12

Demystifying the Learning Organization II

The Case of Chaparral Steel

In the mid-1990s, Chaparral Steel, a “minimill” in Midlothian, Texas, received a great deal of attention for its success in competing with low-cost non-U.S. producers. In its first 8 years in operation, it grew from 250,000 tons of produced steel at start-up time to the 10th largest steel producer in the United States in 1998. Breaking all previous efficiency records in the process, Chaparral received the following lauda-tory analysis in Fortune magazine:

Chaparral is remarkable because, like a sculling crew that pulls in flawless synchronism, it has all the basic elements of good management—customer service, empowerment, quality, training, and more—working in concert. As a result, it produces steel with a record low 1.6 hours of labor per ton, vs. 2.4 hours for other mini-mills and 4.9 hours for integrated producers. (Dumaine, 1992, p. 88)

According to Fortune, Chaparral Steel owed its success to observing general “principles of good management.”

Dorothy Leonard-Barton (1992, 1995), who studied Chaparral Steel extensively, attributed its success to organizational learning. The former CEO of Chaparral Steel Gordon Forward himself stated that “we are a learning organization” whose core competence is the “rapid realization of new technology into products” (Leonard-Barton, 1992, p. 23). Here are some examples of organizational learning at Chaparral:

Managers assume that the performance of any purchased equipment can be improved…. Some improvements are novel enough to be patented. The rolling mill equipment its vendor believed limited to 8-inch slabs is now turning out 14-inch slabs, and the vendor has tried to buy back the design. (Leonard-Barton, 1995, p. 11)

When cooling hoses burst during the first few weeks of starting up the new…project, a group of operators, a welder, a foreman, and a buyer spontaneously gathered to discuss the problem, then scattered to seek solutions…. In this case (quoting a senior operator)… “Everyone telephoned some person they thought might know how to fix the problem—vendors, experts—and within three to four hours, we were getting calls back… and we worked the problem out.” (Leonard-Barton, 1995, p. 10)

These examples reflect the dual mission that CEO Gordon Forward set for all employees: fulfilling their individual functions and continually improving production. In doing so, they were expected to solve production problems, improve production processes and equipment, import new knowledge into the organization, and share their knowledge with others (Leonard-Barton, 1992).

Exemplars of organizational learning like Chaparral are a double-edged sword. On the one hand, they inspire and invite learning through imitation. “By benchmarking against such companies,” wrote Leonard-Barton (1992), “managers can derive principles to incorporate into their own particular visions” (p. 36). These exemplars may even tempt managers to take bold steps to transform their own companies into learning organizations. On the other hand, these exemplars may actually contribute to the mystification of organizational learning by feeding the gap between visionaries and skeptics (see  Chapter 1 ). No matter how inspiring the ultimate vision, studies of these organizations rarely offer a well-marked path for getting there. Furthermore, the reality of promoting organizational learning is demanding, fraught with setbacks, and simply much more mundane than the heroic exemplars. This reality may increase cynicism and the belief that organizational learning is nothing more than another fad or gimmick.

In this chapter, we shall argue that managers can learn from Chaparral Steel's example. In order to do so, however, they need a framework for systematic comparison that enables them to identify both the similarities and the key differences between their organizations and Chaparral. The multi-facet model presented in this book offers such a framework. Our reanalysis will build on descriptions of Chaparral Steel from previous studies and interviews with Gordon Forward rather than on original research (Kantrow, 1986; Leonard-Barton, 1992, 1995; Luthans, 1991; Preuss, 1998). It will show how Chaparral implemented a unique configuration of OLMs, especially the online/internal organizational learning mechanisms (OLMs). It will also identify features of Chaparral's learning culture, managerial policies, and leadership that make these OLMs effective. Finally, it will show how the particular design of learning at Chaparral was shaped, at least in part, by a specific set of contextual conditions. Such an analysis provides a more finely grained basis for comparison with other organizations and for understanding what principles can be adopted and what cannot.

 ORGANIZATIONAL LEARNING MECHANISMS AT CHAPARRAL

The scope and intensity of organizational learning at Chaparral can be attributed to the extensive use of OLMs, especially what we referred to in  Chapter 2  as “online/internal agent OLMs.” These mechanisms fuse work and learning together: The same people who do the work do the learning, and they do both at the same time and place.

