pdf version Garbage Can Decision Making
by David Maister 1983
I have attended a number of partnership retreats held by professional service firms wherein the partners attempt to wrestle with some important choice, such as compensation system design, the issue of nonequity partners or forms of governance.
What is striking about these discussions is how disorganized they are and how far-reaching the debate usually becomes. Partners rarely speak to the topic at hand and frequently use the forum to air old grievances, wrestle with issues not on the agenda and waste time repeating points already made.
The conversation swings back and forth between issues of some importance and trivial matters of administrative detail. What’s more, there appears to be an inverse relationship between the importance of the topic and the amount of time devoted to it.
As a partner in one law firm pointed out to me, “These retreats are just like our normal partnership meetings. We spend hours arguing about the decoration of offices and then open a branch office because it seems ‘right.’ Everyone has to have their say, no one sticks to the point and we always end up talking about the same major issues, which never get resolved.”
This form of decision making is not restricted to professional service firms. It is, for example, also typical of many university faculty meetings I have attended. It occurs most frequently in organizations that attempt to run themselves on democratic, consensus-building principles. In a book entitled Leadership and Ambiguity, Cohen and March identify a type of organization they term as “organized anarchy.” Many of the ideas discussed here derive from their work. These organizations, they argue, have the following characteristics:
- An ambiguity of goals, with inconsistent and ill-defined preferences and a multiplicity of objectives: Because of this ambiguity, few issues can be resolved by appealing to unequivocal and mutually shared (or prioritized) goals.
- An unclear technology: Activities such as teaching, lawyering, consulting and other forms of professional service have ambiguous processes. While there are some regularities of procedure, the activities to be performed remain an “art,” learned by experience, trial and error, imitation and inventions born of necessity.
- Fluid participation: The members of the organization differ in the degree of time and effort they devote to its concerns. Sometimes a partner will focus solely on his or her own work, leaving organizational decision making to others; at other times, they may be exceedingly concerned with organizational matters and want to be involved. The degree of participation is not specified by an organizational chart, but by the issues addressed, the choices to be made, timing and temperament.
- Most issues most of the time have low significance for most of the people: The decisions made often secure only partial and erratic attention from the participants, and a major share of the attention devoted to a particular issue is tied less to its content than its symbolic significance and its impact on group esteem.
- As a consequence of this (previous point), there is high inertia: It takes a great deal of force and energy to get anything changed. There is a tendency to continue with the policies, procedures and patterns of the past.
- Finally, there is a weak information base: The data necessary for informed decision making is not commonly collected (perhaps because of the unclear technology) and not well disseminated (because of the fluidity of participation).
How do decisions get made in an organized anarchy? Michael Cohen, James March and Johan Olsen, in an article published in Administrative Science Quarterly, called it “garbage can” decision making.
In a garbage can process, there are a large number of unresolved issues or problems: “What do we reward around here? Who has status? What clients should we serve? How can we market ourselves better?”
There also exists a related but quite separate stream of choices to be made: “Should we open a branch office? Should we hire a new lateral partner?”
Third, there is a set of stock “solutions” constantly in the air: “Let’s get a computer! Let’s departmentalize! Let’s hire a more high-powered business manager!”
Finally, there are the participants who are involved to varying degrees with particular problems, solutions and choices.
All four of these elements—problems, choices, solutions and participants—are intertwined. There is not (in a garbage can environment) a “clean” decision-making process with circumscribed issues to be decided by a well-defined set of participants with clear, independent choices. Instead, it’s all jumbled up.
Cohen, March and Olsen discovered some major properties of such situations:
- Few problems or issues are dealt with by direct resolution. They are avoided, suppressed or disposed of as a secondary consequence of choice made on some other issue.
- The more issues there are on the table, the more likely individual problems will not get solved or will take longer to resolve.
- Although nominally considering different choices at different times, the decision makers keep encountering the same issues again and again.
- Given the complexities of interrelationships between decisions, issues, actions and people, it matters a good deal in what order issues and choices are confronted.
- Big choices are much less likely to resolve problems than unimportant choices because big choices are frequently dealt with by compromise.
In such an environment as this, how does one manage? Fortunately, Cohen and March do offer some practical advice. After a semiapology for the Machiavellian nature of their proposals, they offer the following guidelines for success in leadership in a garbage can environment:
(1) Spend time. A person, they assert, who is willing to spend time on decision-making activities by exploring issues, choices and solutions lays claim to more tolerant consideration of the problems he or she considers important and becomes a major information source in an information-poor world.
