Understanding Organizations as Learning Systems

With the decline of some well-established firms, the diminishing competitive power of many companies in a burgeoning world market, and the need for organizational renewal and transformation, interest in organizational learning has grown. Senior managers in many organizations are convinced of the importance of improving learning in their organizations. This growth in awareness has raised many unanswered questions: What is a learning organization? What determines the characteristics of a good learning organization (or are all learning organizations good by definition)? How can organizations improve their learning? In the literature in this area, authors have used different definitions or models of organizational learning or have not defined their terms.1 Executives have frequently greeted us with comments like these:

• “How would I know a learning organization if I stumbled over it?”
• “You academics have some great ideas, but what do I do with a mature, large organization on Monday morning?”
• “I’m not sure what a good learning organization is, but you should not study us because we are a bad learning organization.”

Our research is dedicated to helping organizations become better learning systems. We define organizational learning as the capacity or processes within an organization to maintain or improve performance based on experience. Learning is a systems-level phenomenon because it stays within the organization, even if individuals change. One of our assumptions is that organizations learn as they produce. Learning is as much a task as the production and delivery of goods and services. We do not imply that organizations should sacrifice the speed and quality of production in order to learn, but, rather, that production systems be viewed as learning systems. While companies do not usually regard learning as a function of production, our research on successful firms indicates that three learning-related factors are important for their success:

1. Well-developed core competencies that serve as launch points for new products and services. (Canon has made significant investments over time in developing knowledge in eight core competencies applied in the creation of more than thirty products.)
2. An attitude that supports continuous improvement in the business’s value-added chain. (Wal-Mart conducts ongoing experiments in its stores.)
3. The ability to fundamentally renew or revitalize. (Motorola has a long history of renewing itself through its products by periodically exiting old lines and entering new ones.)

These factors identify some of the qualities of an effective learning organization that diligently pursues a constantly enhanced knowledge base. This knowledge allows for the development of competencies and incremental or transformational change. In these instances, there is assimilation and utilization of knowledge and some kind of integrated learning system to support such “actionable learning.” Indeed, an organization’s ability to survive and grow is based on advantages that stem from core competencies that represent collective learning.2

As a corollary to this assumption, we assume that all organizations engage in some form of collective learning as part of their development.3 The creation of culture and the socialization of members in the culture rely on learning processes to ensure an institutionalized reality.4 In this sense, it may be redundant to talk of “learning organizations.” On the other hand, all learning is not the same; some learning is dysfunctional, and some insights or skills that might lead to useful new actions are often hard to attain. The current concern with the learning organization focuses on the gaps in organizational learning capacity and does not negate the usefulness of those learning processes that organizations may do well, even though they have a learning disability. Thus Argyris and Schön emphasize double-loop learning (generative) as an important, often missing, level of learning in contrast with single-loop learning (corrective), which they have found to be more common.5 Similarly, Senge makes a highly persuasive case for generative learning, “as contrasted with adaptive learning,” which he sees as more prevalent.6 The focus for these theorists is on the learning required to make transformational changes — changes in basic assumptions — that organizations need in today’s fast-moving, often chaotic environment. Their approach does not negate the value of everyday incremental “fixes”; it provides a more complete model for observing and developing organizational learning. After periods of significant discontinuous change, incremental, adaptive learning may be just the thing to help consolidate transformational or generative learning.

Another assumption we make is that the value chain of any organization is a domain of integrated learning. To think of the value chain as an integrated learning system is to think of the work in each major step, beginning with strategic decisions through to customer service, as a subsystem for learning experiments. Structures and processes to achieve outcomes can be seen simultaneously as operational tasks and learning exercises; this holds for discrete functions and for cross-functional activities, such as new product development. The organization encompasses each value-added stage as a step in doing business, not as a fixed classification scheme. Most organizations do not think this way, but it is useful for handling complexity. With this “chunking,” we are able to study learning better and to see how integration is achieved at the macro-organizational level. This viewpoint is consistent with a definition of organizations as complex arrangements of people in which learning takes place.

While we have not looked at organizations’ full value-added chains, we selected our research sites so that we could examine learning in different organizational subsets. In addition, we gathered data indicating preferences or biases in investments in learning at different points of the chain and to understand how learning builds, maintains, improves, or shifts core competencies. Do organizations see certain stages of the chain where significant investment is more desirable than at others?

