tl;dr: This post is the fourth in a series surveying the experience of professional culture builders and what their common experiences tell us about building culture in a way that is both bottom-up and intentional. I present five dimensions of culture that practitioners seem to agree on. By breaking culture up into actionable, relatable dimensions, we make it less mysterious. I basically use my judgement (and share my raw data) to coarsely bin “output” dimensions of culture (this post) and “input” levers for changing culture (next post). The 5 dimensions I identify here: interpersonal bonding, shared purpose, shared behavior, shared values, and shared identity.
-
Post #1 laid out the question and the tensions motivating it: Can decentralized organizations do what decentralized organizations do: focus design effort and energy on mechanism, incentives, and the “tangibles” of organization design, or do they have to put energy into intangibles as well?
-
Post #2 shared the results from 5 pre-interviews with culture building professionals: practicing consultants who have worked in enough organizations to develop a general view of what culture is and how an organization can change it. Despite major differences in their backgrounds, they all do basically the same thing. This suggests that culture building, as a set of concrete shared practices, is not a mystic art, but something any community can do with enough energy and commitment.
-
Post #3 describes the method behind posts 4–6: what we did (read a bunch of books by more consultants) and some of the meta takeaways of the effort.
-
Post #4, this post, describes the five qualities or dimensions of culture that authors across our review had the most agreement on. These are concrete qualities things that our efforts at culture building can target with specific, concrete interventions.
-
Post #5 describes 15-20 levers of culture change, ranked by how often our pool of practitioners mentioned them. These would be the intervention that a community’s culture building projects would implement to target each of the dimensions of a good culture.
-
Post #6 argues for explicit management of culture change in decentralized communities. It asserts that culture change is a project that should be managed and that it is possible to do so in a way that respects the bottom-up essence of culture and its changes.
Introduction
After proxying and comparing the mental models of more than a dozen professional consultant teams (5 from interviews and 11 from their books) several themes pop up, despite the fact that these professionals were trained under all kinds of traditions and work for all kinds of companies. I wasn’t sure whether to expect wildly divergent or convergent systems. I’ve since been surprised at how very much they all do the same thing. For example, these four definitions of culture are very similar considering that they came from totally independent authors:
-
“The shared assumptions, values, behaviors, and artifacts that determine and reflect “how and why we do things around here”” — Heskett, Win from Within
-
“Our definition of culture, “assumptions, expectations, beliefs, social structures, or values guiding behavior,”” — Briody Trotter Meerwarth, Transforming Culture
-
“A set of beliefs and resultant observable behaviors that determines—more than any other factor—the performance of the group.” — Happy to work here, DeMarco
-
A culture is “a set of shared experiences, values, and goals that is unique to every… team observed” — Fitzpatrick and Collins-Sussman, Debugging Teams
Given this convergence on what these culture builders build, perhaps there are commonalities across their perceptions of culture’s constituents: the more micro, actionable dimensions that collectively define culture. In a community that has a “good” culture, what are the qualities or strategies that are present or absent?
Some of these qualities are explicit in the definitions, as in Briody et al.'s “assumptions, expectations, beliefs, social structures, or values guiding behavior” but to quibble, even this definition offers an ambiguous listing of the “parts” of culture. Do I parse the sentence to be naming “behavior” or the “values guiding behavior”? And what about a quality like shared identity (also called group identity, collective identity, or sense of belonging by different authors)? It appears in none of these definitions, but does appear in more than half of the books as one of the things that a culture builder tries to maximize.
This was the essence of the second half of the SCRF-supported culture building research: beyond stated definitions, what in practice and experience do culture building practitioners target and how? The work of the project has been identifying similarities and reconciling differences to build a holistic sense of the kinds of things that communities can learn from the professionals when it comes to articulating what they mean by “culture” and what it means to build it.
“Creating a culture takes action because it is not something you are, it is something you do” — paraphrased from Culture Code p.18
Culture and the meaning of life
One early discovery from this work has been that culture builders are not messing around when they talk about building culture. This is not casual Fridays, it’s not “15 pieces of flair”, and it’s not “the customer is always right.” Across informants, people in a group with strong culture have a clear sense of what their lives mean, and they are a part of the group because they feel that that group is full of people who are pursuing the same meaning in the same way. In this kind of group, the best way for your life to mean something, the best way for you to work toward a personally important vision for the world, is to be a part of this group and serve it. Culture building practitioners I’ve talked to feel a very strong sense of mission because they see their work as teaching organizations to align their goals and processes with their employee’s literal meaning of life. I started to admire them myself.