[After the introduction of a new product and production process,] the pulpit controls operator is carefully checking the timing on the line with a stopwatch…. He wants to achieve split-second timing. Asked who suggested he perform this function (which is often given to a process engineer elsewhere), he is surprised at the question: “No one.” He considers it obvious that improvement is always part of his job. (Leonard-Barton, 1992, pp. 26–27)

Online/internal agent OLMs enable employees to produce changes in organizational routines, standard operating procedures, or norms of behavior while carrying out their specific functions.

The heart of online/internal agent learning mechanisms at Chaparral was constant experimentation carried out by everyone from management to the operators. There was no research and development department separate from production because it was considered to be a part of everyone's job (Leonard-Barton, 1992, p. 29). Employees at every level were expected to come up with ideas for improvements and, if possible, try out their own ideas on the spot.

Even operators were given opportunities to acquire and use the most advanced scientific knowledge, even if it meant travel (Kantrow, 1986). There were, of course, engineers and technicians working at Chaparral, but even they had line functions and worked closely with supervisors and operators so that experimentation was a joint project. Forward believed that putting operators, engineers, and maintenance workers in constant contact was the best and quickest means of testing out new ideas (Kantrow, 1986).

These experiments took place at the same time and at the same place as the production process. Innovation and experimentation were simply a part of everyday work routines. In the words of Gordon Forward:

The lab is the plant…. Of course, we don't give the whole plant over to laboratory work, but the whole plant really is a laboratory—even though it's one of the most productive steel mills in the world. We don't stop operations to try crazy things, but we do try to do our research and development right on the factory floor. (Kantrow, 1986, p. 99)

The fact that experimentation and process improvements took place online meant that management had to tolerate disruptions to production. The great benefit was that the results could be implemented immediately without further engineering or testing because the experimental environment and final production environment were the same (Leonard-Barton, 1992).

Chaparral provided a rare and extremely valuable opportunity to observe internal/online mechanisms. Off-line OLMs are relatively easy to comprehend, identify, and institutionalize because they are based on a clear distinction between working and learning. Online/internal agent OLMs, on the other hand, are difficult to observe because they are tightly integrated into the flow of work itself. In many organizations, they are not fixed or formal entities but rather temporary structures that take shape as needed through informal interactions and interpersonal networks. Their informality makes them difficult to create and manage.

At Chaparral, online/internal agent OLMs did not evolve in some naturalistic way but rather from conscious management choice and design. The impetus for these design features was CEO Forward's own learning from experience in a large, bureaucratic organization (Luthans, 1991, p. 67). Forward jokingly compared big, separate research facilities to cemeteries because “good ideas are dying there all the time” (Kantrow, 1986, p. 2). He believed that no matter how much money was invested in R & D (an off-line/external agent OLM), this knowledge did not flow to the people who were supposed to use it. On the other hand, line workers and managers who came up with good ideas for change were unable to get this knowledge adopted and implemented by the organization (Luthans, 1991).

Because Chaparral Steel was created as a completely new, self-contained plant, it offered an opportunity to design an alternative from the bottom up. The idea of making R & D a line function was a strategic choice based on the belief that the key to maintaining competitive advantage as a low-cost steel supplier was combining manufacturing with technology. Thus, the physical and organizational structures of Chaparral were all carefully aimed at integrating work and experimentation through what we call online/internal mechanisms.

One of the key decisions in the design of online/internal mechanisms was keeping the number of employees at Chaparral Steel to less than 1,000. Small size facilitated both knowledge production and dissemination. It enabled employees to know each other personally, influencing the desire and willingness to share information. Even more important, it made it easier for employees to develop “transactive knowledge” (see  Chapter 7  on dissemination). In other words, it was possible for every Chaparral employee to know whose expertise they needed in order to solve a problem or create an innovation. It also enabled them to know who might need or make good use of their own knowledge. Forward considered small size (in terms of number of employees) so important that he stated flatly that he “simply wouldn't consider” getting much bigger in the present location (Kantrow, 1986, p. 97). Rather than increase its scale of operations to stay competitive, Chaparral sought strategic opportunities that enabled it to maintain its small size.