Our last assumption is that the learning process has identifiable stages. Following Huber, whose comprehensive review of the literature presented four steps in an organizational learning process, we arrived at a three-stage model:

1. Knowledge acquisition — The development or creation of skills, insights, relationships.
2. Knowledge sharing — The dissemination of what has been learned.
3. Knowledge utilization — The integration of learning so it is broadly available and can be generalized to new situations.7

Most studies of organizational learning have been concerned with the acquisition of knowledge and, to a lesser extent, with the sharing or dissemination of the acquired knowledge (knowledge transfer). Less is known about the assimilation process, the stage in which knowledge becomes institutionally available, as opposed to being the property of select individuals or groups. Huber refers to the assimilation and utilization process as “organizational memory.” While this is an important aspect of knowledge utilization, it is limited and works better when discussing information, as distinct from knowledge. True knowledge is more than information; it includes the meaning or interpretation of the information, and a lot of intangibles such as the tacit knowledge of experienced people that is not well articulated but often determines collective organizational competence. Studies of organizational learning must be concerned with all three stages in the process.

Early in our research, it became clear that organizational learning does not always occur in the linear fashion implied by any stage model. Learning may take place in planned or informal, often unintended, ways. Moreover, knowledge and skill acquisition takes place in the sharing and utilization stages. It is not something that occurs simply by organizing an “acquisition effort.” With this in mind, we shifted our emphasis to look for a more fluid and chaotic learning environment, seeking less-defined, more subtle embodiments.

The first phase of our research was based on intensive field observations in four companies, Motorola Corporation, Mutual Investment Corporation (MIC), Electricité de France (EDF), and Fiat Auto Company.8 We wanted to have both service and manufacturing settings in U.S. and European environments. We chose two sites where we had access to very senior management and two where we were able to study lower levels. We selected Motorola as an example of a good learning organization; we were able to observe organizational learning during its fourteen-year quality improvement effort.

We did not attempt to study entire firms or to concentrate on any single work units in these four organizations. For example, at Motorola, we began by studying two senior management teams of twenty to twenty-five executives each from all parts of the corporation. Each team focuses on a critical issue defined by the CEO and COO, to whom the groups report. The teams’ structures were designed as executive education interventions and vehicles for “real-time” problem solving. Our objective was to see how these teams reflected and utilized organizational learning at Motorola.

From our interview data, we identified what organizational members claimed they had learned and why. We wrote case descriptions of the learning processes in their organizations, which we shared with the organizations to ensure their accuracy. Using a grounded analysis, we identified categories that reflected learning orientations and then constructed a two-part model of the critical factors that describe organizations as learning systems.9 We have since tested this model in data-gathering workshops with personnel from more than twenty Fortune “500” companies. Our testing led us to revise some of the model’s components, while retaining its overall framework.

Core Themes

Next we discuss the core themes that emerged from our research and provided a basis for our model.

All Organizations Are Learning Systems

All the sites we studied function as learning systems. All have formal and informal processes and structures for the acquisition, sharing, and utilization of knowledge and skills. Members communicated broadly and assimilated values, norms, procedures, and outcome data, starting with early socialization and continuing through group communications, both formal and informal. We talked with staff people in some firms who claimed that their companies were not good learning organizations, but, in each, we were able to identify one or more core competencies that could exist only if there were learning investments in those areas. Some type of structure or process would have to support the informed experience and formal educational interventions required for knowledge acquisition, sharing, and utilization. We found this in both our field sites and other firms. For example, one firm that considers itself to be a poor learning organization because of its difficulty in changing some dysfunction has a reputation in its industry for superior field marketing. It is clear that this group has well-developed recruiting, socialization, training and development, and rotating assignment policies that support its cadre of respected marketing people. Obviously, some learning has been assimilated at a fairly deep level.

Learning Conforms to Culture

The nature of learning and the way in which it occurs are determined by the organization’s culture or subcultures. For example, the entrepreneurial style of MIC’s investment funds group results in a learning approach in which information is made available to fund managers and analysts, but its use is at the managers’ discretion. In addition, there is a good deal of leeway in how fund managers make their investments; some are intuitive, some rely heavily on past performance, and a few use sophisticated computer programs. Thus the fund managers’ use or application of learning is largely informal, not dictated by formal, firmwide programs. Meanwhile, the culture of MIC’s marketing groups is more collaborative; learning is derived more from interaction within and between cross-functional work groups and from improved communication.

