I was recently asked why I call myself an online community “architect.” A fair question. My use of the label “architect” is meant to highlight how intentional communities, such as virtual organizations, can design, assemble, and use cultural practices to become more effective at achieving their goals and their vision.
The “unintentional” communities that we find ourselves belonging to (or resisting) in our society, from our neighborhoods and schools to our language groups and nations, use cultural practices that are grounded in longer histories and broader social structures. These represent the culture we are submerged in from birth; often they are the practices that represent our sense of identity and position in society.
These practices (and associated beliefs) are what anthropologists call “culture.” They remind us that much of this is learned and then used unconsciously, as habits and behaviors that come naturally to the enculturated members of the society. Intentional communities are made up of members that carry their own cultural habits from their society. On top of these habits, an intentional community agrees to add practices that are mutually agreed upon and available for discussion and revision.
Some intentional communities are created as counter-cultural groups, where their practices are designed specifically to conflict with those of the larger society. But many more intentional communities, from workplaces to voluntary societies and virtual organizations simply want to define their core values and to design governance practices to support these.
Often, corporations and networked communities fail to recognize that they need to develop an intentional culture to support their collaborations. They rely on the unconscious cultural habits of their members and on imposed management schemes to compel participation. Management practices can seem more effective and cheaper than the expense of building cultural practices and governance. However, in situations where corporations need to pivot quickly, or where virtual organizations need to rely on volunteers, there is no substitute for shared values supported by governance practices.
The literature on the costs and benefits of a robust corporate culture is quite large. But how can organizations gain expertise in developing intentional cultural practices? Just as a building architect is used to create an efficient and effective envelope of space, a community architect can help an organization create an envelope of practice. Every effective corporate culture is based on shared intention and honest design. It is the role of the community architect to become familiar with those design patterns that can help a new virtual organization build a robust cultural fabric to support its vision and its goals.
Photo Credit: “On The Saturday Before That” by Thomas Hawk on flickr
Building practices for an emerging governance gives your virtual organization community tools to explore how they want to govern their own participation. Too often, governance planners start by thinking of structures: committees, boards, and assemblies. The structures will emerge more organically if planners start with practices and their logic.
Twenty some years ago, anthropologist/sociologist Pierre Bourdieu wrote Le Sens Pratique (Logic of Practice). He noted that cultural activities were not simply reasoned, but also became embedded in their practice. Their performance included tacit knowledge that was not available to the performer, having been learned directly through the body. The cultural norms and skills are carried by their members without the need for reflection or conversation. This is precisely why “culture” is often considered as difficult to change.
How does this theory inform the work of building an intentional community governance? Mainly, I will propose that governance needs to “make sense,” not just as a reason-able activity, but as a practice that feels right. Because this community is intentional, its practices are available for reflection and conversation, and change. This is the real difference between culture out in the world and the culture of a company or a virtual organization. Any organization that loses the ability to direct its internal culture is trapped in a single loop of progress and error and will never learn its way out of this. Your culture belongs to you, and not you to it.
How can practices feel right? How do you know if a practice feels wrong? The main way is to create a small list of provisional core values. These become the touchstones that members use to judge how right or wrong a proposed practice feels. If “maximal transparency” is a core value, then a practice that hides disagreement is going to feel wrong. If “diversity of approach” is a core value, a practice that opens up to a great variety of inputs will feel right.
Don’t feel like these provisional core values will need to be written in stone. Make one of them “community owns its values” and encourage the members to embrace and celebrate them, changing them when they want to.
Remember that governance is not simply about decision making. There are a lot of expressive opportunities that it can enable that will help your members to embody the organization’s culture through their interactions. Governance needs to encourage leadership and resolve conflicts. It should promote best practices and reward service.
Are you ready to gather your fledgling community together and start planning a governance approach? Get them to outline a handful of key values first. Then challenge them to find practices that uplift and promote these. Then you can add in use cases and scenarios for them to build structures that use these practices. Before you know it, your community will have a first draft of their bare-bones governance scheme.
Photo Credit: Elgin County Archives, Wallacetown Women’s Institute fonds on Flickr
Decision making for your virtual organization needs to optimize the decision process and impact. If your organization relies on a wider community of unpaid volunteers, then you will need to find ways to involve this community in your decision process. Where decisions need to be made on a day-to-day basis, you will want to have paid staff with the authority to make these, and to also be accountable to the executive body of your virtual organization. Involving the wider community usually involves two complementary modes of decision making: election and consensus. For large organizations these two modes are sometimes used together in a multi-step process of delegation and consensus.
