reduction of cultural practices within the DAO? How can these infrastructures translate into non-Web3 contexts?

Simulating cultural evolution. Ideas from cultural evolution could be particularly useful in simulating DAO behavior. There is a suite of methods in cultural evolution for simulating how conformity bias, prestige bias, and payoff bias interact with one another in extremely complex ways, with norms co-evolving with these biases. Considerable time and expense has been applied to collect data from social networks that can then be applied to these models, whereas these methods could be applied directly to DAOs based on already-extant data. Such simulations could eventually be incorporated into DAO design.

Theory of cultural evolution. Insofar as they represent interesting examples of self-governing communities, DAOs could inspire new mathematical treatments of agents’ ability to self-determine and create their own games, as opposed to traditional treatments that operate within a tight framework of phenotypes and behavior.

Cultural traditions as cognitive tools. Cognitive tools are artifacts or processes that facilitate cognition. These are not just physical things but also “ways of seeing the world”, i.e. cultural traditions. For example, “mountain calendars” from Mexico (Ezcurra et al., 2022) or “knuckle mnemonics” for months of the year. What kinds of cultural traditions operate as cognitive tools within DAO governance and DAO operations? How can existing infrastructure support such cultural traditions?

Neo-institutional theory

Summary: Building and maintaining legitimacy is a key concern for traditional institutions but especially in DAOs, which often govern through flat hierarchies and informal cultural norms. Neo-institutional theory shifts the focus of research on legitimacy to its communicative aspects and allows for a deeper understanding of how organizations, including DAOs, achieve legitimacy through language and communication rather than mere diffusion of practices. Thus, neo-institutional theory can help practitioners and DAO managers navigate the emergence, intercultural participation, chaos, and complexity that characterize DAO environments.

For decades, institutional theory and its variances have been a dominant research stream within organizational science. A core concept in institutional theory is the concept of legitimacy (Berger & Luckmann, 2011; DiMaggio & Powell, 1983; Friedland & Alford, 1991; Powell & DiMaggio, 1991; Suchman, 1995; Zucker, 1977) (Berger & Luckman, 1967; DiMaggio & Powell,1983, 1991; Friedland & Alford, 1991; Meyer & Scott, 1991; Suchman, 1995; Zucker, 1987). While much work has measured legitimacy as a function of the successful diffusion of organizational practices, the process by which organizations such as DAOs achieve legitimacy is fundamentally communicative. Consistent with the phenomenological tradition in institutional theory (Scott, 2014), legitimacy is “built upon language and uses language as its principal instrumentality” (Berger & Luckmann, 2011, p.64).

A focus on legitimacy as a communicative phenomenon is interesting for two reasons. First, from a theoretical perspective, adopting the view of legitimacy as a communicative concept avoids the suggestion that the diffusion and adoption of organizational structures, practices, and strategies is evidence of legitimacy. Organizational structures, practices, and strategies can be sparsely adopted yet legitimate, and they can be widely adopted yet not legitimate (Green, 2004; Zilber, 2002). Focusing on language also resonates with the linguistic turn in institutional and organizational science (Alvesson & Kärreman, 2000).

Recent research efforts at the intersection of entrepreneurship and institutional theory, such as the work on cultural entrepreneurship (Johnson, 2007; Lounsbury & Glynn, 2001), have started to examine the ways that the legitimacy of organizations is a function of language, suggesting an emerging body of work that scholars can build on and extend. Second, a focus on legitimation as rooted in communication is interesting from practitioners’ perspectives (Buterin, 2021; De Filippi, Mannan, Henderson, et al., 2022). DAOs are internet-native organizations. Coordination is managed digitally in a fast-paced world. When environmental change is high, organizational systems need to adapt quickly, and this work is typically facilitated by people who focus almost exclusively on coordination as opposed to execution —and that is the role of management, or, put into the language of DAOs, that is the role of community managers and delegates (Burton et al., 2017).

Persuasion in DAOs often happens more with stories rather than tables, words rather than numbers, beliefs rather than facts. Managers must become skilled at persuasion with stories, words, and beliefs. Persuasion increases social stickiness, and social stickiness is what keeps organizations as communities and social networks together in the absence of tight, hierarchical structures and vague incentive structures. The current mintage of DAOs is characterized by emergence, intercultural participation, chaos, and complexity. Managing these DAOs requires managers to become skilled communicators. Those who lean on the knowledge gained from research efforts in institutional theory such as those on cultural entrepreneurship may benefit tremendously.

Organizations as complex adaptive systems

Summary: Understanding DAOs as complex adaptive systems (CAS) presents opportunities to study properties such as path-dependency, sensitivity to initial conditions, emergent changes, and episodic shifts. Researchers can employ grounded theory and comparative case study approaches to advance empirical understanding and generate theories around organizational evolution. Additionally, the unique data properties of DAOs offer avenues for fascinating studies on social networks, dynamic tie evolution, and the use of agent-based models to project and understand organizational dynamics over time. This stream of organizational research can further help to address questions of how and why productive self-organization emerges in organizations and DAOs.