In contrast, there is no question that a great deal of organizational learning about quality has occurred at Motorola, but its emphasis on engineering and technical concerns resulted in an earlier, complete embrace of total quality by product manufacturing groups. In a culture that heavily rewards product group performance, total quality in products and processes that require integrated, intergroup action lags behind, particularly in the marketing of systems that cut across divisions.

Style Varies between Learning Systems

There are a variety of ways in which organizations create and maximize their learning. Basic assumptions about the culture lead to learning values and investments that produce a different learning style from a culture with another pattern of values and investments. These style variations are based on a series of learning orientations (dimensions of learning) that members of the organization may not see. We have identified seven learning orientations, which we see as bipolar variables.

For example, each of two distinct groups at both Motorola and MIC had different approaches to the way it accrued and utilized knowledge and skills. One Motorola group had great concern for specifying the metrics to define and measure the targeted learning. The other group was less concerned with very specific measures but, instead, stressed broad objectives. In the two groups at MIC, the methods for sharing and utilizing knowledge were very different; one was informal, and the other more formal and collaborative. From these variations, we concluded that the pattern of the learning orientations largely makes up an organizational learning system. The pattern may not tell us how well learning is promoted but tells a lot about what is learned and where it occurs.

Generic Processes Facilitate Learning

How well an organization maximizes learning within its chosen style does not occur haphazardly. Our data suggest that talking about “the learning organization” is partially effective; some policies, structures, and processes do seem to make a difference. The difference is in how easy or hard it is for useful learning to happen, and in how effective the organization is in “working its style.” By analyzing why learning took place in the companies we studied, we identified ten facilitating factors that induced or supported learning. While we did not observe all the factors at each site, we saw most of them and at other sites as well. Thus we view them as generic factors that any organization can benefit from, regardless of its learning style. For example, scanning, in which benchmarking plays an important role, was so central to learning at Motorola that it is now an integral, ongoing aspect of every important initiative in the company. Although MIC tends to create knowledge and skill internally, it maintains an ongoing vigilance toward its external environment. On the negative side, the absence of solid, ongoing external scanning in other organizations is an important factor in their economic difficulties.

A Model of Organizations as Learning Systems

Our two-part model describes organizations as learning systems (see Figure 1). First, learning orientations are the values and practices that reflect where learning takes place and the nature of what is learned. These orientations form a pattern that defines a given organization’s “learning style.” In this sense, they are descriptive factors that help us to understand without making value judgments. Second, facilitating factors are the structures and processes that affect how easy or hard it is for learning to occur and the amount of effective learning that takes place. These are standards based on best practice in dealing with generic issues. (See the sidebar for definitions of the learning orientations and facilitating factors we identified.)

Both parts of the model are required to understand an organization as a learning system; one without the other provides an incomplete picture. In addition, separating the parts enables organizations to see that they do indeed function as learning systems of some kind, and that their task is to understand better what they do well or poorly. (The idea of assessing what exists is more useful than the pejorative notion that there is only one good way to be a learning organization.) Finally, a refined, detailed list of factors related to organizational learning may help companies select areas for learning improvement that do not demand drastic culture change but, rather, can lead to incremental change over time.

Learning Orientations

In the next section, we expand on the definitions of the seven learning orientations and provide examples of each.

1. Knowledge Source.

To what extent does the organization develop new knowledge internally or seek inspiration in external ideas? This distinction is seen as the difference between innovation and adaptation — or imitation. In the United States, there is a tendency to value innovativeness more highly and look down on “copiers.” American critiques of Japanese businesses often mention that the Japanese are good imitators but not good innovators. In our opinion, both of these approaches have great merit as opposing styles rather than as normative or negative behaviors.

Although our data show a tendency in organizations to prefer one mode over the other, the distinction is not clear-cut. While MIC does scan its environment, it prefers to innovate in responding to customer needs and problems and has been a leader in developing new financial products and services. EDF modeled its nuclear power plants on U.S. technology. Motorola appears to be equally vigorous in innovation and in reflective imitation; it has been innovative in developing new products and adroit at adapting others’ processes, such as benchmarking and TQM procedures. Among firms not in this study, American Airlines, Wal-Mart, Merck, and Rubber-maid appear to be innovative in producing knowledge. And American Home Products is a good example of a highly successful, reflective imitator, as are AT&T’s Universal Credit Card, Tyco Toys (a Lego “copier”), and Lexus and Infiniti automobiles.