Why is consensus important? What type of consensus is the best? How do I create a culture of consensus? Previously, I outlined the arenas where staff and volunteer decision making occur, here I want to focus on the role of consensus. Consensus is primary important as a decision process where this can positively impact the quality of the decision and/or the efficacy of its outcome. The road a consensus decision opens up the discussion to include every member’s perspective and intuition. This process brings in the full range of the group’s knowledge to bear on the issue. During this discussion, aspects of the problem may be illuminated that were previously obscure. The result can be a decision that is stronger or more astute. Even when the discussion leads to a compromise, that compromise can be based on real-world limitations, and so, might avoid trouble after implementation. Where the implementation of the decision will require the active participation of the larger community (e.g., a decision to support a certain data/metadata format) consensus carries an invaluable mark of community involvement and ownership for the decision.
Just enough of a consensus
Absolute consensus may be unreasonable, given the range and interests of various stakeholder groups in the membership. If this is the case, and it would probably become evident during the start-up of the community, then some sort of “working consensus” (or rough consensus) might be a reasonable alternative. This would be a type of super-majority that would allow for a few opposing views to be not included in the final decision, but to be included in a durable report of the decision process as a minority perspective. The logic is to be able to move ahead, while maintaining the conflict that emerged in the decision process on the surface of the final decision. This type of working consensus would need to have at least a 81% majority: in a group of 20, no more than three people can disagree with the final decision. The goal is always to achieve a total consensus, with the working consensus as a fall-back.
Consensus decision making requires a consensus-aware culture for interaction within the group. There are some established cultural practice guidelines for consensus organizations. One of my favorites is the Seeds for Change organization in the UK. The consensus decision process challenges each member to listen fully to the arguments, to state their own position clearly, and to be aware that they have more than just an option to support or block a decision. A member can also abstain, withholding their outright support and refusing to block a decision. It is important in the discussions leading to a decision that the facilitator (often a staff member) is also a process mentor, reminding members of the need for open minds and hearts in the process, and a clear-headed, well-founded motive when a member decides to block a decision.
Consensus decision making does not directly scale beyond a couple-dozen members. In a larger community, each stakeholder subgroup (this requires a fractal subsetting to sub-groups of no more than a couple-dozen members) is granted a representative to an executive council where the consensus discussion is held. At the point of a vote, the representative returns to their subgroup and outlines the issues and the proposed decision. Once a consensus is acquired within the subgroup, this is carried back to the executive group.
The process of arriving at a consensus is at the same time a process of listening to the best ideas and the strongest fears of the community’s stakeholders, and a means to forge a better solution as a final decision. A decision based on consensus carries the trust and the will of the entire community.
Your virtual organization will work harder and better when everyone has a say about policies, and not just about procedures. Policy setting happens best in the second loop of a double-loop governed organization. That way the policy can respond directly to the vision of the organization.
A friend, a sysadmin at a major research university, was on his way to DEF CON when I ran into him on the street. He enjoyed hearing about the new security hacks he would be facing, and, as usual, he was not happy about the security measures on his campus. After describing how schools, departments, and centers had managed to grab the right to host their own Internet content—with predictably inconsistent results—he concluded that even his own department could use some new policies.
“Unfortunately, I can’t make policy,” he said, “I can only recommend procedures.” Faculty make policy, he explained. Staff implement this as well as they can. “I can talk all day, but some new PhD who thinks he knows enough will want to run his own server and connect into the department’s databases on my servers.” My friend has far more knowledge about computer security than he can implement, and he sees trouble ahead against which he cannot defend by making and enforcing better policies. When the system fails, when data are stolen or lost, he will be asked to explain and tasked to repair the damage. When the fan and the feces collide, he only hopes the precious work of graduate students is not collateral damage.
I could just recommend that you do not follow the model of a large academic department in a research university when you create your virtual organization. However, I doubt any of you (particularly those of you who have worked in a large academic department) had plans to do so. My point here is that every member of your organization can contribute to its policies and help defend it from failure.
Here are three ways you can make policies more effective for your organization:
1) When you create an executive panel or committee to make policy, be sure that this body is well connected with the larger membership.
