Learning Consortia: How to Create Parallel Learning Systems for Organization Sets
Edgar H. Schein
Nov. 1994, Revised Aug. 1995
Table of Contents
AbstractOrganizational Learning as a Change Process
Proposition 1
Proposition 2
Proposition 3
Proposition 4
Proposition 5
Proposition 6
The Leaders Group--Diffusion Upward
The Liaison Officer's Group--Partial Lateral Diffusion
The Semi-Annual Meeting--General Diffusion
Conclusions and Implications
References
The purpose of this working paper is to develop a model of how organizational learning can be diffused and sustained from an initial successful learning project or lab to the rest of the organization. The main point is to demonstrate that sets of organizations who are involved in learning can play an important role in helping each other.
We have many examples of getting learning started in pilot projects of various sorts only to see them atrophy or be rejected by other parts of the organization. In this paper I will not address how to get a learning system started. The focus will be on how to build systems and structures that will sustain such systems once they are successful and will diffuse the learning process to other parts of the organization.
I use the term "organization" in the broadest sense to cover the public and private sector and communities that have some sense of their own identity and defined social borders. The term "organization set" (Evan, 1967) refers to a cluster or group of organizations that are connected to each other through some common activity or through a need to interact. For purposes of this paper an organization set is defined as a set of companies all of whom are trying to become "learning organizations" and are using a common external organization, MIT's Organization Learning Center (OLC), as their point of reference and source of help.
Organizational Learning as a Change Process
I will assume that any form of organizational learning is a change
process of some sort. Whether we are talking of adaptive responses to maintain
equilibrium, growth, innovation, generative learning, or major transformation,
some measurable change in the organization system is always involved. At the
most general level, therefore, we must have a theory of change that
encompasses all of these forms.
I will assume that any change involves the creation of something new. If the organization is merely trying to survive by building the capacity to do what it has always done in the face of changing environmental presses, the building of that capacity, the development of adaptive responses is, by definition, a creative act. Any theory of change must, therefore, contain a theory of what makes it possible for a system to create something new--a new process, a new capacity to respond, a new product, or eventually a whole new structure.
What then are the most basic conditions for any learning or creativity to occur? I have previously argued that the system must first be unfrozen by 1 ) experiencing some disconfirmation which 2) produces anxiety or guilt under 3) conditions of enough psychological safety to overcome the defenses produced by the anxiety associated with change. If these three conditions are met, members of the organization are open to taking in new information which allows them to reframe or cognitively re-define some aspect of their psychological field. Such re-definition in turn leads to new behavior designed to reduce the disconfirmation (Schein, 1990, 1992, 1993).
This model accounts for creativity at the individual level and describes a generic change process. If we seek to understand organizational change and learning, we have to have a further model to reflect the empirical reality that only some people in the organization will experience disconfirmation and launch into a learning process at any given time. If one studies cases of successful organizational transformation, i.e. major system wide learning, one finds that in every case the organizational learning first began in an individual, then diffused to a group, and only gradually diffused from the group into the main body of the organization.
Observing this process carefully from a clinical perspective reveals that the group which initially learns becomes a "parallel system" or what Zand has called a "collateral organization" for a period of time (Bushe & Shani, 1991; Zand, 1974). In other words, such a parallel system becomes for the organization the functional equivalent of the R & D organization or a pilot plant in the product development arena. For organizational learning to be successful, therefore, requires us to understand how parallel systems are created, sustained, and eventually re-absorbed into the main body of the organization.
Proposition 1: In order to learn something new an organization
must have enough "slack" to create a parallel system.
If everyone is fully engaged in the basic primary task of the system, they
will continue to act in ways that may become maladaptive if the environment
changes. Disconfirming information must be noticed and the organization must
have the capacity to act on it, i.e. overcome the anxiety associated with learning.
Most organizations recognize the need for such slack in the product
development area and, therefore, create R & D departments and/or tolerate the
presence of some percentage of employees whose immediate function is not
completely accounted for or necessary (i.e. some corporate staffs, internal
consulting groups, etc.). The concept of lean and mean can be dangerous if it
implies the removal of all slack.