2. Product-Process Focus.

Does the organization prefer to accumulate knowledge about product and service outcomes or about the basic processes underlying various products? Many observers have stated that one reason Japanese companies are so competitive is that they make considerably more investments in process technologies in comparison to U.S. companies. The difference is between interest in “getting product out the door” and curiosity about the steps in the processes. All organizations give some attention to each side; the issue is to organize for learning in both domains.

Motorola makes learning investments on both sides. The executives we observed spent roughly equal amounts of time in collaborative learning about processes and outcomes. They paid less attention to “people processes” than to “hard” or technical processes, but many of them accepted the importance of process issues. MIC, EDF, and Fiat have traditionally focused almost exclusively on product issues but are now making greater learning investments in process issues.

3. Documentation Mode.

Do attitudes vary as to what constitutes knowledge and where knowledge resides? At one pole, knowledge is seen in personal terms, as something an individual possesses by virtue of education or experience. This kind of knowledge is lost when a long-time employee leaves an organization; processes and insights evaporate because they were not shared or made a part of collective memory. At the other pole, knowledge is defined in more objective, social terms, as being a consensually supported result of information processing. This attitude emphasizes organizational memory or a publicly documented body of knowledge.

MIC’s investment funds group focuses on a personal documentation style, eschewing policy statements and procedure manuals. In keeping with its entrepreneurial orientation, MIC makes it possible for individuals to learn a great deal, but there is little pressure to codify this. Though engaged in a business that values “hard data,” the group supports subjective, tacit knowledge in decision-making processes. And at Fiat’s Direzione Technica, where the individual has historically been the repository of knowledge, efforts are being made to establish a memoria technica, or engineering knowledge bank. Motorola shows evidence of both approaches but works hard to make knowledge explicit and broadly available.

4. Dissemination Mode.

Has the organization established an atmosphere in which learning evolves or in which a more structured, controlled approach induces learning? In the more structured approach, the company decides that valuable insights or methods should be shared and used by others across the organization. It uses written communication and formal educational methods or certifies learning through writing the procedures down. In the more informal approach, learning is spread through encounters between role models and gatekeepers who compellingly reinforce learning. In another approach, learning occurs when members of an occupational group or work team share their experiences in ongoing dialogue.10

MIC’s investment funds group clearly prefers informal dissemination in which learning develops and is shared in loosely organized interactions. This method occurs in other MIC areas, although the marketing groups are becoming more structured in their dissemination. Motorola supports both approaches, though it invests heavily in structured, firmwide programs when senior management wants a basic value or method institutionalized. It considered quality so critical that it now includes vendors and customers in its dissemination. (Recently, some vendors were told that they had to compete for the Malcolm Baldrige Quality Award in order to be on the company’s approved vendor list.) EDF prefers formal modes, emphasizing documented procedures that all share. Fiat’s Direzione Technica formally spreads knowledge by accumulating it in specialist departments and then disseminating it to cross-functional design teams.

5. Learning Focus.

Is learning concentrated on methods and tools to improve what is already being done or on testing the assumptions underlying what is being done? Argyris and Schön call the former “single-loop learning” and the latter “double-loop learning.”11 They have rightfully argued that organizational performance problems are more likely due to a lack of awareness and inability to articulate and check underlying assumptions than to a function of poor efficiency. In our opinion, these learning capabilities reinforce each other. Organizations may have a preference for one mode over the other, but a sound learning system can benefit from good work in both areas.

Our research sites displayed a range of behavior. EDF is primarily focused on incremental issues and does not question its basic assumptions. It prides itself on being the world’s major nuclear power utility and devotes significant resources to being the most efficient, safe operator through small improvements rather than transformations. Though similar, Fiat’s Direzione Technica is beginning to question assumptions about its new product development process. Since 1987, MIC has been in a transformational mode, particularly in the way that its marketing groups have focused on a questioning learning style. Motorola is fairly well balanced in its orientation; the founding family has historically accepted the concept of organizational renewal, which has led to far-reaching changes in the company’s product lines through the years and to an inquisitive style. On the other hand, its strong dedication to efficiency learning often precludes questioning basic assumptions.

