In this paper we propose conceptual foundations for defining a DAO community and propose how a DAO community’s health can be measured.

Specifically, we seek to answer the following questions:

1. What are the features of a community?

We begin very broadly, explaining how we understand what a community is, and what its central features are. We conclude that it is a system of individuals at the micro-level. However, it has several subsystems and communities themselves belong to the larger ecosystem. This nested systems or network science perspective on communities guides our work, in addition to a social identity perspective. We also note that there is a dark side to communities (e.g., negative views on outsiders, group-think, etc.)

1. What characterizes DAO communities?

After agreeing on what a community is, we explore what defines  DAO communities so we may compare them to other forms of community. We use several guiding questions to address this question, such as by whom it is influenced and controlled, and how cohesive the community is.

1. What is DAO Community Health?

Having asserted that DAO communities have specific properties that differentiate them from traditional communities, we delve into the question of what does it mean when a DAO community is healthy? How can you distinguish health from an unhealthy community?

In line with our system view of community, we understand community health to be regenerative and define it as "The state of being, interaction, and integration of individuals and nested subsystems of a DAO Community as they work to achieve individual and collective goals".

1. How can we evaluate a DAO community’s health?

In this section, we explore how we can operationalize community health and go from a definition to measurements. Using the analogy from medicine, we propose that there are several vital signs that can be measured to create a diagnosis of a community’s health, and suggest engagement, participation structure, and sense of community. We acknowledge that these vital signs measure primarily the symptoms of a community’s health. To better understand the causes we identify specific avenues for further research and the need for close collaboration with the community to gain context-specific knowledge.

1. How can we think about measuring DAO Community Health?

In the final section, we provide a high level description of what data we are collecting to measure a community’s vital signs.

---

Q1: What are the features of a community?

1.1 Defining a community

Community has been explored across multiple disciplines and perspectives, from biology to marketing, social psychology, network science and beyond.

Some examples of how communities have been described are as follows:

• A community has common values, norms, rules and regulations (Chung, Kim, and Shin, 2020).

• A group of people who are connected through relationships and have a shared identity, and set of values (Ospina, 2017)).

• Composed of: membership, influence, reinforcement and shared emotional connection. (Martiskainen, 2016).

• A self-organized group of people who make a commitment to be there for each other; they participate not only for their own needs but to serve the needs of others (Wheatley & Frieze, 2006).

• Community as Gemeinschaft (Ferdinand Tönner 1957 [1887]), where personal and informal ties are common and interaction is influenced by social values.

• Groups of people informally bound together by a shared expertise and passion for a shared enterprise (E. C. Wenger & Snyder, 2000).

• A tight cluster in a network of relationships (Hu et al 2008)).

Common among these definitions is that a community is composed of people, and that these people develop social relations with each other. A consequence of these two features is that people within one community develop a shared sense of identity strengthened through common values, and (deep) emotional investment for each other and the community.

Several sociologists (e.g., Willmot, 1985; Lee & Newby, 1983) agree with a three-fold classification of community as involving:

1. A place where the community exists and community members interact with each other. This is not limited to physical space.

2. A social structure unique to it based on the social relationships between community members (that both arises from the members interactions and shapes said interactions).

3. A meaning or an identity.

Crow and Allan (1995) argue that these three facets influence each other and should not be seen in isolation. In addition, the authors argue that a 4th facet, time, should be considered when analysing communities. By including time, it is possible to account for the evolution of communities. Communities contain forces which push some people out of the community and pull others in (Warwick & Littlejohn, 1992). The inclusion of time opens the door for including a shared history into the examination of communities. Communities are not only shaped by recent events, but also by interaction that happened many weeks, months or years ago. The dimension of time is also necessary to examine how a community reacts to external or internal perturbations, how it recovers and adapts.

1.2 The Telos of Community

Humans participate in social collectives to increase their chances of satisfying their needs through coordination. From this axiom, we assert that a community will continue to exist as long as the members deem their participation in the community to be beneficial, compared to other options to satisfy their needs.

Importantly, although a taxonomy of human needs is beyond the object of this paper, there's one need that's of special interest to the study of community: under its different names, the need for human connection, belonging, and relatedness. As Baumeister and Leary find, "existing evidence supports the hypothesis that the need to belong is a powerful, fundamental, and extremely pervasive motivation" (“The Need to Belong,” 2017). Community is not only a means to an end but also an end onto itself.

Finally, humans’ sense of self (and hence the definition of one's needs) is a complicated affair: we simultaneously desire to be unique, while we also want to belong to a collective (Brewer & Gardner, 1996).

We identify with several collectives, just by the nature of who we are. However, how strong we identify with a social collective determines its impact on our lives. For example, in religious and ethnic conflicts, a strong identification with a social collective can lead an individual to perceive the needs of the collective as their own, to the point where they might sacrifice themselves (or others) to serve said collective.

There is a feedback effect occurring between individuals and the community: whatever individuals believe or do bubbles up to create an emergent community. From there, this emergent community shapes and directs the actions and thoughts of individual members.

“They weave a web of reciprocity, of giving and taking. [...] Through unity, survival. All flourishing is mutual.” - Kimmerer, Robin Wall. Braiding Sweetgrass

1.3 Systemic view on Community

Just like a body is composed of various body parts which in themselves are composed of cells composed out of biomolecules, so is a community. And just like a body is part of a larger society and ecosystem, so is a community. According to the Nested System Model (Holling 2004), communities exist within a hierarchy of nested systems: individuals within relationships, within subgroups, within the community, within society, within the biosphere, etc.