The physical design of Chaparral Steel, which was related to size, provided an infrastructure conducive to the development of online/internal agent OLMs. First, all of Chaparral's operations were located in one plant, enabling employees to exchange information face to face without travel or sophisticated knowledge management technologies. Other features of the plant's physical design facilitated accidental meetings among all levels. The plant headquarters were situated in close proximity to the furnaces and mills as well as the employees’ locker room (Leonard-Barton, 1992). Workers and line managers with ideas or problems did not have to go far to meet with higher levels of management. Indeed, they were likely to run into each other at least once or twice a day so that no one would have to wait very long to move information up or down. In fact, there was nothing really “accidental” about these meetings at all because they were facilitated by conscious design.

The organizational design of Chaparral complemented the physical design in enabling online/internal agent OLMs to take shape just about anywhere or anytime. The one-story headquarters building was symbolic of the flat organization that kept hierarchical distance and formality to a minimum. From the standpoint of management, informal onthe-spots meetings were neither interruptions nor distractions but rather integral parts their work. In addition, there was an intentional absence of the normal bureaucratic barriers to innovation. Chaparral employees were not required to follow bureaucratic procedures or receive management approval before trying out improvements. Decision making was delegated down the hierarchy so that even line supervisors could authorize tens of thousands of dollars for experiments without receiving permission from above (Leonard-Barton, 1992).

Rather than emphasize clear roles and a strict division of labor, management assigned both units and individuals to numerous tasks. Everyone had line responsibilities and operators performed 40% of the maintenance work. The company had a marketing department, but every employee was considered to be in sales (Leonard-Barton, 1992). Another unusual strategy for blurring boundaries was requiring all employees to work night shifts regardless of seniority. This strategy meant that organizational knowledge was not concentrated during particular hours but well dispersed across the entire 24-hour day (Leonard-Barton, 1992).

Requiring employees to fulfill multiple functions and develop multiple skills enlarged the scope of their interests, understanding, and involvement in Chaparral's operations and activities. It also increased the likelihood that employees would encounter and learn from people outside their functional specialty. This design feature made the organization more flexible, counteracted competition over turf, and encouraged knowledge sharing.

Not all organizational learning at Chaparral took place through online OLMs. There were also regularly scheduled, off-line meetings for discussing problems and ideas for improvement. However, a foreman estimated that “90 percent of the problems never make it to the morning meetings” because they are solved online (Leonard-Barton, 1995, p. 8). In other words, most problems were dealt with through the online mechanisms. This quote attests to the power and efficiency of these mechanisms.

Another common off-line OLM at Chaparral was travel. As mentioned, the company encouraged all employees to travel for the purpose of acquiring new knowledge. Teams of managers, foremen, and workers visited suppliers and competitors in the United States, Japan, and Europe. Gordon Forward observed that many important breakthroughs occurred during travel because the time together led to openness (Luthans, 1991). In addition, research projects were cosponsored with U.S. academic institutions to import cutting-edge knowledge into Chaparral Steel.

Whether or not new knowledge was acquired internally or from without, the efficient dissemination of knowledge was generated by the same features of the organization that promoted online/internal OLMs for experimentation. The intentional blurring of horizontal and vertical boundaries facilitated intensive networking and knowledge sharing among employees. The lack of bureaucratic barriers meant that results of experimentation and knowledge imported from outside could be quickly implemented and disseminated. Organizational members had no trouble imitating others and adopting their performance improvements as standard practice (Leonard-Barton, 1995).

OLMs for dissemination, however, were not left to chance. Chaparral used some well-planned strategies for making sure that knowledge was shared:

For instance, in “commissioning” (ramping-up to problem-free production) of [a new mill] only two teams of operators are being trained. Each team works a twelve hour shift (with paid overtime). After the initial eight weeks of grueling schedule, these operators will be dispersed among the rest of the crews, to diffuse the knowledge they have created and assimilate about the idiosyncrasies of the [production] process. (Leonard-Barton, 1995, p. 13)

This method of disseminating knowledge by dispersing a highly skilled group of workers is common practice at Chaparral.

As this example illustrates, knowledge at Chaparral was transmitted primarily through people and interpersonal interactions. Face-to-face knowledge transmission was quicker and more efficient than engaging in a process of making knowledge explicit, generating written reports, translating it into formal procedures, and then transmitting it through training and support. It was also self-reinforcing and a bit risky. Because so much knowledge was transmitted informally and could not be accessed through written documents, people needed to interact and communicate directly with others if they wanted to stay on top of things.