Many DAO participants believe DAOs offer new opportunities for organizational systems to self-organize. There is a real opportunity for academic work that may help practitioners understand how, when, and why productive self-organization emerges. Much existing work is rooted in views of organizations as complex adaptive systems evolving at the edge of chaos. Indeed, the contributions of scholars such as Herbert Simon and James March were as fundamental for the development of complex systems as organization science (Simon, 2019), while Elinor Ostrom explicitly bases her foundational concept of institutional diversity on the proto-CAS thought of cybernetics (Ostrom, 1995). Interestingly, empirical research in this domain has been slowed down by the lack of appropriate empirically-grounded data and methods to test researchers’ theoretical hypotheses. DAOs present an opportunity to not only test these hypotheses but also to inform and generate new ones.

Complexity, in brief, views systems such as organizations as wholes that are more than the sum of their parts and even as archetypes for the multi-scale structure that characterizes emergent complexity (Ostrom, 1995). The literature on complexity highlights properties such as path-dependency, sensitivity to initial conditions, emergent (uncertain but not random), and episodic changes (Beinhocker, 2007). Of particular interest to practitioners in the DAO space might be how to manage complex organizations while staying true to emergent self-organization (Ostrom, 2009).

DAOs are a fertile ground with plenty of exceptional data to study self-organization. On-chain data does not suffer from left-censoring; however note the learnings from the Curve wars regarding manipulation of the governance that occurs on-chain. Certain datasets have perfect data, obviating the need for sophisticated statistics to address gaps. The recent explosion of DAOs also offers sufficient freedoms of observation. This presents rich opportunities. On one hand, grounded theory and comparative case study approaches would advance our empirical understanding of DAOs. It may also generate theory around organizational evolution.

Models of organizational change as punctuated equilibrium traditionally invoke change as developing from a growing discrepancy between an organization and its environment. The causal mechanism typically invoked is organizational inertia. Yet considering organizations as complex adaptive systems suggests alternative mechanisms. For example, organizations may change via punctuated equilibrium because patterns of both small and large changes over time can often naturally lead to a pattern of change in the form of a punctuated equilibrium.

Moreover, a unique research opportunity rests in the data properties around DAOs. Empirically inclined researchers may create fascinating new studies that leverage this data:

Researchers may analyze social networks of agents not only at a given point in time, but dynamically as these social networks evolve over time, and agents and their organizations begin to co-create each other (Zhong & Frey, 2022). Much research in social networks is based on the presence or absence of connections between some or the other actor. Data related to DAOs allows researchers to see how ties dynamically evolve. This can be similarly studied at the dyadic, triadic, cluster, or network levels. Researchers viewing organizations as complex adaptive systems can easily measure, more easily than ever, how a focal actor’s behavior in the current time period affects the behavior of other actors in the next time period.

An even more fascinating approach would be to train or condition agent-based models on real-life data points. DAOs may permit faithful implementations and rigorous tests of classic computational models of organizations, from the anarchic garbage can model (Cohen et al., 1972) to Carley’s ORGAHEAD of organizational adaptation (Carley, 2000) For example, an agent-based model’s first iterations modeled using real data and the revealed dynamics could continue projecting an organization’s evolution in the next few iterations. Such models would approach scientific work with practitioner work, both tremendously benefiting from each other via superior, shared understanding and greater empirical and theoretical groundedness. Organizational methodology in the era of complete data Summary: The questions scholars ask about organizations tend to be a function of the limited data they have access to. As DAOs provide complete data from the level of individuals to the level of populations of organizations, the completeness and transparency of this data promises to advance organizational scholarship, particularly as organizational scholars increasingly work to reconcile and integrate disparate research frameworks.

Organizations are difficult to study. They are typically too big, dynamic, and unobservable to provide the elements necessary for controlled scientific inquiry. DAOs correct for this by being highly instrumented, entirely transparent organizations that offer systematically rich, granular, and longitudinal data. The comprehensiveness of DAO data is atypical of what academics have encountered to date. As a result, DAOs represent not just a novel empirical setting but an opportunity to develop and test novel empirical methods that were previously impossible to apply.

As DAOs rely on on-chain processes for membership, governance, and operations, they generate a wealth of granular, longitudinal data about the evolution of their social structures and practices. Often this on-chain data does not suffer from typical left-censoring or sampling issues, and, apart from community moderation, other public data from governance and operational interactions is uncensored. We have already discussed the longitudinal tracing of how social structures evolve via on-chain elements, dynamics, and histories such as smart contracts, incentive schemes, and forks, but many other opportunities remain.