2) Use a federated election process to preserve the voices of minority factions and edge groups, and empower contributions from across the membership.
3) Get additional feedback from the membership before you implement a new policy.
As a bonus, you will find that policies that are enacted with these practices are likely to be followed with greater rigor and care than those that simply appear in an email from the top.
In March, 2013, David Reid at Palantir tweeted a quote: “Drupal’s community is a bus driven by the collective screams of its passengers.” At first glance, this seems really unfortunate. But on reflection, there are some great things going on here. First, it’s a bus, not a limo. Second, it’s filled with passengers. Third, they are passionately engaged in the destination of the bus. And fourth, the bus is responding to them. In short, the Drupal community is doing a lot of things right.
This tweet took me back to a sunny afternoon in Berkeley at BADCamp. I was in front of Dwinelle Hall (where I first met my wife way back before there was an Internet) talking with Jacob Redding, who, at the time, was the Executive Director of the Drupal Association. I had recently finished Jono Bacon’s book, The Art of Community, taken from his experience helping to lead the community of developers who support the Ubuntu distribution of Linux. I mentioned to Jacob that, compared with Ubuntu, Drupal did not seem to be grounded in a common vision that was written down as a statement somewhere.
Jacob said that the Drupal community doesn’t have a single vision statement but the leaders of the project each may have their own. The Drupal project does have a mission statement (drupal.org/mission). In the project there is a lot of non-visible leadership. The Drupal Association is a good example of that. The Association does not guide or direct the project, but it certainly does watch what is happening with the project and provides the necessary support and services to either sustain momentum on areas in the project or encourage more involvement. “You can say the Drupal Association is leading from behind,” he said.
There are visionaries across the Drupal developer network, dozens of active people who are building new tools and pushing for better code. There are occasions in the year (at big Drupal Camps or at DrupalCon) and moments in the development cycle (sprints and telecons, in discussions on Drupal.org and elsewhere) where these leaders would come forward with their visions, and, at some point, a decision would happen. But the Drupal Association, which runs DrupalCon and hosts Drupal.org, is not out in front breaking trail for these decisions.
I called Jacob last week to get some clarity on this “leading from behind” technique. I might be involved in a grand experiment on governance that the NSF has fostered. It’s called EarthCube, and it proposes to corral all of the geosciences and environmental sciences under a new organization that can manage and promote the multi- and transdisciplinary sharing of data and knowledge. To bootstrap this effort, a seed of governance will be funded. In the first months, this governance kernel will need to get out in front and open up online opportunities for thousands of scientists to step into new leadership positions and get to work. In the subsequent months, the funded governance kernel will need to step back and lead from behind as the community finds its own governance footing. In the final phase, the funds for governance will be passed directly to the community, and kernel will become a front office. But this transition from leading from the front to leading from behind is truly experimental.
“Part of it is allowing chaos to reign,” Jacob told me. The point, I would propose, is to not spend a lot of effort (and time) establishing and then defending a set of top-down decisions. He continued,“a lot of things go wrong when you have a benevolent dictatorship that says ‘this is the way’.” When the community feels the need to make a decision, the lack of a top-down process enables people step up to offer solutions.
Behind the scenes, Jacob noted, the leaders of the Drupal project pay very close attention to all of the blogs and the threads on discussions and comments, and keep track of who was working well with others and who was influencing their peers. It would be vital to have these people in the room for important decisions. This internal eco-system is fostered through support for travel to Drupal events, and through email lists and other back-channel communication. Out in the front channels, the Drupal bus riders are shouting out their visions and their peeves.
The other eco-system to which the Drupal project and Organization pays close attention is the larger CMS world, where big corporations and government agencies are looking to make long-term investments. “This way we can counsel corporate users, and use their feedback to counsel Drupal developers,” Jacob noted. One big part of leading from behind is also watching the road ahead.
If you’ve familiar with object-oriented coding using any agile methods, you understand that the customer is at the center of the software development effort. The customers problems, her needs and goals, laid out in stories that the programmers return to every day: these keep the software from feature creep and UX failure. Delivering working software in small increments helps the customer reveal the moment they don’t need or understand something. The programmers can toss out that new feature and go back to the customer stories again with fresh insight.