All system levels are interconnected, and interaction between the scales is such that change (or evolution or transformation or reorganization) at one scale affects change at other scales. Aspects about the way that systems are connected and interconnected with systems on other levels are often consistent among the different levels. On the level of a community, when considering what the optimal ways of connection are and what types of connection lead to “disease”, we can take inspiration from systems on other levels that have been observed and studied extensively like ecosystems, cells or the brain. Taking inspiration from the networks of interaction in biological systems when designing or improving human systems is applying “Network Biomimicry” (van der Molen, 2022). Throughout the text, our perspective on communities and their health will be supported by network biomimicry examples taken from various natural systems.

Coexistence with other communities and individuals

In addition to the "vertical" interactions between nested systems (interactions with the level above and below), there are "lateral" interactions with other entities.

"From an ecological perspective, ecosystems consist of various species that interact with each other. These interactions can be of different forms.

• Antagonistic interactions: one species is harmed while the other benefits. For example, the predator-prey interaction between a lion and a gazelle. The lion hunts on the gazelle in order to be able to proliferate.

• Mutualistic interactions: both species benefit from each other. Like the interaction between a flower and a pollinator. The flowers provide food for the pollinators which in turn help in the reproductive process of the flower.

• Competitive interactions: both species negatively affect each other. For instance, two different sea mammals that compete for the same limited food sources." van der Molen (2022)

In our case, there can be antagonistic, mutualistic or competitive interactions at every level. For example:

• Members trying to allocate limited resources face choices of antagonism (e.g. scamming other members), competition (e.g. fighting to complete a bounty sooner), and collaboration (e.g. creating a proposal together).

• Sub-groups can prey (e.g. taking over the work and rewards of another group), compete (e.g. trying to gain a bigger budget for their guild), and collaborate (e.g. develop a new capability together).

• And similarly, communities can prey (e.g. attacks on another community to destabilise it), compete (e.g. arms race offering better rewards for participation), or collaborate (e.g co-organised events and other mutually beneficial initiatives).

Joiners and leavers

Finally, we assert that communities are open systems where members, ideas and resources can enter or leave the community. A community exists in the balance of these flows, and especially that of (different types of) members who can then enable or constrain other flows. The biological parallel is how proteins have a higher turnover than cells, cells have a higher turnover than organisms and organisms have a higher turnover than ecosystems (but individual species can also survive ecosystem collapse).

1.4 The Dark Patterns of Community

It's important to note that communities are not necessarily always positive. At the extreme end also lie gangs, cults, and violent sectarianism. Communities can harvest multiple dark patterns, and have detrimental consequences on the ecosystem, including:

Negative views of outsiders

Members who strongly identify with a community can develop a negative view of outsiders, those people who are not part of the same community. A good example of this is European football fans during Derbys. Strong identification with a community thus leads to intergroup stereotyping and trash talking (Hickman & Ward, 2007). Strong identification with a group can lead to exaggerated positive attitudes toward insiders and unrealistic negative attitudes toward outsiders. According to Molenberghs (2013, pg 1), from a neurological perspective, "we perceive the actions and faces of in- and out-group members differently, and we emphasize more with in-group members.". And from a cell biology perspective, a common symptom of cancer is that the tumor cells distance themselves from healthy cells by removing the gap junctions between them (Aasen et al., 2016).

Compromising members' wellbeing

The desire to preserve the group can suppress dissenting voices. For example, Barclay et al. (2004, pg 1) found that "rural communities have informal social norms for tolerating certain types of crime and for prescribing the reporting of such crimes. Many victims of crime suffered in silence". And Fox-Rogers finds that destructive patterns, such as clientelism and corruption, can be exacerbated by uneven distributions of power in communities (“The Dark Side of Community,” 2019). Within healthy biological systems, a strongly uneven distribution of power is hardly ever observed. Instead, these systems self-organise in a decentralised manner (Camazine et al., 2020).

Ossification

Through identification with a community, we pass our instinct for self-preservation to cultural symbols, rituals and ideas. These markers might have arisen as functional responses to an issue, later on becoming ingrained in the identity fabric of the community. Even when the issue is no longer relevant, and hence the markers have lost their functional value, we tend to resist change as a way of preserving the identity. As Fisher & Sonn (2002, pg 1) note, "the challenge of managing change is how to build forward, maintaining those markers of real social value, and incorporating the new ones that are brought by newcomers, and those that are developed together". When looking at natural ecosystems, the species that are able to adapt to the constantly changing environment will prevail. Meanwhile, species that are unable to adapt and continue their old behavioural patterns are most likely to go extinct (Darwin, 2003).

Groupthink

As we depend on our communities to satisfy needs and desires and experience a sense of reciprocity, we can come to value harmony as a (superficial) indicator of community health and hence of our own safety. Additionally, a desire for consensus can disincentivize disagreeing with the group (or at least expressing a difference of opinion). In consequence, communities can experience groupthink, a phenomenon where alternatives to the group’s (perceived) opinion are systematically dismissed (Solomon, 2006). Group members who share unique information are often evaluated more poorly (Thomas-Hunt, 2003) because they voice something unfamiliar to others.  Stress and isolation (the group being poorly linked with outsiders) significantly increase groupthink (Breitsohl et al., 2015). And unfortunately, “different states of anonymity do not have a strong effect on the likelihood of conforming to group opinion” (Tsikerdekis, 2013).

Unstable growth rate

A well-functioning community that serves the needs of its members can attract an ever-increasing number of new (potential) members. This is especially the case for communities not bound to a specific location. Virtual communities can grow with limited constraints thanks to the scalability of virtual platforms. However, fast growth can undermine the intimacy, cohesion, and generally sense of community of the early days (Slemp et al., 2012). As attention is a finite resource, the exponential growth of a community can drain and extinguish other communities, which in turn can destabilise the ecosystem that sustains the growing community. The risk of unstable growth is also reflected in cancer. The rapid growth of tumor cells brings the whole body out of balance, eventually leading to the death of all the cells in the body, including the tumor cells (Houten & Reilley, 1980).