The dimensions below are important parts of culture because they are each a part of creating a group of people who feel that they are doing something important for the world that they couldn’t do anywhere else. In the world view of culture builders, that’s the crux of culture.
For a decentralized organization to do the same — and many clearly do so very effectively in their early days — seems very worth working for.
Consensus on nodes, not on links.
My ideal output for this project was going to be a table with dimensions of culture as the columns (outputs) and common culture building practices/tools/levers as the rows/inputs. Progress was promising: I have succeeded in getting meaningful dimensions and interventions. But a framework’s nodes of meaning contain much less of its view of the world than the proposed pattern of relationships that link them. How are dimensions related to each other, and how are they mapped to by their levers?
Although the parts come out clearly, the relationships our informants describe between them are a confused muddle.
-
A company puts effort into valuing its individuals and as a result its individuals feel valued,
-
They continually refine their fit to their ideals and as a result they demonstrate continuous refinement,
-
They share a common identity and as a result they gain a shared identity.
-
Shared values and Adaptability are qualities of a good culture, except that Adaptability is also one of the values that a culture can share.
It continues: Under the mental model of author/informant Coyle, meaningful bonds among teammates causes shared identity, while under Vogl’s, shared identity causes meaningful bonds. Vogl speaks of behavioral standards equally as a means and an end. Practitioners Perlov and Heskett independently identify vulnerability as the mechanism for trust, but Perlov follows recent schools of thought by making the feeling of psychological safety the central driver of cultural development, while Heskett is much more pluralistic in its sources and means.
Should we trash it all as literally nonsensical axiomatic circularity, or forgive it for being an ecosystem of mutually causal factors operating on each other over time in a complex system?
I’m pretty sympathetic to the difficulty of studying complex topics, so I lean towards forgiveness when I can. Heskett ends up as the best showcase of how hard it is to come up with a clear system of relations between the inputs and outputs of culture. I experienced within his system the same thing that I experienced comparing across all of the systems: concepts piled into a mess in which everything affects everything else. “Everything affects everything” is actually a fine insight, and probably a fair gloss of the truth, but it ends up confusing his desired end state. Let’s say it’s fine that adaptability, learning, trust, and transparency are treated as both the inputs and the end states of desirable culture. That doesn’t have to be circular. Maybe building culture is bootstrapping these elements from each other, reciprocally building each other up from nothing. Fine.
But if everything causes everything else then you know as little as you knew when nothing affected anything. It’s not wrong, it’s just not helpful. So unfortunately I can’t say much firm about relationships of these things to each other. I’ll focus on the dimensions, clearly labeling mny ideas and speculations for how they are ordered and how they interact.
The dimensions
Figure 1: The dimensions of culture, by author overlap
Here are the dimensions that constitute culture, ranked in terms of how many authors named them as core elements of strong cultures. From top to bottom:
- Bonding, trust. (9 authors out of 11)
- Shared behavior. (7 authors)
- Shared purpose. (7)
- Shared values/beliefs. (7)
- Shared identity. (6)
- and one runner up: Adaptability (5 authors).
These may seem obvious, but you have to appreciate what is not present. Happiness and fairness/transparency both got cut for only being mentioned by 2/11 authors as core elements of strong cultures. Two is much too hard to tell from noise. And diversity was only centered by one. I drew the a line at “mentioned by at least 50% of authors”, which left out out “supportive work environment” and “psychological safety” (both of those also occurred in the “inputs” list and seemed too actionable to be treated as outcomes).
I’m going to treat these out of order.
Shared purpose
Shared purpose is a shared sense of what the group is trying to accomplish in the world: the world that the group is working toward. People often get a sense that their lives are meaningful from identifying a purpose in the world that they want to serve. And when your peers serve the same purpose, you get meaning from their activities as well and you get access to a signature of strong culture: the feeling that you are an integral part of something special. When a group has a shared purpose that purpose can be used to define the group as a coherent social unit oriented around something that matters for the world.
Most authors identified shared purpose as a defining element of culture, and of the “shared X” dimensions it is tied with “shared behaviors” for the most commonly mentioned. Those who didn’t mention it seem more likely to have mentioned the adaptability quality instead, but I don’t have an interpretation for that that is better than “noise”. Several authors that I lumped into shared purpose said “shared goal” or “common goal” instead. If it’s a mistake to merge “purpose” and “goal” under one dimension, it’s because a goal can be a bit more micro and achievable than a purpose: “provide the best customer experience in the auto parts industry” can be a goal and not a purpose. Some authors seemed comfortable with this, but I merged them because of the overall sense I got that the more lofty and purpose-y the goal is, the better is represents strong culture.