6. Value-Chain Focus.

Which core competencies and learning investments does the organization value and support? By learning investments, we mean all allocations of personnel and money to develop knowledge and skill over time, including training and education, pilot projects, developmental assignments, available resources, and so on. If a particular organization is “engineering focused” or “marketing driven,” it is biased in favor of substantial learning investments in those areas. We divided the value chain into two categories: internally directed activities of a “design and make” nature, and those more externally focused of a “sell and deliver” nature. The former include R&D, engineering, and manufacturing. The latter are sales, distribution, and service activities. Although this does some disservice to the value chain concept, the breakdown easily accounts for our observations.

At MIC, the investment funds group focuses on the design and make side. While this is balanced by learning investments on the deliver side in the MIC marketing groups, there is a strong boundary between these groups, and the fund management side is regarded as the organization’s core. Motorola’s total quality effort clearly recognizes the importance of value-added at both sides, but “design and make” is significantly ahead of “deliver” in learning investments in quality. Fiat’s Direzione Technica is clearly oriented toward design and make, although its new system of simultaneous engineering is balancing its approach with increased sensitivity to the deliver side. EDF nuclear operations focuses squarely on efficient production. While not in our study, Digital Equipment Corporation’s learning investments traditionally were much more heavily focused on “design and make” than on “deliver.”

7. Skill Development Focus.

Does the organization develop both individual and group skills? We believe it helps to view this as a stylistic choice, as opposed to seeing it in normative terms. In this way, an organization can assess how it is doing and improve either one. It can also develop better ways of integrating individual learning programs with team needs by taking a harder look at the value of group development.

MIC designed the investment funds group to promote individual learning, which seems to fit with its culture and reward system. Heavy investment in team learning would probably improve its performance. On the other hand, MIC’s marketing groups, more supportive of collective learning, are now investing in team development as one way to improve its total effectiveness. Fiat’s Direzione Technica has been oriented toward more individual development, but, with its new reliance on cross-functional work teams, group development is increasingly more important. Recently, Motorola has become more team oriented and is making heavier investments in collaborative learning. It designed the two executive groups we observed to foster collective learning on two strategic issues affecting the entire company. EDF develops both individual and group skills, especially in control-room teams. All EDF employees follow individual training programs for certification or promotion. Control-room teams also learn, in groups, by using plant simulators. Some other firms that emphasize team learning are Federal Express, which invests heavily in teams for its quality effort, and Herman Miller, which stresses participative management and the Scanlon plan.

We view the seven learning orientations as a matrix. An organizational unit can be described by the pattern of its orientations in the matrix, which in turn provides a way to identify its learning style. Given the characteristics of the sites we studied and other sites we are familiar with, we believe it is possible to identify learning styles that represent a distinct pattern of orientations. Such styles may reflect the industry, size, or age of an organization, or the nature of its technology.

Facilitating Factors

The second part of our model is the facilitating factors that expedite learning. The ten factors are defined in the sidebar.

1. Scanning Imperative.

Does the organization understand or comprehend the environment in which it functions? In recent years, researchers have emphasized the importance of environmental scanning and agreed that many organizations were in trouble because of limited or poor scanning efforts. Thus many firms have increased their scanning capacity. Five years into Motorola’s quality program, a significant scanning effort showed it what others, particularly the Japanese, were doing. In reaction, Motorola substantially changed its approach and won the first Baldrige Award four years later. By contrast, the mainframe computer manufacturers (Cray, Unisys, IBM) and the U.S. auto companies in the 1970s failed to respond to developing changes that sound investigative work would have made painfully visible. Recent changes at Fiat result from a concerted scanning effort in which fifty senior managers visited the manufacturing facilities of world-class auto and other durable goods companies.

2. Performance Gap.

First, how do managers, familiar with looking at the differences between targeted outcomes and actual performance, analyze variances? When feedback shows a gap, particularly if it implies failure, their analysis often leads to experimenting and developing new insights and skills. One reason that well-established, long-successful organizations are often not good learning systems is that they experience lengthy periods in which feedback is almost entirely positive; the lack of disconfirming evidence is a barrier to learning.

Secondly, is there a potential new vision that is not simply a quantitative extension of the old or goes well beyond the performance level seen as achievable in the old vision? One or more firm members may visualize something not previously noted. Awareness of a performance gap is important because it often leads the organization to recognize that learning needs to occur or that something already known may not be working. Even if a group cannot articulate exactly what that need might be, its awareness of ignorance can motivate learning, as occurred at Motorola after its 1984 benchmarking. Currently, this “humility” is driving Fiat’s Direzione Technica to make a major study of what it needs to know.