Looking ahead, this data could be used to power multiscale studies that bridge (or challenge) Hannan and Freeman’s five levels of organizational analysis: members, subunits, individual organizations, populations of organizations, and communities of populations of organizations (Hannan et al., 1996). This is important because the major paradigmatic differences driving contemporary organizational scholarship are more a function of data availability than any epistemological differences. An ideal example lies in firm- versus population-level theories of organizational survival. Frameworks like the firm-level “resource-based view” rely on behavioral data of members of organizations to discover internal determinants of organizational success. By contrast, frameworks like the population-level organizational ecology view rely on data about populations of interacting organizations to reveal environmental determinants of success. While scholars acknowledge that both must be true (Goll & Rasheed, 2005; Leiblein & Miller, 2003; Volberda et al., 2012), the difficulty of collecting large, complete, multi-scale datasets has made it impossible to understand how internal and environmental theories of organizational performance interact and complement each other. From this perspective, DAOs promise to be the substrate upon which organizational scholars develop a unified multi-scale conception of firm success and even emergent social order.

Looking even further ahead, it is important to recognize that the people who start and operate organizations are themselves organizational scholars engaged in the challenge of building a social system capable of learning and adapting (Senge, 2006). Classic work by March emphasizes the spotty and imperfect nature of trying to theorize about a single organization while situated within it (March et al., 1991). With their automation, instrumentation, standardization, and transparency, DAOs promise to transform organizational learning by making it easier for members to automatically register and learn from each others’ mistakes both within organizations and across many organizations. Although it has long been recognized that organizational learning can happen at the level of communities of populations of organizations (Baum & Ingram, 1998), even lessons at that scale tend to be anecdotal, and lack the automatic, systematic comprehensiveness of potential future DAO-to-DAO organizational learning systems.

Organizational ethnography

Beyond analyzing DAOs through their on-chain and quantitative data, qualitative research methods are widely employed throughout organization studies. In this section, we make a case for ethnography, a specific type of qualitative research method, as a means to surface DAO-based social phenomena and practices that may push the boundaries of existing disciplines.

Ethnography is a qualitative approach in which the researcher is immersed in the phenomena and community they are studying. It is typically used to make visible the invisible, surface new and recurring questions and themes, and challenge normative assumptions with an eye to identifying discrepancies that exist between what communities say they do versus what they actually do. The process involves producing “thick description” (Geertz, 1973), meaning an interpretative account not just of actions, but the meanings, symbols and structures that inform behaviors. As ethnography looks to “first-hand” knowledge rather than relying exclusively on second-hand accounts (Hine, 2015), researchers will spend significant time with the people, communities, and contexts they are researching to understand their language, design, rules, and customs. In addition to observation field notes, ethnographic data may include interviews, focus groups, collecting data across multiple media sites (Atkinson & Hammersley, 1998) or even computational data (Maddox, 2020).

In the case of DAO research, the ethnographer’s task lies in documenting and understanding the social dynamics underpinning a particular context and the interactions among both human and non-human actors constituting a field site. A DAO community is not always the thing being studied; the ethnographer may be tracing various phenomena such as values, standards, infrastructures, and automated policies by observing how they occur in and through DAOs.

The ethnographer may choose to extend upon traditional ethnographic research methods by including the study of computer mediated social interactions, the study of online communities and the use of digital technologies in data collection and analysis (Abidin & de Seta, 2020; Pink et al., 2015). For instance, ethnographers studying in DAOs may need to follow both human and non-human actors (Seaver, 2017), and make sense of their interaction with each other and the community’s wider social and material context (Latour, 2007). Specific techniques such as situational mapping, arena mapping, and positional mapping (Clarke, 2003; Clarke et al., 2016) may be used to synthesize empirical observation without excluding technical actors, infrastructural components, discursive dynamics or the broader context.

Conducting ethnographic research in DAOs and Web3 comes with specific challenges. Firstly, ethnographers are required to define, access and make sense of the deluge of data gathered through their field site, including online forums such as Discord servers. In the context at hand, this requires us to ask questions such as: Where is the DAO? Where, when and how can we see it, and where not? Secondly, while ethnographers are relatively free to choose the specific techniques most appropriate to making sense of their data, it is not always easy to get informed consent from all actors being observed in a DAO. One example of an attempt to overcome this challenge is the Telescope consent bot, developed by Rennie et al. (Rennie et al., 2022) to automate parts of the consent process for data collection within Discord chats.

To date, researchers have used ethnographic methods to explore the informal aspects DAO governance (DuPont, 2017; Shorin et al., 2021), cultural dynamics and beliefs in cryptocurrency communities (Brody & Couture, 2021; Faustino et al., 2022), activism in darknet marketplaces (Maddox et al., 2016) and the use of Web3 technologies as a form of ‘self-infrastructuring’ (Nabben, 2023). Researchers employing ethnographic methods have made valuable contributions to DAO science by revealing unformalized and somewhat messy off-chain governance dynamics, anchor on-chain components within specific context and meaning to understand their mediating capabilities, and by developing new participatory methods to involve DAO practitioners in the research process. Ethnography is therefore a promising approach for identifying emergent ‘open problems’ in DAO science."