I’d like to treat shared purpose solely as an output of culture building work. But because of the endemic circularity and mutual influence of all of these concepts, it’s worth entertaining it as an input to culture rather than a quality of it. When your peers serve the same purpose, you have a reason to serve them, bond with them, and work and learn with them. This makes shared purpose a credible supporting mechanism for the other qualities of bonding, shared behavior, shared identity, and beliefs.
It’s not a bad perspective to treat purpose as one of the more designable or actionable of the dimensions. I’m not saying that because it is easy to give a person a purpose that they deeply identify with. That’s hard. But mutability is not the only thing that makes a quality designable. In the case of shared purpose, communities can actively attract, select, or self-select for people who have already internalized their shared purpose. Decentralized communities, though often technically open to all, seem very comfortable understanding themselves as selecting for a certain type of person. Defining “certain type of person” as “person with this certain purpose” aligns selection tools in the direction of strong culture.
I have more to say about this in the next post, but briefly a community can work toward shared purpose by
- publishing a self-description that includes a vision for the world,
- setting an expectation that participation is based on internalization of that purpose, and
- using the community’s purpose to target recruitment and drive the collective’s actions.
The sooner this happens in a community’s lifecycle, the easier it will be to reach and maintain a clear consensus view of why the community exists.
Shared values and beliefs
Values or beliefs are the qualities that communities use to narrow down the means they are using and concrete outcomes they seek. Values and beliefs act a bit like purpose, as a source of constraints on the strategies that a group will center on. But they are different from purpose in a bit of a “how” sense and a bit of a “what” sense. Two people may have a purpose of using decentralization to empower the world’s people, but what that looks like is very different if you constrain that outcome to represent diversity, equality, and fairness values, or if you have a conveniently simplistic worldview focused entirely on, say, freedom (“one value to rule them all”). But like purpose, people use their values as a source of life meaning (“without my integrity I am nothing”), so a group with shared values is one whose members can involve each other in their pursuit of meaning in life. That makes shared values a driver of precisely the kind of soft, insubstantial, but compelling collective efficacy that makes culture a worthwhile things for groups to pursue. That is interesting because creating shared values, despite being so important, is so far from what decentralized organizations are by definition good at creating (like mechanisms and voting systems).
Above and beyond the value of a community sharing values is the value of sharing “good” ones. Informants shared no common core of values to strive for, but although I saw practitioners harnessing competition, they ultimately put much more emphasis on cooperation, teamwork, trust, fairness, openness, and adaptability (which was mentioned enough to be a runner-up for its own dimension). To keep a statement of values from devolving into a list of pleasant words and empty slogans, a community should be selective in focusing on the few of them that most bubble up from the membership.
I combined shared beliefs and shared values into one construct, but they are a bit different from each other. There is a bit of global-to-granular thing: “purpose is to goal as value is to belief.” But that’s not exactly right.
Anyhow I’m a bit uncomfortable merging them. I did it anyway with encouragement from the data. On their own, values and beliefs are identified as important by the fewest authors. They are both mentioned by only four of eleven authors. But looking more closely, only one author commits to them both as distinct concepts (Figure 1). Every other informant who identifies one as central omits the other. This comes off to me as evidence that they are being treated as alternatives or synonyms, or maybe as unsatisfactory alternative words for a cluster concept that each imperfectly represents. At the end of the day, the words “purpose”, “behavior”, and “identity” are already so vaguely overlapping, and already occupy enough of meaning space, that the negative space leftover for the more marginal “values” and “beliefs” may be an awkward union of sundry little semantic bubbles that remain.
As I’ve mentioned, informants have a common understanding that these qualities of culture bootstrap each other: they aren’t just outputs, they are also inputs to each other. What does it look like to treat values and beliefs as “inputs,” as things a culture builder can directly tweak to affect other qualities? They are both constraints on the strategy and behavior that one will use to accomplish a goal. So when people have strong shared values and beliefs it may be that much easier to converge on shared behaviors. And to the extent people identify with their values, a community with strong shared values is selecting for people who are prone to finding a shared identity.
Shared values and beliefs aren’t bad as inputs. Like shared purposes, a decentralized community can converge on shared beliefs and values by stating them concisely and using them as a litmus test for membership and a basis of shared identity. Still I think of them as a less effective lever than shared purpose because it seems like an organization should have one purpose but can have lots of values and beliefs. More qualities to match on or recruit for means less clarity about who is or isn’t an ideal fit. Additionally, values and beliefs are a bit secondary to purpose in determining a group’s activities. Purpose literally defines a community’s activities.