Influence is obtained by doing the work, serving on the committee, having more facts than others and freeing others from having to worry about management issues.
(2) Persist. In a garbage can world, few issues are resolved once and for all. If a proposal has been rejected today, it may be accepted tomorrow. Decisions are made as a result of a series of episodes involving different people in different settings, and they may be unmade or modified by subsequent episodes.
(3) Exchange status for substance. Cohen and March note that in the universities they studied (as in the professional service firms I have studied), “governance is simultaneously a system for making decisions and a system for certifying status. Participation is an end in itself, not just a means.” They suggest that some substantial elements of the governance of universities (and, I would argue, of professional service firms) can best be understood in the light of the hypothesis that “most people are most of the time less concerned with the contents of a decision than they are with eliciting an acknowledgment of their importance within the community.”
Accordingly, to get things done, exchange status for substance—a simple rule, but a powerful one.
(4) Facilitate opposition participation. Since in garbage can situations most people do not participate much or very frequently, and since such organizations are generally information weak, many participants’ expectations about what can and cannot be achieved, internally or externally, have a tendency to drift away from reality. Accordingly, “educating the opposition” is an important task. Involving them in decision-making committees will tend to correct this situation.
(5) Overload the system. Within an organized anarchy, Cohen and March note, it is a mistake to become absolutely committed to any one project since there are innumerable ways in which the processes will confound the cleverest behavior with respect to a single proposal, however imaginative or subjectively important.
What such organizations cannot do is cope with large numbers of projects. Someone with the habit of producing many proposals, without absolute commitment to any one, may lose any one of them but cannot be stopped on everything.
(6) Provide garbage cans. Since garbage can situations are ones where any choice or decision can provide the opportunity to raise any number of unresolved problems or issues, it is pointless to try to react by attempting to enforce rules of relevance, which are generally somewhat arbitrary.
Instead, Cohen and March suggest providing “garbage can topics” to draw the attention of those who wish to raise the extraneous issue and divert these issues away from the topic at hand. The first item on a meeting agenda is suggested as a perfect garbage can.
(7) Manage unobtrusively. As is obvious by now, direct confrontation is unlikely to succeed in a garbage can process. The energy and forces at work in the organization cannot be suppressed, but they can be redirected to different purposes. Minor actions can produce major effects, although these effects are not perceived by the organization when the “unimportant” choice to take a minor action is made.
Cohen and March point out that the major instruments of unobtrusive management are bureaucratic. For example, if you want to get an organization used to departments, don’t attempt to enforce a formal change. Instead, try changing the format of the accounting reports so that results for different groups are separated out.
In time, the groupings will come to be seen as natural. No war will be launched over the format of reports, but the attitudes of the players will be subtly changed. Once this is accomplished the burden of overcoming inertia will be placed on the opposition.
(8) Interpret history. Since most events in organized anarchies are complex and vague, they are subject to various interpretations. (Why did we lose that client? Why are profits down?) He who gets to interpret history gets to influence the future. According to Cohen and March, minutes should be written long enough after the event so as to legitimize the reality of forgetfulness.
If you try to write history (in other words, provide the meaning) too soon, opposition will arise, but after a suitable period of time, the “official” history is accepted. By being in command of the written history, the basis can be laid for subsequent independent action in the name of collective action.
The best professional service manager I ever met proceeded in the following way: He never brought an issue to an open forum until he had visited each partner on a one-on-one basis. He listened to their position, reasoned with them in private and framed his proposals for partnership meetings only when this process was completed.
Each person felt consulted and involved, as if he or she had participated in the decision-making process. But the chairman controlled the agenda at meetings and dealt with objections in private. Only when he had discovered what would “sell” did he raise a topic in open debate. Votes on his issues were always unanimous, and issues were not rehashed in public.
There is an art to building consensus. Garbage can situations need leaders who know how to make the organization work for them rather against them.
As I review the advice Cohen and March provide to leaders of garbage can organizations, I am struck by the analogies between the behavior they recommend and that exhibited by an astute politician.
It is almost as if they are summarizing the tactics a congressman uses in winning support for a bill on Capitol Hill. He or she too will win “ownership” of an issue by investing time on behalf of others, persisting, exchanging status for substance (that is, offering coauthorship of a bill), co-opting powerful opponents early on and managing unobtrusively until ready to “go public.”
Consensus-based “democracies” are political arenas, whether they are law firms or national assemblies, and the secrets to getting things accomplished are the same. Garbage can leaders in the law could usefully study political biographies (such as Robert Caro’s work on Lyndon Johnson), not only for amusement but for sound managerial advice.