So far, we have defined what a community is and asserted that it will continue to exist as long as the members deem their participation in the community to be beneficial, compared to other options to satisfy their needs. We have introduced the Nested System Model which describes communities as existing within a hierarchy of nested systems. We also introduced Network Biomimicry which is taking inspiration from biological networks when constructing or improving human networks. We described antagonistic, mutualistic and competitive interactions as the different ways in which components of a system can interact with each other and we presented unhealthy aspects of communities and how they relate to unhealthy systems in biology. In the next sections, we will give a definition of DAOs and describe DAO communities in depth, we will define what DAO community health is and we will give an overview of how DAO community health can be studied and improved.

---

Q2: What characterizes DAO communities?

2.1 What is a DAO

Before characterizes DAO communities, we need to describe briefly what a DAO is.

In What is a DAO? Conceptual Foundations, (Ospina & Bohle Carbonell, 2022) put forward the following definition of a DAO:

Collectives that exhibit organisationality, expressed and evolved through communication events and processes, and shaped by an Ethos that highlights:

• Decentralised power: no single source of authority.

• Autonomous: self-sovereign, not bound to an external coercive force.

• A common goal, vision or set of values that are (being) worked towards.

• A shared treasury controlled by a decentralised voting mechanism.

From this definition, we can then look for typologies of communities and compare them to DAOs and other collectives that exhibit organisationality to identify similarities and differences between these concepts and thus identify conceptual foundations to meaningfully discuss DAO Communities and avoid reinventing the wheel.

2.2 Understanding DAO Communities

Communities exist in many different shapes and forms. As we’ll see, we can differentiate DAO communities from other categorisations of communities (e.g., kinship communities), and we can also differ DAO communities from each other (e.g., the structure of social DAO vs Protocol DAO). To structure our discussion on the properties of DAO communities we use the following five questions:

1. Who influences and controls the community?

2. Who participates in the community?

3. Where is the community gathering and interacting?

4. What is the structure and how bonded and intentional is the community?

5. What defines the Community's identity?

Who influences and controls the community?

Communities can begin with a group of people settling in an area or meeting regularly to engage in their favourite activities. Meanwhile, other communities are formally planned and sponsored by a specific actor who desires the community to exist. Lauden & Traver (2003, as quoted by Porter, 2004 page n.d.) distinguished between member-initiated and organisation-sponsored communities: While member-initiated communities are created and managed by community members, organisation-sponsored communities start with a sponsor (commercial or non-commercial organisation such as a government, non-profit, educational institution). Organisation-sponsored communities have key stakeholders and/or beneficiaries (e.g., customers, employees, students) who are important to the sponsoring organisation’s mission and goals and as such will shape the community.

We can not conclude that DAOs start necessarily as either member-led or organisation-sponsored, having examples of both cases (i.e. a community progressively organising into a DAO and a centralised organisation building community and progressively decentralising into a DAO). That being said, the motives for creating the community can be different if it’s formed bottom-up or top-down. However, given the real-world examples, we can not claim that DAO communities are always member-led or organisation-sponsored.

However, it's clear from our definition of a DAO that there's an aspect of Organisationality as well as a "A common goal, vision or set of values that are (being) worked towards" and "shared treasury controlled by a decentralised voting mechanism". As such, we can leverage a comparison to communities that exist in symbiosis and overlap with an organisation, and note the ambition for the community to (eventually) become self-governed.

Who participates in the community?

Compared to traditional organisations, DAOs value decentralisation and a shared treasury controlled by a decentralised governance mechanism (Ospina & Bohle Carbonell, 2022). As a consequence, aspirationally, DAOs are governed by their community and their communities are made to be their governors. This, in turn and as we shall see, encourages specific patterns of participation and membership in DAO communities.

To date, the mechanisms for community-led governance rely most frequently on the use of digital tokens, and although the design of the token system varies, there are a few common patterns: users often receive tokens (via airdrops and other mechanisms), labour providers are also often compensated (partially or wholly) in the DAO's token, and investors receive tokens instead of share certificates. The tokens may be designated as “utility” (required to access the community or use the product, services, or platform) or as “governance” (enabling token holders to govern the DAO). As they can be transferred, one may treat them as assets or utilize them as investment vehicles.

The distribution of tokens among those using the platform, those providing capital, and those providing labour enables each to use tokens toward the platform services, participate in its appreciation of value, and govern the DAO, which means that DAOs tend to merge stakeholder classes. In consequence, compared to traditional communities, the types of stakeholders of a DAO are more enmeshed and hopefully aligned, increasing the chances of them merging into a single identity and community: DAO members.
As such, traditional lenses to study communities - such as communities of practice, brand communities or even workers communities - risk being too reductive in isolation to understand the variety of motivations of DAO community participants and the resulting variety of interactions (e.g. knowledge sharing, peer support, co-creating, co-shaping, etc).

We can thus conclude that all stakeholders, regardless if they are investors, workers or users, may often choose to participate in the community.

Where is the community gathering and interaction?

The early literature on Community emphasised the importance of shared geography (i.e. co-located communities). A community was tied to a specific neighbourhood with its socio-economic characteristics, and often kinship ties between members of the community.  But with the advent of the internet and social media, virtual communities have multiplied and so has the research on them.

A key differentiating factor between co-located and virtual communities is the manner in which community members interact with each other. Offline communities benefit from the richness of offline interaction (e.g., non-verbal communication, touch, smell), visual membership signals (e.g., clothes, accessories, murals), and physical proximity. To add richness and texture to interaction, virtual communities have to develop new ways to substitute the dry text-based communication mediums. Currently, this is achieved through emojis, memes, gifs, videos, and increasingly 3D avatars). This distinction between virtual and co-located communities is important, as interaction mechanisms are crucial for community members to build and reinforce relationships between each other.