In our findings, EDF provides perhaps the best instance of a performance gap leading to adaptive learning. Due to the nature of the nuclear power business, performance variations became the catalyst for a learning effort to again achieve the prescribed standard. We also found that future-oriented CEOs encouraged performance-gap considerations related to generative learning at Motorola and MIC (parent company).

3. Concern for Measurement.

Does the organization develop and use metrics that support learning? Are measures internally or externally focused, specific, and custom-built or standard measures? The importance of metrics in total quality programs has been well documented and is used in target-setting programs such as management by objectives.12 Our interest is in how the discourse about measurements, and the search for the most appropriate ones, is a critical aspect of learning, almost as much as learning that evolves from responding to the feedback that metrics provide.

Motorola executives believe that concern for measurement was one of the most critical reasons for their quality program’s success. At three or four critical junctures, reexamination of measurement issues helped propel a move to a new level of learning. They are applying this factor to new initiatives, a major concern of the executive groups we observed. At EDF, the value of metrics is clearly associated with the performance gap. Its nuclear power plants are authorized to operate at certain specifications that, if not met, may suggest or predict an unplanned event leading to shutdown. Each occasion becomes an opportunity for learning to take place.

4. Experimental Mind-set.

Does the organization emphasize experimentation on an ongoing basis? If learning comes through experience, it follows that the more one can plan guided experiences, the more one will learn. Until managers see organizing for production at any stage of the value chain as a learning experiment as well as a production activity, learning will come slowly. Managers need to learn to act like applied research scientists at the same time they deliver goods and services.13

We did not see significant evidence of experimental mind-sets at our research sites, with some notable exceptions at Motorola. At its paging products operation, we observed the current production line for one product, a blueprint and preparation for the new setup to replace the line, and a “white room” laboratory in which research is now underway for the line that will replace the one currently being installed. Motorola University constantly tries new learning approaches; the two executive groups we observed at Motorola were also part of an experiment in executive education.

We have seen evidence of experimental mind-sets in reports about other firms. For example, on any given day, Wal-Mart conducts about 250 tests in its stores, concentrated on sales promotion, display, and customer service. Although a traditional firm in many ways, 3M’s attitude toward new product development and operational unit size suggests a strong experimental mind-set.

5. Climate of Openness.

Are the boundaries around information flow permeable so people can make their own observations? Much informal learning is a function of daily, often unplanned interactions among people. In addition, the opportunity to meet with other groups and see higher levels of management in operation promotes learning.14 People need freedom to express their views through legitimate disagreement and debate. Another critical aspect is the extent to which errors are shared and not hidden.15

Perhaps the most dramatic example of openness in our findings is EDF, where abnormalities or deviations are publicly reported throughout the entire system of fifty-seven nuclear power plants. The company treats such incidents as researchable events to see if the problem exists anywhere else and follows up with a learning-driven investigation to eliminate it. It then disseminates this knowledge throughout the company. While this openness may be explained by the critical nature of problems in a nuclear power plant, we can only speculate as to what would be gained if any organization functioned as though a mistake is potentially disastrous and also an opportunity to learn.

6. Continuous Education.

Is there a commitment to lifelong education at all levels of the organization? This includes formal programs but goes well beyond that to more pervasive support of any kind of developmental experience. The mere presence of traditional training and development activities is not sufficient; it must be accompanied by a palpable sense that one is never finished learning and practicing (something akin to the Samurai tradition). The extent to which this commitment permeates the entire organization, and not just the training and development groups, is another indicator. In many ways, this factor is another way of expressing what Senge calls “personal mastery.”

MIC does an excellent job of exposing its young analysts to developmental experiences. Its chairman also seeks knowledge in many areas, not just direct financial matters. Motorola has a policy in which every employee has some educational experience every year; it has joint ventures with several community colleges around the country, joint programs with the state of Illinois for software competence development and training of school superintendents, and on-the-job and classroom experiences for managers up to the senior level. The company spends 3.6 percent of its revenues on education and plans to double this amount.16 Among firms not in our study, General Electric, Unilever, and Digital Equipment Corporation have valued continuous education at all levels for many years.

7. Operational Variety.

Is there more than one way to accomplish work goals? An organization that supports variation in strategy, policy, process, structure, and personnel is more adaptable when unforeseen problems arise. It provides more options and, perhaps even more important, allows for rich stimulation and interpretation for all its members. This factor helps enhance future learning in a way not possible with a singular approach.