They are testing all the code every day and releasing every week. So there’s never a time when the changing the software costs more than the last time. They are free to pivot toward some new capability that just might provoke delight in their customers; something they would never have tried if the cost of change was growing. Finally, they are solving their customers’ problems and building into the software new opportunities that might not have existed anywhere else in this way ever before. OK, so it rarely works out that well. But the philosophy of agile code development is sound and it offers valuable lessons beyond writing code.
Virtual organization community leaders would do well to consider how the Manifesto for Agile Software Development might be tweaked to be a working Manifesto for Agile Community Development:
Individuals and interactions over processes and tools
Working software (working volunteers) over comprehensive documentation
Customer collaboration (member collaboration) over contract negotiation
Let’s look at these one by one. Individuals and interactions are the daily work of the community leader. Staying in touch with members is more important than whatever tool (email, forum, listserve) is picked to accomplish this. Working volunteers build value and direction for the virtual organization. Having a grand, detailed volunteer guide is less important than having a team of members who want to move the organization ahead. Member collaboration means giving the membership as much ownership as possible. Staff are around to take care of day-to-day business, but budgets, planning, and the real work of the VO belong to the members. The staff are the sails, the members are the wind. Responding to change means having a full double-loop governance system and pivoting this to stay ahead of changes surrounding the organization. If you’re only following a plan, you are already lost.
Agile: it’s not just for software anymore!
photo credit: From flickr user magia3e http://www.flickr.com/photos/magia3e/6236962059/sizes/z/in/photostream/
Here is a talk I gave at the recent (Summer 2012) ESIP Federation meeting. Sharing Creativity:
I am hoping that this talk will lead to some conversations over the potential for virtual organizations to achieve, with more efficiency and effectiveness, a capacity for creativity and predictable innovation. This capacity—in large part due to Internet-enabled capabilities for coordination and collaboration— can, I believe, rival (at various scales) the capabilities of dedicated R&D facilities/programs such as Bell Labs and Xerox PARC on the corporate side, and the Manhattan Project and the Apollo program on the government side.
Large research and development operations such as these were built as national laboratories, with hundreds or thousands of employees and forefront facilities. They were designed to assemble a critical mass of talent and direct this toward innovation. They were also enormously expensive: the Apollo program had cost more than $25 billion by 1973 (in 1973 dollars). The successful ones are rightfully famous.
Today’s top technology companies (Apple and Google, for example) often add to their innovation potential by buying forefront start-up companies, as much for their talent as for their technology. Their goal in a highly competitive market is to own enough talent, enough intelligence, enough creativity, to stay ahead of their rivals.
The basic business-school rule for improving the odds for successful innovation is to assemble a requisite variety of knowledge: a range of knowledge at least as large as the problem being tackled. The three ways to do this are the following: Hire it (add to your team); Grow it (reeducate your team); and, Buy it (purchase a rival company/team). All of these methods assume that you need to own the requisite variety of knowledge.
Science, on the main, has only one rival: the unknown. Scientists are relatively free to seek out new collaborators from anywhere. And, through Internet-based services, they are now enabled to become collaborators everywhere. This is one reason why the NSF has been promoting virtual organizations and research networks as the future of science collaboration (instead of building new centers at institutions). A good part of the potential that virtual organizations offer government and private funding agencies comes from a new logic for innovation: assemble and share the requisite variety of knowledge. With the right sort of organizational governance and funding, a virtual organization can achieve what the older “think tank” R&D centers could: predictable, successful innovation.
There are some social aspects of the ESIP Federation that might be key to this capacity for creativity. These aspects are not secret, however, and can be fully copied and applied in other arenas. They are also not expensive (the Federation budget is remarkably small), but they are of great value, in that they have been worked upon by dozens of volunteers over the course of more than a decade.
Virtual organizations (VOs) come in many forms and sizes. The science of building and managing VOs is still being explored. There are many examples of early failures, and only a few examples that herald their potential success. Members of virtual organizations need to be sufficiently engaged to build collective intelligence. Take a look at the YouTube video and let me know what you think.
I once spent several years studying how communities assembled festival productions in the streets of their towns and cities. This includes almost three years in Kyoto, Japan, and other experiences from South India and the US. The impetus behind this (no, it wasn’t the beer, but that helped) was to see a form of cultural production that was forced to renew itself regularly, and to describe and explain how that cadence of activity could sustain an enormous volunteer effort every year. I was also interested in festivals as they often included activities and behaviors that were not available or advisable on the streets during the rest of the year (aspects of comedy, nudity, violence, politics, joy, and sex: stir together and stand back). The time and space of the festival did not just occur on the street, it transformed the street. Much the same could be said for the impact of the event on the lives of the participants.