At the individual or group level we articulate this same principle by noting the importance of "reflection" and "practice fields" as essential components of a learning process (Senge, 1990). When we are in a reflective mode, we are creating within ourselves a "parallel system" in which we can experiment, contemplate, and reframe our thoughts and, thereby allow new thoughts to arise. When we "practice" we are creating a "parallel system" in which to test new behavior, allow ourselves to make errors, repeat desired responses until they become habitual. In either case we have to make time for these activities and create psychological space for them to occur.
Proposition 2. New learning (creativity) initially occurs through
individuals or groups who have enough slack to create a practice
field and experiment with new responses to the new environmental
conditions. If those responses are successful, the new unit will be
operating for a time by principles. rules. or norms other than those
of the rest of the organization.
Such units are called pilot plants, experimental units, or in other ways
named as being different to legitimate the different things they are doing. For
our purposes all such units can be thought of as "parallel systems." As Von
Hippel (1988) has shown in the product development arena, such parallel
systems often involve what he calls "lead users" who are individuals or groups
in the customer organization. As such customer groups involve the
manufacturer in the new product or process ideas they collectively become a
parallel system to the R & D function itself and have the problem of diffusing
their new learning into first the R & D function and then the manufacturing
function.
The inability of an organization to absorb the learning of the parallel system is illustrated very well in the innovation literature where examples abound such as Xerox Parc's development of innovative user interfaces that were only later adopted by Apple. In other words, precisely because the parallel system develops norms and cognitive models that differ from those of the main organization, they become a potential threat and source of anxiety. The organization often throws its immune system into gear and attempts to isolate or destroy the threatening parallel system. A model of organizational learning must, therefore, develop concepts that deal with the conditions under which parallel systems survive and are able to diffuse their learning back into the organization.
Proposition 3. Unless the parallel system is positioned near the
top of the organization or reaches a critical mass. it will continue to
learn for a time. but its learning will not automatically diffuse back
into the main organization. In fact. the organization's "immune
system" will isolate and possibly destroy it because the new
responses threaten the traditions and culture of the main part of the
organization.
One condition for the survival of the parallel system and diffusion of the
learning process is that the parallel system has the power to impose its new
learning on the organization. Typically this condition occurs when a CEO is the
initial learner, creates a group around him or herself to function as a parallel
system and then patiently and persistently creates the conditions necessary for
the rest of the organization to learn. This process requires patience and
persistence even from a high power base because the change process at each
level is the same. People have to experience disconfirmation, anxiety, and
psychological safety before they will become reflective enough to learn
something new. Examples of this kind of organizational change and learning
would probably be found in Herman Miller, Harley-Davidson, Philips U.S.,
General Electric, and other companies where the CEO is clearly part of the
parallel system. Ciba-Geigy's turnaround from 1979-1982 illustrates this
process in detail (Schein, 1992, Ch. 17).
If the parallel system is somewhere in the middle of the organization as in the case of the Ford Product Development Team, the Federal Express sales organization, the various Leading Learning Communities, the DEC logistics group, and so on, one can immediately observe the immune system's defensive dynamic in operation. One obvious prescription for such parallel systems is to figure out how to involve the power centers of the organization, but this takes time and strength on the part of the parallel system. Often the CEO or other senior executives are the ones most threatened by the new behaviors of the parallel system because it reveals how their own current behavior may be dysfunctional. How to disconfirm the CEO and provide psychological safety upwards is as yet little understood.
We have observed in the evolution of the OLC a mechanism that may help in this process, either by providing time and/or strength, or by providing new tools and processes for involving the power centers. However, in the meantime it becomes crucial to determine how parallel systems can survive and thrive without explicit support from the top of the organization.
Proposition 4. If the parallel system is to survive long enough to
influence the main organization. it must have support from some
other systems either inside or outside the organization.
Ideally, if the parallel system is learning something new and valuable,
support should come from other parts of the organization but, because of the
immune reaction, such support is rarely forthcoming from inside. How then can
the support be provided from the outside? One obvious mechanism that
organizational parallel systems use is consultants. Not only are the consultants
initially useful in reframing the problem and developing new behavioral
responses, but their presence can stabilize and legitimate the parallel system
and they can become a bridge to other power centers.