Most DAO communities to date operate primarily online. As such, we'll refer primarily to virtual communities literature in the construction of our framework.

How bonded and intentional is the community?

Henri & Pudelko (2003) classified Virtual Communities according to two different dimensions that vary on a continuum:

• The strength of the social bonds: the strength of social cohesion between community members i.e. communities can be a group of loosely coupled people or a tight-knit one. The authors argue that depending on the intentionality of the community’s goals and objectives, members have loose and weak connections with each other or form a tight-knit cohesive social group.

• The degree of shared intentionality: refers to the existence of common objectives and interdependence among the participants (Bock, Ahuja, Suh, & Yap, 2015; Gangi & Wasko, 2009; Meirinhos & Oso ́rio, 2009). The more intentional people are with why they are forming and participating in a community, the stronger the community will be.

DAOs vary across both dimensions with, for example, Social DAOs being primarily about member-to-member relationships and Protocol DAOs being primarily about a shared goal of developing or maintaining the protocol.

DAO communities can blend stakeholder types and may have rather fluid boundaries, however, they also tend towards autonomy (see Ethos of DAOs, (Ospina & Bohle Carbonell, 2022) with the DAO maintaining its own platform and controlling the organisation's membership and access. In consequence, DAOs often include a highly aligned and bonded core group (who maintain and develop the platform and shape the organisational aspects), followed by ad-hoc contributors and a periphery that resembles more a gamer or learning community (where members have low social bonding and low shared intentionality, interacting primarily for personal entertainment or learning).

We conclude that DAO Communities vary in shared intentionality and strength of social bonds both across DAOs and between subpopulations within each DAO.

What defines the Community's identity?

Although each DAO and each DAO community will have unique elements to its identity, generally speaking, we can reference the four qualities highlighted by the Ethos of DAOs (section 2.1) as common factors that characterize DAO communities and might differentiate them from other communities (Ospina & Bohle Carbonell, 2022).

From the theoretical review above, we can arrive at five characteristics of DAO Communities:

• Exist in symbiosis with an organisational aspect (the DAO).

• Include multiple stakeholder classes and tend to blend them into one.

• Gather primarily online (virtual communities).

• Vary in shared intentionality and strength of social bonds both across DAOs and within each DAO's Community

• Subscribe to the Ethos of DAOs (decentralisation of power; autonomy; shared goals, vision or values; and a shared treasury managed by the community).

---

Q3: What is DAO Community Health?

Borrowing an analogy from the medical domain, we view a community to be, in a way, like a living organism. Although traditional definitions of health were based on the absence of illness, more recent definitions go beyond and assert health as an emergent property and multi-dimensional construct:

“the complex systems on which our lives depend — ecological systems, communities, economic systems, our bodies — all have emergent properties, a primary one being health and well-being” (Goodwin et al., 2001, p.27)

3.1 The Regenerative view on Community Health

20th century views on health tended to consider the body as a single organism. However, this organism consists of an interaction of organs which in turn consists of interacting cells. And more recent research (e.g. Pflughoeft & Versalovic, 2012) has shown that beyond these cells,  humans consist of microorganisms that coexist alongside human cells to enable human life, not to be considered in isolation. Equally, some traditional views of community health depict a community as a single unit or simply a collection of individuals. However, as we have argued, a community is composed of multiple nested systems that interact and depend on each other, namely:

1. Individual members

2. The relationships between individuals (reciprocal or not) forming community specific participation structure

3. The sub-groups (formally created) and cliques (informally arising)

4. The overall community

5. The overall ecosystem in which the community is embedded

As these nested systems are interdependent, we assert that a community is only healthy insofar as its subsystems and supra systems are also healthy. Consequently, a community that can support and contribute to the health of these nested systems is deemed (regeneratively) healthier.

3.2 Health as Resilience, Adaptability, and Transformability

Importantly, health goes beyond the absence of illness. Communities exist within an ecosystem. This ecosystem is constantly in motion: members are interacting within a community and across communities, people are joining or leaving, and environmental changes in the ecosystem are impacting members and communities. To be healthy, beyond an ephemeral state, a community needs to deal with these internal and external forces shaping it. This is in line with Carillo’s (2017) suggestion that a community is only healthy if it is able “to function effectively, to cope adequately, and to change appropriately in response to internal and external stimuli".

Although naming conventions have fluctuated across fields, we can highlight a few necessary properties for communities to remain healthy. Communities that possess these properties remain healthy or regain their health status after a period of turbulence. Borrowing from Walker et al. (2004) discussion on sustainable development and change of social-ecological systems, three attributes describe the developmental trajectory a community can experience. These attributes describe an increasing level of agency in a system.

• Resilience: The capacity of a system to absorb disturbances and reorganise while undergoing change so as to still retain the same function, structure, identity, and feedbacks*.

• Adaptability: The capacity of actors in the system to influence resilience.

• Transformability: The capacity to create a fundamentally new system when ecological, economic, or social structures make the existing system untenable.

*Note: some definitions of resilience also include Adaptability and Transformability. We chose to retain this distinction for the practicality of having a detailed language to discuss Community Health evaluation in the following section.

Community Resilience

Community resilience is the idea that, after an internal or external event shocks the community, the community can regain its former status quo (Matarrita-Cascante et al., 2017). The community is thus responding to events, processing the created tension while seeking to regain its former stable state. Here, the developmental trajectory is purely reactive: The system is dealing with the events so that it can regain its current status.