We did not see a great deal of variety at our sites. EDF, perhaps due to the importance of total control over operations, shows little variation. Fiat’s Direzione Technica follows similar response routines, although the change to a new structure should lead to greater variation because of its independent design teams. An exception is MIC investment funds group, where we identified at least three different methods that fund managers used in making investment decisions. Senior management, although a bit skeptical about one of the methods, seemed willing to support all three as legitimate approaches.

8. Multiple Advocates.

Along with involved leadership, is there more than one “champion” who sets the stage for learning? This is particularly necessary in learning that is related to changing a basic value or a long-cherished method. The greater the number of advocates who promote a new idea, the more rapidly and extensively the learning will take place. Moreover, in an effective system, any member should be able to act as an awareness-enhancing agent or an advocate for new competence development. In this way, both top-down and bottom-up initiatives are possible.

One of the authors participated in two significant change efforts that failed, largely because there was only one champion in each case. One highly frustrated CEO said, “It doesn’t do me or the company any good if I’m the only champion of this new way of doing business.” At Motorola, we found that a major factor in the quality effort’s success was the early identification, empowerment, and encouragement of a significant number of advocates. In a current initiative we observed, Motorola is enlisting a minimum of 300 champions in strategic parts of the company. Digital Equipment Corporation has had learning initiators throughout the company since its early days. Digital’s problem has been in assimilating and integrating the lessons of its myriad educational and experimental efforts, rather than in creating an environment that enables broad-scale initiation. MIC’s investment funds group encourages many individuals to initiate their own learning but not to proselytize.

9. Involved Leadership.

Is leadership at every organizational level engaged in hands-on implementation of the vision? This includes eliminating management layers, being visible in the bowels of the organization, and being an active, early participant in any learning effort. Only through direct involvement that reflects coordination, vision, and integration can leaders obtain important data and provide powerful role models.

At Motorola, CEO Bob Galvin not only drove the quality vision, he was a student in the first seminars on quality and made it the first item on the agenda at monthly meetings with his division executives. Much-admired Wal-Mart CEO David Glass spends two or three days each week at stores and warehouses; employees can call him at home and are often transferred to his hotel when he is in the field. Mike Walsh of Tenneco (formerly of Union Pacific Railroad) meets with groups of employees at all levels in what Tom Peters calls “conversation.”17

10. Systems Perspective.

Do the key actors think broadly about the interdependency of organizational variables? This involves the degree to which managers can look at their internal systems as a source of their difficulties, as opposed to blaming external factors. Research in the field of systems dynamics has demonstrated how managers elicit unintended consequences by taking action in one area without seeing its dynamic relationship to its effects.18

Despite its importance, this factor was relatively lacking at our research sites. MIC and Motorola are structured so that there are strong boundaries between groups and functions. Both have changed their perspectives recently, MIC as a consequence of unexpected internal problems related to the October 1987 stock market crash, and Motorola after experiencing difficulties in selling large-scale systems (as opposed to discrete products). In a 1992 survey of 3,000 Motorola employees that asked them to evaluate their unit based on Senge’s five factors, they rated systems thinking the lowest and the one that required the most work to improve organizational learning. In contrast, Fiat’s Direzione Technica took a systems approach to understanding the consequences of its structure on new product development. As a result, it changed the structure to establish mechanisms for simultaneous engineering. To reduce the new products’ time to market, functions now work in parallel rather than sequentially.

General Directions for Enhancing Learning

We have divided the seven learning orientations and ten facilitating factors into three stages — knowledge acquisition, dissemination, and utilization. Figure 2 shows the orientations and factors within this framework. Within our two-part model, there are two general directions for enhancing learning in an organizational unit. One is to embrace the existing style and improve its effectiveness. This strategy develops a fundamental part of the culture to its fullest extent. For example, a firm that is a reflective imitator more than an innovator could adopt this strategy with heightened awareness of its value. A company that has benefited from heavy learning investments on the “make” side of the value chain would see the value of those investments and decide to build further on them. This approach builds on the notion that full acceptance of what has been accomplished is validating and energizing for those involved. It is similar to the appreciative inquiry numerous organizational change consultants advocate.19 The task is to select two or three facilitating factors to improve on.