I also saw a lot of festivals that failed to transform the street, that had been, somehow, reduced to parades and pageants, to a semblance of their former glory. In these events participation was simply another chore. As such, participants required payment. Volunteerism declined. The vestigial energy of the festival was provided by the music and the movement. But the ability to transform the lives of the participants was lost.
The craft and the logic of a festival event are in many ways similar to that of any virtual organization that hopes to engage its members. The face-to-face meetings or your organization may not achieve actual festivity, but the power of your organization to transform the capabilities and empower the hopes of its members should never be forgotten.
Several years ago I wrote a piece for the Kyoto Journal, “Imagine the festival as a building,” in which I asked readers to imagine their community-based organization (festival or virtual) as a building, a house that is rebuilt from scratch every year. The main messages of this article outlined how the process of designing and constructing the house each year supported the neighborhood in vital ways, and explored how this event would certainly be destroyed when its festival logic is violated, even if an event kept happening.
The essence of a festival logic is honest and free voluntary participation and expression. A radical form of intimacy emerges. Some scholars tie this back to ancient forms of human physicality (e.g., Cox 1969; Stallybrass and White 1986); however, Anthony Giddens (1992) anchors this form of intimacy also as a hallmark of current modernity. We all want to achieve more moments of intimacy.
Festival logic informs the whole process of design and performance. When you incorporate a similar festival logic into your community-led virtual organization you unleash new levels of member engagement. Street festivals generally focus this energy into artistic forms, while your organization might focus this on product development, creativity, and innovation. If you can imagine your VO as a street festival, you might discover its festival logic. If you cannot, then you might want to ask yourself how to add this logic to your organization.
What can you do when your research team says, “This is way too complex.”
When your government agency or university laboratory looks to innovate in a world where multiple/large data inputs are coming on line, how can you stay ahead of the inherent complexity of the systems you are creating/interrogating? One way to look at this problem is through Ashby’s principle/law of requisite variety. A principle of cybernetic management, requisite variety notes that unless the control system has at least the variety of the environment it controls, it will fail. Which actually means that some part of the environment will be controlled elsewhere. Elsewhere is also where innovation happens; because unless you can corral the inherent variety of the problem you face, it will seem too complex for your team to innovate a response (Kofman ). You can either go out and hire a bigger team, or you can borrow enough requisite variety just long enough to bring your own team up to speed. That is a great use for a virtual organization (VO).
Theorists of knowledge management have applied Ashby’s law in various modes, including a thread of interest in what is called a “learning organization” and mode of business communications management known as “systems thinking.” [There is a great amount of information about this available at the Society for Organizational Learning http://www.solonline.org/.] The point they make is that the team you build to tackle a tough problem needs to have enough of a portfolio of knowledge and skills to address all parts of the problem. Not only that, but they need to communicate their skills and knowledge to one another so that each team member shares in this collective intelligence. Andrew Van de Ven put it this way, “Requisite variety is more nearly achieved by making environmental scanning a responsibility of all unit members, and by recruiting personnel within the innovation unit who understand and have access to each of the key environmental stakeholder groups or issues that affect the innovation’s development.” (Van de Ven, 600).
Virtual organizations include online communities, research collaboratories, open source software programmer collectives, and other groups in a great variety of arenas and professions. What they offer is an open network of common interest and complementary talents. When your business or agency is looking to innovate in a world where data are more plentiful than insights (Abbott, 114) then it makes great sense, in terms of time and effort, to join a VO and gain enough requisite variety to conquer complexity and kickstart some innovation.
Abbott, Mark R. (2009) “A New Path for Science?” in The Fourth Paradigm: Data-Intensive Scientific Discovery. Hey, Tony, Stewart Tansley, and Kristin Tolle, eds. Pp. 111-116. Redmond: Microsoft Research.
The 5 ways that double-loop governance can save your organization from itself
Institutional guilt (see Part 1 below) is routinized violation of your organization’s values, vision, rules, or policies. It is symptomatic of dysfunctional communication strategies inside an organization. It leads to distrust of staff and disengagement from the organization’s vision. Staff and volunteer disengagement/disenchantment is a prime reason non-profit organizations fail (Duckles, et al 2005).