In network science, resilience is described as the time it takes for a system to return to an equilibrium after a perturbation (Okuyama & Holland, 2008; Thébault & Fontaine, 2010), a change occurring in the system. Two forms of resilience can be specified:

• Resilience to removal (species extinction): From a DAO perspective, a perturbation could be a contributor involuntarily or voluntarily leaving (e.g. stopping to participate in the community). Other DAO members might have relied heavily on the presence of this person and, missing them, the other members might leave in turn. The persistence of a community reflects the number of community members that remain after a new equilibrium is reached in response to a community member that leaves. The parallel in ecology is a species becoming extinct, leading to an extinction cascade.

• Resilience to addition (species invasion): From a DAO perspective, a perturbation could be contributors joining a community. While most communities will be able to absorb the perturbation from a single contributor joining, the impact of a large number of contributors joining could send negative reverberation through the community. Imagine if a large number of inexperienced or value-unaligned members join a DAO, diluting the community's identity. Similarly, a community's health will suffer if malicious actors enter trying to capture its governance.

The network perspective on resilience can be augmented by adding insights from other research fields. For example, the psychological perspective focuses on the resilience of individuals and defines resilience as the ‘‘existence, development and engagement of community resources by community members to thrive in an environment characterised by change, uncertainty, unpredictability and surprise” (Berkes & Ross, 2013). We consider individuals as the lowest level of our system view of community health. Naturally, their resilience impacts the resilience of other systems, just like the resilience of cells impact the resilience of an organism, and how the resilience of organisms impacts the resilience of an ecosystem. Another view is the socio-ecological perspective on resilience. This view considers resilience as intentional action to enhance the personal and collective capacity of community members and institutions to respond to and influence the course of social and economic change (Berkes & Ross, 2013).

Adaptability: Learning and Antifragility

Beyond the ability of a system to resist or recover from shocks (i.e. resilience) lies the ability of a system to adapt such that it can avoid future shocks and increase its resilience. In this developmental trajectory, the community is adopting a more proactive stand and is aiming to “do something” with the external shocks, and not just to go back to the original state.

To achieve adaptability, DAO communities need to be learning; in other words, they need to adopt elements of a Learning Organisation (Senge, 2006). Organisational learning is "the process by which the organisation constantly questions existing product, process and system, identifies strategic position, and applies various modes of learning" (Wang & Ahmed, 2003, pg. 14).

As per Wang & Ahmed's definition, Organisational Learning gives us a (somewhat) holistic perspective on Adaptability. A community can only adapt if its members are (continuously) learning and its culture is evolving. In addition, it needs to have processes in place to change its processes, capture and distribute its knowledge, and improve continuously as a collective.

More recently, as the concept of organisational learning has increasingly expanded to include aspects of creativity and (disruptive) innovation, it ties with another concept with applicability to communities: antifragility - the ability of a system to produce a response to stressors such that it leads to more benefit than harm (Taleb 2013). Said otherwise, systems which are antifragile have learned (adapted) in such a way that they benefit more from negative shocks and perturbations than from positive events. A negative shock is perceived by Taleb (2003) as a random event, potentially an error in the system or one of its sub-systems. From our nested system perspective, we would like to include shocks from supra-systems (the environment) as potential sources of negative shocks. A community that is not antifragile will deal with the resulting tension from the shock and return to its erroneous default state. Therefore it has not learned from the perturbation and will suffer future setbacks due to subsequent negative events.

Enabling Adaptability (and its advanced state Antifragility) requires enablers at the individual and collective level, and between them:

• At the individual level this relates to individual capabilities such as cognitive flexibility (Dane, 2010), as well as resources such as social capital (Nahapiet & Ghoshal, 1998) have been shown to enable an individual to learn and adapt to new work and career circumstances (Armanda Hamtiaux & Claude Houssemand, 2012; Oh et al., 2022).

• At the intersection of systems, Adaptability is enabled by the ability of individuals (or sub-systems) to make and execute choices that lead the collective to adapt i.e. bottom-up governance or grass-roots movements.

• At the collective level, Adaptability requires

  • Collective Intelligence (i.e. collective sense-making, meaning-making, and choice-making) and Collective Leadership (Por, 2008),

  • Collective Capacity i.e. Community Capacity as the ability to get things done to implement the changes (Laverack, 2005).

Transformability

Finally, beyond the ability of a system to adapt, systems need to transform. This occurs when a system reaches an evolutionary dead alley - a point from which it can no longer adapt to continue to serve its purpose. A community which is on the developmental trajectory of transformation is fundamentally changing itself to become a fully new system.

In biology, an evolutionary dead alley leads to death and decomposition. However, the composing elements are the basis for generating new life. In science, we can draw parallels with changes in scientific paradigms where the old worldview accumulates exceptions and inconsistencies until it falls apart and is replaced by a new one. This new one might retain elements of the old one.

Traditionally, we have been more concerned with Resilience or Adaptability than with Transformability. A Google Scholar search on  “community resilience” yields 3,000,000 results, “organisational resilience” 188,000, “community adaptability” 624,000, “organisational adaptability” 112,000, while “community transformability” yields only 18,500 and “organisation transformability” 17,200. And importantly, the word “transformability” is often used in the found papers with a definition that more closely resembles Adaptability than actual Transformability. However, in DAOs some primitives have been identified that directly enable Transformability:

• At the individual level, community members can Rage Quit, leave the community and take their share of the collective assets with them.

• At the subgroup or clique level, a subset of community members can fork, copy a DAO's smart contracts into a new organisation.

And some traditional barriers to Transformability (e.g. non-compete agreements and non-disclosure agreements) are largely absent in DAOs.