The second direction is to change learning orientations. The organizational group would make more learning investments at a different part of the value chain, try to be an innovator if it is now more of an imitator, and so on. These are different changes from those involved in enhancing the facilitative factors, and the tactics will be different. Some changes will be seen as an attack on the organization’s basic values, and it may be possible to avoid this by moving toward balance between the two poles, so members of the organization will support the existing style and advocate the “new look” as a supplementary measure.

Supporting the Learning Orientations

In the second phase of our research, in which we worked closely with personnel from more than thirty Fortune “500” companies to identify their learning orientations, we validated our notion that organizations learn in varied ways. The singular “learning organization” should be a pluralistic model.

Looking at “what is” in a descriptive rather than normative way has another advantage in that you see better what you are not by examining better what you are. In the gestalt approach to dealing with resistance to organizational change, it has been well documented that change comes more readily if the targets of change first become more aware of and more accepting of their resistance.20 In other words, it is important to gain full knowledge and appreciation of your organizational assumptions about learning whether you want to build on them or alter them.

This model may also be used to identify the complementarity of styles between coordinating organizations and to recognize that circumstances may dictate conditions and orientations in particular settings. For example, EDF’s nuclear operations are constrained from transforming real-time operations due to the potentially dire consequences (e.g., the Chernobyl disaster) of operating under novel assumptions. However, at EDF, testing system assumptions is characteristic of its R&D division, which uses new technologies in the design of new plants. Thus changing one’s style needs to be considered from a systems perspective; it may also be associated with the stage of organizational development.21

Strategies for Improving Organizational Learning Capability

When starting to improve its learning capabilities, an organization may decide to focus on any stage of the learning cycle — knowledge acquisition, dissemination, or utilization. While it may be possible or necessary to look at all three phases simultaneously, focusing on a single area is more manageable. The next task is to select an option for focus:

1. Improve on learning orientations. There are two reasons for selecting this option. First, the organization may decide to shift its position on one or more learning orientations. Second, the current pattern of learning orientations has resulted in identifiable strong competencies, so improving or expanding them may be the best way to enhance the unit’s learning capabilities. This focus assumes that facilitating factors meet an acceptable standard and that more can be accomplished by adding to the strong base established by the learning orientations.

2. Improve on facilitating factors. In this option, the organization accepts its pattern of learning orientations as adequate or appropriate to its culture and decides that improving the systems and structures of the facilitating factors is the most useful course. This option assumes that maximizing the facilitating factors would add more to the organization’s learning capabilities than enhancing or changing the current learning orientations.
3. Change both learning orientations and facilitating factors. An organization should select this option when it sees the other variables as inadequate. This option assumes that large-scale change is necessary and that changing one group of variables without changing the other will be only partially successful.

Each organizational unit or firm must make the decision to pursue one strategy or another for itself. While there are no rules for making this decision, the three options are incrementally more difficult to implement (i.e., one is the easiest to implement; three is the hardest). From the first to the third options, the resistance to change within the organization increases significantly. It is one thing to develop a plan for improving what is already done reasonably well; it is another to engage in nothing less than near-total transformation. It is one thing to stay within accepted, assimilated paradigms; it is another to replace institutionalized models.

Whatever the organization’s choice, we offer three guidelines for developing and implementing a chosen strategy:

1. Before deciding to become something new, study and evaluate what you are now. Without full awareness and appreciation of current assumptions about management, organization, and learning, it is not possible to grasp what is being done well and what might be improved or changed.
2. Though the systemic issues and relationships in organizational life require that change be approached from multiple directions and at several points, organizations can change in major ways if people experience success with more modest, focused, and specific changes. As with many skills, there is a learning curve for the skill of managing and surviving transitions. Large-scale change requires that many initiatives be put into place in a carefully designed, integrated sequence.
3. Organizations must consider cultural factors in choosing and implementing any strategy, particularly when considering how it does specific things. For example, in a highly individualistic society like the United States or the United Kingdom, skill development focuses on individual skills; in comparison, more communitarian societies such as Japan or Korea have traditionally focused on group skill development. Moving from one pole to the other is a major cultural change; to simply improve on the existing orientation is much easier.

To help managers better understand the learning capabilities in their own organizations, we have developed and are testing an “organizational learning inventory.” This diagnostic tool will enable an organization’s members to produce a learning profile based on our model. The profile can guide managers to their choices for improving learning capability. Through further research, we intend to show how learning profiles vary within and across different companies and industries.