Institutional guilt is something that will ruin your virtual organization. It poisons the culture and it drives away volunteers while it demoralizes your staff. Implementing double-loop governance is a good way to build in protection against institutional guilt. You also need to be sure that your employees and volunteer committees do not fall into the trap of violating your own values and policies for some immediate purpose. Double-loop governance opens up learning capabilities and communication channels to help limit and repair occasions where volunteers or staff do stray from your organization’s vision and values.
1. Double-loop governance makes every member a caretaker of the vision and values for the virtual organization.
Your values are not just a bulleted list on your website nor a poster on the wall. They are the deep logic of why your organization exists. When you create the knowledge loop that includes questioning and reaffirming your values into every decision, then your staff and volunteers can celebrate these values. Membership includes embracing the values, and entering into the ongoing conversation about them that keeps them current and vital.
2. Double-loop governance makes a virtue out of transparent decision making.
Transparent here means available to all members (not necessarily public). Practically, transparency includes time and place availability. Members are told when and where a decision is being made. For a virtual organization, this might be a set period of time to edit a certain wiki, or a set period in which to vote online. The management of critical-path decisions may (and usually should) devolve to active subgroups charged with delivering the outcomes. These subgroups need to maintain their own transparent decision process. A great example here is Wikipedia, where each entry contains the edited text, a history of edits, and a discussion page about the text and its edits.
3. Double-loop governance brings conflict to the surface.
Conflict avoidance is a major source of “unusual routines” (Rice and Cooper 2010) in general, including those that create institutional guilt. Conflict can arise in many forms. Personal issues surrounding time commitments, responsibility and authority, and expectation management cannot be avoided through double-loop governance alone, but they can be openly addressed and resolved in a manner that promotes reflective learning among those involved. Evaluation conflict avoidance happens when tests of deliverables are either postponed, curtailed, or done in private. Double-loop governance supports open and thorough testing, and the disclosure of competing interpretations. Conflict is rapidly promoted to the surface of discussions, where voices of dissent become available to all members. Resolution is commonly achieved through a working consensus, not 100% agreement, but something more robust than a simple majority. Conflicts over the underlying assumptions of the organization can result in new values and a new vision: the organization is free to pivot toward a novel direction at any time.
4. Double-loop governance accelerates failure to ensure success.
Remember that double-loop governance supports double-loop learning. Single-loop learning focuses on avoiding failure. Double-loop learning focuses on using failure to recalibrate the underlying assumptions of the activity, this promotes the act of failing as a learning device, and a logic of rapid iterations of activities with open testing. In software development efforts, double-loop governance actively supports agile development decisions. In all endeavors, the ability to fail quickly and recover takes the fear out of trying new strategies. This almost guarantees a better final result.
5. Double-loop governance supports do-ocracy and emergent leadership.
While not all double-loop governed organizations are strict meritocracies, the best find ways to recognize and reward achievements and contributions. One of the benefits of the network effect is an ability to reach out beyond the founding team and find people who have similar interests and valuable skills. As the network expands, the chances of encountering tomorrow’s leadership improves. When these people become engaged in activities and outcomes, they need to have a clear path to leading subgroups and then larger groups, and ultimately the organization.
Final Thoughts: Double-loop your organization and forget the guilt
Remember that decisions that don’t get made by the people who are supposed to make them get made anyhow by the people who need them. Even the decision not to decide today is made by someone. When decisions are guided by the values and vision of the organization, when the process is transparent, when the conflicts appear on the surface, when failure is just another chance at success, and when leadership opens up in front of those who have proven their worth: that is when institutional guilt has no purchase on the logic of your organization.
Duckles, Beth M., Mark A. Hager, and Joseph Galaskiewicz (2005) “How Nonprofits Close: Using Narratives to Study Organizational Processes.” Pp. 169-203 in Qualitative Organizational Research: Best Papers from the Davis Conference on Qualitative Research, ed. Kimberly D. Elsbach. Greenwich, CT: Information Age Publishing.
Rice, R. E. & Cooper, S. (2010). Organizations and unusual routines: A systems analysis of dysfunctional feedback processes. Cambridge, UK: Cambridge University Press.