3.3 DAO Community Health

Having covered the regenerative view on community (nested, interdependent system) as well as resilience, adaptability, and transformability, we'll now aim to bring these concepts together to form a unified definition of DAO Community Health and qualify health (multidimensional spectrum of health). As such, we define DAO Community Health as:

Healthy or not:

As per the above definition, a DAO community is considered healthy (positive health) if the state of the community is such that it:

1. Contributes (or at least does not destabilise) its nested systems and itself:

   1. Satisfies the needs and aspirations of the members (including alignment with their values e.g. the Ethos of DAOs);

   2. Promotes healthy relationships, and functioning subgroups and cliques;

   3. Advances its collective goals (community capacity to generate value for the DAO as a whole);

   4. Contributes towards a healthier ecosystem.

2. Is resilient, adaptable, and has the capacity to transform to better fulfil the above.

Conversely, an unhealthy community is one that fails to fulfil either #1 or #2 above. And as such, would likely exhibit some of the patterns, we covered in the section on The Dark Patterns of Community: negative views on outsiders (hyper-competition), ossification (lack of adaptability), groupthink (collective stupidity), compromising members' wellbeing and unstable growth rate (non-regenerative).

Finally, we can define a range of stages in between (between theoretical extremes of perfect positive health and perfect negative health) based on the ability of a community to fulfil the different components.

---

Q4: How can we evaluate a DAO Community’s Health?

4.1 Evaluating a DAO community’s health

As per our definition above, we assert that Community Health is a function of members' actions and the interaction between the community's nested systems, and we can evaluate it with a multilayered model of Community Health measurement.

By the way, we've built a tool to make it super easy for you to assess the health of your community: TogetherCrew.com

4.2 The Vital Signs: Engagement, Structure, and Sense of Community

We assert that the health of a community can be evaluated in a snapshot (at a specific point in time) through three lenses:

1. Members' level of engagement.

2. The structure of members' interactions (the structure of the sub-systems).

3. Members' sense of community.

These snapshots can be taken for different time periods (e.g., daily/ weekly/ monthly; pre-season, season, post-season). By combining frequent passive data collection (e.g. daily collection of messages on chat platforms) with less frequent active data collection (e.g. quarterly surveys and weekly pulse surveys), we can balance depth of insights with member disturbance. These snapshots add to a dynamic view of Community Health, providing the opportunity to connect specific actions and events to changes in community health. And as such, enable community managers to evaluate their impact and communicate it to other stakeholders.

Our definition of Community Health includes five aspects:

1. Community members

2. Relationships between two members

3. Cliques (informal groups) and subgroups (formal groups)

4. The community

5. The larger ecosystems of communities

As we adopt a  regenerative system view on community health, the health state in one system (positive or negative) will impact the adjacent systems.

For our selected vital signs, engagement measures the health of community members in terms of their activity and not their holistic health (physical, mental etc., and includes not only the treatment of illness but also its social origins (McKee, 1988; Saylor, 2004). We assume that a person’s holistic health will impact their activity. As such, directly measuring a person’s holistic health will be undertaken in an exploration of the Enabling Factors for long-term Community Health. Similarly, measuring the health of the larger ecosystem is a vital but herculean task, as we would need to measure participation structures across DAO communities. However, to keep the ecosystem in mind, we will track general events that have an impact on DAO communities (e.g., rise/fall of cryptocurrency, sentiment analysis of news impacting web3).

Finally, we will be measuring the health of relationships between community members, cliques and subgroups, and the complete community. Our guiding assumptions and selected metrics are described below.

4.2.1 Engagement

Engagement is directly tied to an observable behaviour of community members: Their level of interaction with other community members and content created by the community. A lack of engagement is visible through the absence of voices in the conversation, content that is not shared, and events that are not attended. Engagement is thus synonymous with community members' involvement in a community.

According to Etienne Wenger (1999) members have different trajectories of participation in a community:

1. Peripheral trajectories are those that never lead to full participation

2. An inbound trajectory describes the path of newcomers who are committed to becoming full participants within the community and whose identities are invested in their future participation.

3. Insider trajectories are held by those members that have reached full participation. The trajectory of their participation is focused on the evolution of the community and the renegotiation of their place within it.

4. Boundary trajectories span different communities as participants work to sustain an identity across community boundaries.

5. Outbound trajectories entail movement out of the community.

Another distinction in participation can be made between public and private interaction. Public interaction is visible to all, whereas private interaction happens behind closed doors and is only visible to the selected few. Interactions that happen in private messages or small gated communities give community members a chance to build intimate and deep relationships (bonding social capital) with others (Donath, 2007), while interactions in open spaces enable community members to interact with many others and build a wide net of brief acquaintances (bridging social capital) (Lee et al., 2014). Further research has shown that communicating with others in an open space stimulates people to create deep and shallow relationships, whereas private communication only helps to strengthen relationships (Li & Chen, 2022). Both forms of social capital, bonding (deep relationships) and bridging (acquaintances) help community members develop a sense of community (Li & Chen, 2022).

Importantly, engagement should not be limited to the interactions described above. As we have seen, DAO Communities thrive through an ecology of interactions between the diversity of stakeholders that are merged together, and this generates health as an emergent property. In consequence, we'll strive to expand over time the range and types of interactions that can be collected in the data to enrich our measurements of DAO Community Health.

4.2.1.1 Lurking

Missing from the list above is lurking as a form of passive participation (Nonnecke & Preece, 2001). This is a form of participation originating in virtual communities. Community members who lurk are visiting the community’s virtual space (e.g., forum, Discord, Reddit, Twitter or other social media channels), but do not visibly interact with any content in any form: They do not reply using emojis or text and do not attend events.

It is important to consider lurkers, because as much as 90 per cent of the individuals joining a community never become active in the first place (Nonnecke and Preece 2000; Schneider et al. 2013, as quoted by Trier 2014). It's presumed that lurkers read posts and thus consume the community's content, potentially sharing it in other communities and integrating it into their daily practice (Takahashi et al., 2003).