Shared behavior
Shared behavior is a lot what it sounds like in the abstract, and is a large umbrella, including organizational processes (convergence of behavior on how the organization operates in a way that communicates the legitimacy of those processes), rituals (convergence of behavior around actions that reinforce collective identity), and teamwork (shared behavior in a different sense: not common behavior of many people, but coordinated behavior of many people to a team-level “behavior.” The power of “flow states” for happiness and job satisfaction give one vivid mechanism for how behavior/action becomes a tool for personal life meaning. Sharing those states with others, and generally working in close concert with others on a community’s activities, intertwines members’ sources of meaning with a contribution above and beyond the binding effect of shared purpose, beliefs, and values. For example, behavioral evidence in evolutionary anthropology shows that cooperative effort and synchronization of behavior are common elements of ritual. According to this literature, the shared behavior of a ritual literally expands a person’s experienced sense of where their self ends in space to include the bodies of others.
Shared behavior was next to shared purpose as the second most mentioned element of culture that our practitioners independently identified as a core element of strong culture (7/11). Nearly every author who identified shared identity or shared values also mentioned shared behavior (as opposed to shared purpose and beliefs, which occurred more with each other than either did with behavior)(This is just an observation, I don’t have a satisfying interpretation).
As a lever or input (rather than output or quality), shared behavior is harder to use as a basis of boundaries/selection/self-selection than shared purpose or values, but it is easier to change. That said, changing the behavior of a group of people, though easy relative to other shared qualities, is hard in an absolute sense: it takes smart design, a lot of buy-in, monitoring, reinforcement, and iteration. Habit, routine, and the actions of others are very powerful determinants of behavior, so when bad behaviors are entrenched, they can be very hard to change. Despite steady advances in the behavioral science, the best baseline predictor of a person’s behavior in the future is their behavior in the past.
The best tool for fostering shared behavior in a community may be teamwork. Dividing an organization’s activities across many stable small teams forces people to coordinate their individual behavior toward collective ends in a way that the work of behavioral monitoring and reinforcement can be distributed to the membership. Experiencing success as part of a team reinforces the value of shared behavior, and emphasizes the idea that the group is the individual’s best vehicle for pursuing their personal sense of life meaning.
Bonding/trust
Bonding is the quality of feeling kinship with another person. I distinguish it from shared identity below because identity as I talk about it here is a relationship between the individual and the group with bonding is a relationship between individuals. A closely-knit team of four people might have four manifestations of group identity (each of the four identifying as the whole group) but six pairs of bonds (a fully connected network of 4 nodes has 6 links). Feeling a bond to many people in a group helps associate one’s personal sense of life meaning with the group because we can rely on people we are bonded to to support us in our goals and challenges. Many authors named trust, which I lumped with bonding because most talk of trust was as an output of a bonding process, usually because of work in an effective small team. With trust in place, people can reorient effort from monitoring each other to producing with each other. Under a condition of trust, another’s actions toward the community’s purpose are a contribution to one’s own personal sense of meaning within the group.
The importance of this dimension is apparent from the fact that it was mentioned by over 80% of authors: 9/11. This is the closest to consensus I encountered in this work, either on the side of cultural dimensions or inputs to culture change.
Bonding and trust are feelings. However hard it is to change what someone does, it is harder to change how they feel. So it makes sense that I treat bonding/trust as one of the least actionable of the dimensions. It is much more a result of culture building than lever or input to it. Still, like purpose, values, and identity, bonds/trust can be the basis of selection for new members. The mechanism is a bit different. Purpose, values, and identity can be used for recruitment via “broadcast” means: a group’s impersonal public description of itself can engage an aligned person even if they have no prior relationships in that community. By contrast, bonding and trust aren’t suited to selection by broadcast (I can’t tell you in a blurb that you trust me) but they can be selected for by focusing growth and recruitment on the friends and colleagues of existing community members (word of mouth growth). A newcomer begins in the community with one or two bonds of trust with those who recruited them, and their integration in the community can focus on increasing their count of interpersonal connections, by including them in a team, supporting social events, or with other levers here.
Because teamwork and co-work are widely recognized as a primary mechanisms of bonding, everything that fosters teamwork (clear roles, emphasizing prosocial characteristics, leadership, and inclusive decision making) can contribute indirectly to bonding. Two authors also named vulnerability as a primary cause of trust and bonding, which is worth mentioning because vulnerability can be displayed unilaterally by anyone, making it actionable at the individual level.