Another mechanism, illustrated by the OLC, is for the organization to join a consortium devoted to learning and innovation that provides insight, stimulation, and support in two ways: 1 ) the involvement with researchers, trainers, and coaches; and 2) the involvement with other learners.
Proposition 5. Critical support for the parallel structure must come
from outside the organization. either from consultants. research
centers or other parallel organizations going through the same
learning process.
If a number of organizations ally themselves with a research center
devoted to creating organizational learning, the various parallel systems that
come into being in this organization set can become a critical source of learning
and support for each other. For this to work, the organizations must begin to
see themselves as an organizational set that can be mutually supportive, must
build a parallel system for the entire organizational set and must learn how to
use that parallel system to enhance the learning in each organization. In this
second order parallel system, stage 2 learning can occur as described in
Proposition 6.
Proposition 6: For new concepts and Skills (know-how) to be fully
utilized. a two-stage learning process Is necessary where the
second stage is based on learners in the parallel systems learning
from each other and from their counterparts in the organization set.
Most radically new ideas and the skills sets or know-how that are needed
to implement them are too complex to be acquired by practitioners from
academics or consultants. The learning process begins in the parallel system
through the input and training of consultants, but for full utilization a second
stage learning process is needed where the practitioners learn from others in
the learning set. It has to be others in the learning set to overcome the
automatic restrictions that the members of a given parallel system impose on
themselves through their embeddedness in their own organizational culture.
The final coaching has to be done by practitioner peers because they are the
only ones who understand the opportunities and constraints afforded by the
culture of the occupational community in which they operate, i.e. in this case the
community of "managers."
The MIT Organization Learning Center has created both an organization set and a set of mutually supporting parallel structures for the entire set. These are the Leaders Group, the Liaison Officers Group, and the Semi-annual Meetings. In the remainder of this analysis I would like to spell out how these groups and meetings serve the functions of support systems and, thereby, increase the probability of diffusion back into the main part of the organization.
The Leaders Group--Diffusion Upward
The Leaders group consists of a carefully selected group of 5 CEO's who
meet for two full days quarterly with senior staff of the OLC around agendas
decided by the group from quarter to quarter. The intent is to combine periods
of generative dialogue (Isaacs, 1994) with focused learning exercises and
topical discussions with selected visitors such as Karl Henrik Robert, the
Swedish environmentalist. For the group itself it is an opportunity to reflect,
reframe, and re-generate their own commitment to learning. As a support
system for the parallel systems in each of their companies the group functions to
legitimize the learning projects and to insure that they survive and diffuse.
One thing these CEO's share is the recognition that formal power is not enough to start a learning process in their own organizations. Each of them has voiced great frustration at one time or another about the slowness of change in their companies and the difficulties of bringing other layers of the organization on board. From a theoretical point of view it is clear why this is so. The CEO has to produce the same conditions for change that launched him or her, namely to provide for disconfirmation, the creation of anxiety or guilt, and the provision of psychological safety. The easiest thing to do from a position of power is to disconfirm and create anxiety or guilt. The most difficult step is to provide enough psychological safety to enable the immediate subordinates and the layers below them to begin a learning process themselves. In addition the CEO has to legitimize a certain amount of slack in the organization so that people can take time out to be reflective and reframe issues.
The Leaders Group is not only supportive to the parallel systems in their own organizations, but they provide a role model for other CEO's in the learning set. CEO's who might be very threatened by disconfirmation from within their organization might be able to gain their psychological safety by checking out what is going on in other learning organizations where the CEO's are centrally involved. This group then becomes a critical force to support the learning of the entire organization set even though it may represent only five companies.
The Liaison Officer's Group--Partial Lateral Diffusion
The Liaison Officers Group (LOG) was originally created 1 ) to be an
ongoing interface with the OLC, 2) to provide a structure for companies to get
acquainted with each others' projects, and 3) hopefully to help each other to
learn by collaborating on projects, by providing training and coaching to each
other and to provide a broad support base for the whole enterprise. Senge's
original model calls for lateral diffusion by creating partnerships not only with
MIT but among the sponsoring companies so that eventually they can carry the
learning projects by themselves with each others' help and without the explicit
intervention of MIT staff.