In the case of DAOs, lurkers might contribute significantly by serving as ambassadors and evangelists across communities, thus helping to direct attention and resources. However, it is also possible that numerous lurkers consume information and insights and port them to other (potentially competing) communities without giving back. Such a form of cross-DAO information sharing can be beneficial for the complete DAO ecosystem and, assuming a balanced inflow and outflow of information, also for individual DAOs (Thébault & Fontaine, 2010; C.-C. Wang et al., 2017).

In aggregate, the free flow of information is likely to facilitate the development of the overall ecosystem but poses questions around community-level sustainability that each community needs to contemplate according to their unique situation.Additionally, as lurkers remain anonymous, their presence can impact trust and a Sense of Community:.

Lurkers do not post, and therefore do not create an identity in the online community. Consequently, they remain anonymous, leading other community members to wonder with whom they are sharing this online space. Finally, lurkers themselves suffer from remaining silent. A lack of participation in public conversations reduces their Sense of Community by not building up any bonding social capital and deep relationships with others (Li & Chen, 2022).

4.2.2 Participation Structure

The engagement levels of individual community members create a specific social fabric that is unique to the DAO community. This social fabric weaves the different strands of interactions together to create a community's participation tapestry. In the following text, we will be calling this participation tapestry a social network or just a network.

Social networks, interaction patterns between humans, have been studied for many years. Researchers have investigated the impact of network structure on people’s choices (e.g., adopting an innovation, eating a BigMac) and outcome (success vs failures) (Borgatti & Halgin, 2011). The general assumption among social network scientists is that the connections between people help them exchange things such as information, energy, and goods (network flow model), or help them coordinate action and beliefs such as peer pressure and protests (network bond model).

As we view DAO Community Health from a regenerative lens (DAO Communities as interdependent nested systems), it's worth considering network biomimicry. Various biological model systems can be applied, but we would like to especially highlight network research from ecology due to the extensive research in this field, spanning decades. As we've mentioned, a DAO community has several participating stakeholders (investors, workers, users, etc), which ecologists would call ‘species’, and we can also conceptualise different functional skill groups (developers, marketers, designers, etc.) as species. Ecology, compared to social science network research, focuses on factors that lead to the growth or decay of populations within an ecosystem, and as such can give us a lens to understand dynamics within the community that can lead to its success or failure. Similarly, types of stakeholders could be seen as cell types or brain regions and so examples related to these types of systems will be briefly mentioned too.

Early network studies using computer models have shown that networks tend to become less resilient when they grow in size and complexity (May, 1972). Whereas observational studies within ecology show an increase in the resilience of an ecosystem with increasing complexity (Hedgpeth, 1954; MacArthur, 1955). In the decades that followed, extensive research elucidated that this increased resilience was a result of the specific network structure that is present in ecosystem networks (Landi et al., 2018; van der Molen, 2022). Thus, ecosystems that grow in size and complexity can become more resilient if the interaction patterns between members of the ecosystem are supportive. For example:

• Moderate amounts of clustering: In networks, a cluster is a group of community members who interact predominantly with most or all members in that group. They form a sub-system which arises bottom-up without formal incentives. When this clustering is present in moderation, members from different clusters also have limited interactions with members from other clusters. This is referred to as the ‘small world effect’ (Watts & Strogatz, 1998). A community is described as a small world if two conditions are met: it has cliques giving members space to discuss niche topics which facilitates local information processing, and boundary spanners that connect these niches to ensure global information integration. Interestingly, protein interactions networks in cells and brain region interactions are also small world like (Bassett et al., 2006; Goldberg & Roth, 2003).

• Variation in the number of relationships that members have with other members: Some members of the ecosystem are connected to many other members, while others interact only with a few. In ecology, the members who are connected to many other members are described as generalists. In social science, they are called popular members or work-horses. Online communities often show this uneven long tail distribution and this is observed in many other biological systems too, including cells and the brain(Almaas & Barabási, 2006; Tomasi et al., 2017).

• A skeleton of strong relationships but a majority of weaker relationships: Certain members have a strong influence on each other, creating a stable framework for the community. However, the majority of the members only infrequently interact with each other allowing the community to be flexible to changes. This structure reflects a balance between bonding and bridging social capital (Li & Chen, 2022). Similar skeletons of strong interactions supporting a larger number of weaker interactions are also observed in brains and self-organised brain organoids (Sharf et al., 2022; Song et al., 2005).

• Connectivity: When people have more relationships with others in the community, it leads to a more stable community. Information can travel more easily through the community due to the redundancy in the ways in which a message can pass from one person to someone else (Thébault & Fontaine, 2010). A higher number of independent paths that connect every pair of members in a community leads to a higher social cohesion within this community (White & Harary, 2001), and is a sign of long-term community stability (Quintane et al., 2013). Similarly, redundancy in brain networks also support cell and brain function (Bettinardi et al., 2017; Cutler & McCourt, 2005).

• Both mutualistic and competitive interactions: Some interactions are beneficial for both individuals involved whereas other interactions are competitive. Both of these types of interactions contribute to the stability of a community (Mougi & Kondoh, 2012). For instance in DAOs where members contribute their unique skills in a collaborative effort, working towards the overarching goals of the DAO (mutualistic). At the same time, DAO members can compete with each other in coming up with the best solutions and implementations (competitive). For example, by filing competing proposals in attempts to obtain a bounty, promoting high-quality proposals. Similarly, in biological networks many proteins have as sole or main function to enhance or silence other proteins (Duan & Walther, 2015).