Shared identity
Shared identity can mean three things:
- that people in your group personally identify with the same broad social labels as you (Being in the same ethnic group as others)
- that a social label has emerged that is synonymous with the group’s membership (being a member of metagov.org makes me a metagovernor. There is no other way to be a metagovernor, and the label didn’t exist before metagov).
- that your personal sense of where “you” end includes the physical bodies of other people.
The authors in this review tend to mean the second, but all three are a source of aligning an individual’s pursuit of meaning in life with the activities of a group of others who are pursuing meaning for themselves. Identity is powerful, and many fundamental theories of human culture and sociality take group identity as a primitive. Although it wasn’t any kind of majority view, at least three authors mentioned how in-group psychology, by supporting group identity, becomes a powerful tool for engaging culture. This may sound dangerous: a valued and humanized in-group seems to imply a devalued and dehumanized out-group, but several authors independently described a mechanism that seems to avoid this: they keep the basis of shared identity more on the community and its positive shared values and purpose, and less on the negative characteristics of outsiders. For a group to feel that they are important, it is not necessary to feel that all outsiders to be awful.
The third sense—the evaporation of boundaries between you and other people—is the most eerie and unfamiliar, but it does exist and, where it does, is an especially powerful driver of culture. It is well documented that rituals around the globe target this sense specifically, and it may be a powerful target for experimentation among decentralized communities looked to intentionally cultivate a strong sense of culture. The same research community cited for the power of shared behavior and ritual identifies one effect of ritual as the creation of a visceral feeling of oneness with others in the ritual.
Shared identity was identified as a core quality of strong culture of 6/11 authors, just over half. Nearly every author who named shared identity also named bonding/trust: which makes sense as the two seem like they would be strongly mutually self-reinforcing. Another observation is that the six authors who named shared identity as core were the six who associated culture with the largest number of qualities: every author who named shared identity reported a mental model that names at least 4 of the 5 qualities we report here, while none of the authors who converged on three or fewer included identity in their three (or fewer). One interpretation of this is that shared identity is squarely in the centroid of mental models (being named by the informants with the most “typical” understandings of culture). Regardless of the interpretation, shared identity seems to be a typical quality of the most maximalist and pluralistic mental models in the study.
Ritual, teamwork in small teams, and use of collective language in group communications are all straightforward tools for cultivating shared identity. But overall I understand this dimension as less of an input lever and more of an output cultural quality. Communities that use all of those tools—ritual, teamwork, “we” language, and more can still easily fail to develop collective identity. It seems likely that shared purpose and bonding are necessary pre-requisites to the more powerful and encompassing sense of shared identity, especially the deeper and more visceral varieties of it. Ultimately, shared identity seems more “downstream” from the other qualities because it is hard enough to change what someone does, and even harder to change what they feel. On the other hand, identity is like purpose and values in that it can be selected for during recruitment, but only in its first sense, which is in terms of broader pre-existing social labels. The deeper sense we target here, perfect overlap between group membership and its associated identity group, by definition can’t be a basis for self-selection: if members can adopt that identity before they are members of the group then there isn’t perfect overlap between the group’s membership and it’s community of identifiers.
Runner up: Adaptability
Adaptability is an ability of a group to change with changes in its membership, environment, purpose or other elements. Practitioners also report it in terms of a culture of learning, a culture of open-mindedness, a culture of continual refinement/improvement/quality, and a “flat” culture of openness to the contributions of all members.
I honestly don’t see adaptability as a dimension or quality of culture in the way that the others are, but it was mentioned by enough authors (5/11, just less than half) that it seems important to include tentatively. It isn’t clear here that “adaptability” can be a property of a healthy culture in any more meaningful way than everyone internalizing the value of adaptability. Put another way: it may be that when we talk about the dimension of “shared values,” we aren’t talking about just any values, but “good” ones like “adaptability” (like commitment to learning, open-mindedness, and flatness named above). Under that way of looking at it, the cleaner framework deletes “adaptability” as a standalone quality, and instead amends “shared values” to something like “shared values that are good” or “shared values from whitelist X”).
One strength of including it is that it seemed to temper the vulnerability of shared purpose, values, beliefs, behavior, and identity to ossification. Many of these authors see a healthy community as something that changes and grows, and that means leaving space for a community’s shared commitments to change and grow as well.
On the list or not, it is hard for me to rate adaptability as actionable because I’m still not sure how it manifests except as a shared value. Fortunately, there are lots of levers for making culture work actionable, which we will get to in the next post.
Thank you for your ideas and engagement.