We found that by sharing problems and issues with the LOG (e.g. how much to charge a sponsoring company; what deliverables had to be there to justify the fee, how often to meet, how to design the meetings, etc.) we enlisted a level of commitment and involvement on the part of its members that made the learning process itself richer. LOG members all had gone through the basic competency course and were individually on a learning track. They were therefore capable of contributing significantly to questions of how to design meetings, learning projects, and learning tools. During the meetings they become well enough acquainted with each other to begin to call on each other for help and support. It is this last function that is crucial to sustaining the parallel systems. As projects become successful and run into difficulty in their home organizations, they can share their dilemmas with members of the LOG and get genuine help and support. Not the least of such support is the discovery that successful projects run into similar kinds of defensiveness in widely different kinds of organizations. Members of the LOG can then reflect on and invent collective solutions based on a broader understanding of organizational dynamics.
The Semi-Annual Meeting--General Diffusion
The semi-annual meeting is attended by a much larger audience.
Sponsoring companies are allowed to send large teams and visitors from
potential sponsors or other learning centers world wide are often invited.
Besides hearing reports of the status of the various learning projects they attend
focused sessions on such topics as Dialogue or Organizational Culture. The
meeting provides ample opportunities for networking and following up on
interesting ideas that may have come up in an earlier session. The insights
gained by members of learning projects that are far enough along to be validly
labelled parallel systems are widely shared in this context, providing not only a
stimulus to others in their own organization but to other organizations. At the
same time, some of the problems and hurdles that are reported may have been
creatively dealt with in other organizations and those ideas can then filter back
to the projects.
Conclusions and Implications
I have argued that generative learning takes place initially through the
creation of parallel systems and that those parallel systems when they are parts
of larger organizations need to be supported long enough to begin to diffuse
their insights into the main body of the organization. I have also argued that
such support is usually difficult to get inside the organization unless the parallel
system includes the power centers, but that significant support can be provided
by other parallel systems from other organizations in the set that is learning
together. The creation of an organizational set that is jointly committed to
organizational learning and the creation of a parallel system to that organization
set therefore becomes necessary as a forum where project members can share
their dilemmas and difficulties and learn from each others' experiences. This
sharing process leads to Stage 2 learning where the concepts and skills that
are originally transmitted from academics and consultants now get embedded
in the practitioner's repertoire through the coaching from peers who are in their
own learning process but who come from other organizations and, thereby, can
help learners overcome the blind spots of their own organizational cultures.
The Leaders Group, the Liaison Officers Group and the Semi-Annual Meeting of the OLC function as such parallel systems to the organization set and should therefore not only be stabilized and supported by the OLC but should be studied in their own right as an essential mechanism of organizational learning. It may well be that the most important empirical outcome of the OLC program will be data on how to build an effective set of learning organizations and how to help them to create their own parallel systems to sustain and diffuse their learning. The actual tools used in the learning process are important, but the structuring of the parallel systems may, in the end, be the crucial determinant of whether organizations as a whole can learn generatively.
References
Bushe, G. R. & Shani, A. B. Parallel Learning Structures. Reading, MA.:
Addison Wesley, 1991.
Evan, W. M. "The Organization Set--Toward a Theory of Interorganizational Relations." In J. D. Thompson (Ed.) Approaches to Organizational Design. Univ. of Pittsburgh Press, 1967.
Schein, E. H. Innovative Cultures and Adaptive Organizations. Sri Lanka Journal of Development Administration. 1990, 7, No. 2, 9-39.
Schein, E. H. Organizational Culture and Leadership (2d Ed.). San Francisco: Jossey-Bass, 1992.
Schein, E. H. How Can Organizations Learn Faster? Sloan Management Review, 1993, 34, 85-92.
Senge, P. M. The Fifth Discipline. N.Y.: Doubleday, 1990.
von Hippel, E. The Sources of Innovation. N. Y.: Oxford Univ. Press, 1988.
Zand, D. E. Collateral Organization: A New Change Strategy. Journal of Aeelied Behavioral Science, 1974,10, 63-89.