4.2.3 Sense of Community

In contrast to the observable health measure of engagement, a Sense of Community (SoC), can not be observed, but only inferred from other data points. SoC is an emergent state describing the emotional interconnectedness between community members

Researchers consider that a Sense of Community is one of the main elements that stimulate participation in an online community (Blanchard, 2008; Luo et al., 2017; Talò et al., 2014). The Sense of Community creates the motivating force for people to log into the online community. In addition, it creates an emotional bond with other community members, a similar type of connection people feel towards friends they meet in real life (Abfalter, Zaglia, & Mueller, 2012; Luo et al., 2017).

According to Blanchard's (2008) research, important predictors of a sense of virtual community are the existence of shared norms, observing people supporting each other, and creating an identity in the online community. Drawing a parallel to this definition, we conceptualise the 'Sense of a Community' as composed of:

• Membership: a sense of belonging and identifying with the community, a common symbol system.

• Shared norms: Members have (some) shared understanding of what behaviour is acceptable. Behaviour here refers to communication style in text-based and video-mediated communication. Text-based communication includes words, pictures (e.g., gifs) and emojis.

• Community-specific identity: Members develop an identity specific to the community and through this, perceive to be accountable to the community. This community-specific identity can be different to the identity a member develops in a different community. To not break down their association with the community, members are cognizant of not developing a too unique identity and of resembling, to some degree, other community members.

4.3 Enabling Factors and additional lenses on the nested systems

As we have seen, the Vital Signs already give us a surprising wealth of information across the different nested systems and even have some predictive capacity. However, it's always possible to go deeper. A non-exhaustive list of constructs that can be used to provide some additional insights are:

• Agency, Autonomy, and Sovereignty.

• Trust and Psychological Safety.

• Incentivisation and Motivation perspectives.

• Technology Acceptance Model.

• Psychometric profiling

• Holistic Health evaluation

• Cross community measurement to derive ecosystem health.

---

Q5: How can we think about measuring the Vital Signs?

Part of the vital signs for Community Health are individuals' perceptions about the community, while others are individuals' actions in the community. For this reason, we are combining different data sources to measure the health of a community.

Engagement of individual community members is measured through their participation in the community. This refers to their posting behaviour and event attendance. Specifically, we are measuring their position in the community and how it evolved over time to calculate the proportion of community members who are on inbound (new joiners), insider, or outbound (potential leaver) trajectory. Finally, we will also measure the proportion of community members who span different subgroups and cliques in the community. These individuals hold key positions as they enable the exchange of information between smaller groups.

When measuring the Participation Structure, our goal is to identify the ability of the community to sustainably grow in size and complexity. As a data source we are using community members’ interaction data. Interaction data are the posts members make in public channels in the community’s online forum. By using this data source, we are reducing the burden of community members, as they do not have to fill out a survey and recall their participation patterns. In addition, this data collection strategy is less prone to recency and salience bias. The leading indicator of Participation Structure is the small world metric which reflects a balance between clustering in the network and the presence of boundary spanners that connect these clusters into a cohesive whole. Additionally, we will investigate the distribution of the number of interactions that different members have and the distribution of the strength of these interactions to assess community resilience.

To measure Sense of Community we are using a scientifically validated survey. By using a survey, we are able to gain insights into an individual's perceptions and feelings about the community. The Brief Sense of Community scale has been developed by David McMillian, one of the authors first describing the theoretical framework. From the original 8-item survey, in our pulse survey, we have decided to focus only on the two questions assessing individual members’ emotional connections. We decided this, as a pulse survey is normally very brief (2 to 5 questions). The dimension of emotional connection was selected as it has the strongest association with the overall construct of SoC.

Notes on Individual and Ecosystem Health and on Resilience, Adaptability and Transformability.

Although measuring these dimensions falls outside the scope of the Vital Signs (our starting point for Community Health), we can derive information on them thanks to the data we're already collecting.

For example, Sense of Community has been shown to be an indicator of community resilience and individual wellbeing.

Additionally, to measure the resilience of a community, we can calculate how the participation structure changes over time. A community that regains its previously (healthy) engagement levels, participation structure, and sense of community, is considered resilient. The accuracy of the resilient measure depends on the selected time-frame and internal and external events that occurred during the data collection period. It is important to first establish what is the baseline level of the community, and if this baseline level is considered to be healthy for the specific maturity stage of the community, before measuring the resilience of the community.

Finally, we have seen that structural elements that enable effective information flows are directly linked to a community's ability to adapt (Matarrita-Cascante et al., 2017), and disempowerment (linked to poor adaptability via its impact in community governance's effectiveness) can lead to dis-identification and hence reflected on the Sense of Community.

---

Avenues for further research

As we've seen, the topic of DAO Community Health is vast, and we're excited to bring this project to light and contribute to this stream of research.

For our next phase of research, we'll be collecting data from different communities to test and refine the model of Community Health. If you'd wish to participate with your community, please join the RnDAO Discord server or the telegram group for this initiative and let us know.

Additional research projects (by us or others), could expand upon the health assessments of the different nested systems with the additional constructs we have highlighted. Equally, further research could explore the enabling factors for long-term Community Health.

---

This research would serve little purpose if we couldn't translate it into action. So we made a tool for you. Let's build healthier community together!

Try TogetherCrew!

---

Acknowledgments

We thank the RnDAO community that has provided fertile ground for the inspiration and cradling of this project and we thank our sponsors, Aragon and Polygon for their generous support.

---

Bibliography

A Typology of Virtual Communities: A Multi‐Disciplinary Foundation for Future Research—Porter—2004—Journal of Computer-Mediated Communication—Wiley Online Library. (n.d.). Retrieved July 1, 2022, from https://onlinelibrary.wiley.com/doi/full/10.1111/j.1083-6101.2004.tb00228.x

Evaluating community capacity: Visual representation and interpretation | Community Development Journal | Oxford Academic. (n.d.). Retrieved July 15, 2022, from https://academic.oup.com/cdj/article-abstract/41/3/266/494888