 PRODUCTIVE LEARNING AT CHAPARRAL STEEL

Viewed through the multi-facet model, the existence of online/internal agent OLMs provided a necessary but insufficient condition for productive learning. Understanding productive organizational learning requires a systematic analysis of the culture, the psychological climate, specific management policies, and the context in which learning occurs. In this section, we look at each of these facets as they apply to Chaparral Steel.

The Cultural Facet

As described in  Chapter 3 , the cultural facet refers to the behavioral norms and the underlying values that regulate behavior in OLMs. The various accounts of Chaparral Steel provide ample evidence of the existence of the five norms that we consider essential for productive learning: inquiry, issue orientation, transparency, integrity, and accountability.

Norms of inquiry were clearly evident in Gordon Forward's description of the influence of his educational background on his role as the CEO:

I feel my research background [showed me] how to deal with the mystery and intrigue of solving the puzzle. To me, the inquiry, the breaking through to new vistas, the finding of solutions, is what it's all about. (Luthans, 1991, p. 64)

This same norm of inquiry was reflected in numerous comments from workers to the effect that they were always asking questions and constantly trying to learn more about what they were doing (Leonard-Barton, 1992). It was also evident in the example given earlier of employees who doggedly pursued information and knowledge from a variety of sources in order to fully understand a problem and arrive at a solution.

Issue orientation was manifested by egalitarianism and respect for the individual, which were shared values in Chaparral Steel (Leonard-Barton, 1992). The strength of issue orientation at Chaparral was summed up nicely by this statement: “Knowledge is valued not so much for the pedigree of its source but for its usefulness” (Leonard-Barton, 1992, p. 36). Everyone's ideas were taken into consideration, regardless of function or level in the hierarchy. Indeed, the ideas and opinions of operators were especially valued because they were the closest to the work itself. Another element of issue orientation is the fact that innovations were usually seen as team efforts involving multiple inputs rather than individual accomplishments (Leonard-Barton, 1992).

Transparency was closely coupled with inquiry and issue orientation at Chaparral. Employees at all levels were encouraged to say what they think—even to the CEO. Gordon Forward put it this way:

I actually like to deal in incomplete thoughts—in other words, a thought that you haven't really fleshed out yet, but one that you want to bounce off of someone. It's kind of like brainstorming. I've seen… more stilted structures stifle this kind of interaction. The approach was, “Don't talk to me until you've come to a conclusion.” (Luthans, 1991, p. 70)

Chaparral employees repeatedly stressed that they freely shared their ideas even if they were unsure about their quality (Leonard-Barton, 1992).

Forward also gave expression to the norms of integrity (and inquiry), which we have defined as a willingness to admit ignorance or error:

When you're operating in a technical field… trying to go one step beyond research, one of the things you learn fast is that you can't fool yourself…. You've got to be open in your questioning…. And you can't succeed by pretending to know things that you really don't. You have to go find them out. You have to try an experiment here, and experiment there, make your mistakes, ask your questions, and learn from it all. (Kantrow, 1986, p. 101)

The norm of integrity was also manifested in a millwright's report that he was actually pleased when someone less knowledgeable than he caught him in an oversight because it saved him from making a costly error (Leonard-Barton, 1992).

Finally, in Chaparral Steel's culture, accountability was expressed in a number of ways. The proper response to making mistakes was to admit them and then take action to fix them (Leonard-Barton, 1992). Because of blurred functional distinctions and the general ownership of problems, everyone was responsible for ensuring results, and no one could hide behind the claim that “it's not my job” (Leonard-Barton, 1992, p. 27). Accountability was also reflected in the wide latitude for experimentation but also in the demand to apply successful results. As illustrated in the description of the online/internal agent OLM, all employees were expected to act on and disseminate new knowledge.

The Psychological Facet

The published accounts of Chaparral Steel did not delve deeply into the psychological climate of the organization. Nevertheless, there was enough evidence to make inferences about both psychological safety and commitment to the organization. The clearly evident norms of transparency and integrity implied that employees experience a high degree psychological safety. In one telling quote, Forward provided an interesting perspective on this facet:

In a research laboratory, risk is accepted as the norm, since the cutting edge is fraught with uncertainty. In contrast, risk is usually anathema in a production environment…. Chaparral managers avoid riskless projects because a “sure thing” holds no promise of competitive advantage…. This positive attitude toward risk permeates the company…. If everyone experiments, learns, and innovates, then neither success nor failure can be heavily personalized. (Leonard-Barton, 1992, p. 32)

To the extent that this statement reflected the thinking of organizational members in general, it contains two critical assumptions that apply to psychological safety. First, mistakes were considered not only normal but actually necessary occurrences, so there was no shame in making and admitting them. Second, everyone made mistakes, and most innovations were collective efforts. There was little fear of being blamed personally for errors.

The multi-facet model also maintains that people will expend the extra effort and take the risks involved in learning to the extent that they feel committed to the organization. Leonard-Barton (1992) described the workforce at Chaparral as “highly dedicated” (p. 33). The organization was deliberately understaffed and demanded much more of employees than most organizations, but employee motivation appeared to be very high and absenteeism very low (Leonard-Barton, 1992), providing some indication of high commitment.

The psychological climate and the cultural norms at Chaparral reinforced each other. Their positive strength accounted for the effective functioning of the online/internal agent OLMs. In this respect, Chaparral differs from organizations in which effective OLMs constitute cultural islands where learning norms may be more operative in OLMs than in the organization as a whole. Just as the online/internal agent OLMs were tightly integrated into the flow of work, the learning norms appeared to dominate the organizational culture as a whole.

Leadership

By this point, it should be obvious that the leadership of Gordon Forward, who was with the organization from its founding in 1973 until his retirement, played a key role in shaping Chaparral Steel as a learning organization. This is evident from the impact on the organization of his training as a researcher and his personal experience. It is interesting that the descriptions of Forward and interviews with him say little about his personality or leadership style. Instead these sources indicate two channels of influence (see  Chapter 6 ) through which Forward and his top management team promoted the strategy of online/internal agency learning in Chaparral Steel. The first channel was Chaparral's overall mission. As we have already noted, workers in Chaparral Steel were assigned the dual task of performing their work and improving their production processes. The second channel was a set of policies that either promote organizational learning directly or supported it through human resource practices.

The Policy Facet

The multi-facet model points to three issues that need to be addressed by organizational policy in order to promote learning: tolerance for error, commitment to learning, and commitment to the work-force. In his interview with Fred Luthans, Forward said, “We don't have policies” but rather “basic ideas.” This was, of course, stretching a point. However, Forward was expressing his disdain for bureaucratic rules and procedures that are “designed to catch the 3 percent who [are] trying to cheat in one way or another” (Luthans, 1991, p. 69). By minimizing the number of formal rules and procedures, Chaparral Steel manifests trust in its workforce, trust that they will rise to and meet the challenge of beating the competition through continuous innovativeness and superior efficiency.

From the founding of Chaparral Steel, management instituted policies to support the organization's multiple missions of production, problem solving, improvement of production and products, importing new knowledge into the organization, and sharing knowledge with others. These policies facilitated the development of online/internal agent OLMs and cultivated the cultural norms and psychological climate conducive to effectiveness.

Developing policies that promote a tolerance for error is one of the most difficult areas for management because it is not always possible to clearly distinguish between errors that stem from legitimate risk taking and those that stem from negligence or incompetence. Nevertheless, Chaparral had a policy that provided employees with the freedom to take risks. Gordon Forward explained this policy as follows:

People say [risk taking] isn't very smart. But I respond that I think what they are doing is really risky. They ask, “What do you mean?” I say, “To me, the biggest risk, particularly in the business world of today, is to do nothing.” Plenty of firms are afraid to take risks and I see this reluctance as a major risk itself…. The other thing is that people need the freedom to [achieve their goals]. We all figure we'll make some mistakes along the way, but that freedom is important. (Luthans, 1991, p. 69)

Tolerance for error is manifested in Chaparral Steel by the preference for new recruits who are risk takers and for deliberately taking conscious risks.

This policy was also evident in the story of one of the company vice presidents who, as a supervisor, purchased new equipment that totally failed and cost the company $1.5 million. He himself was not punished for his error. Furthermore, he explained that employees were rewarded for having new ideas by being given the freedom to test them out, even despite the obvious risks. The policy of tolerance for error clearly trickled down to managers who, as mentioned before, avoided projects that entailed little risk but also gained little. By the same token, mistakes were admitted and corrected rather than covered up (Leonard-Barton, 1992).

Chaparral's belief in people and commitment to the workforce was clearly articulated by Forward and expressed through its employment policies from the beginning of the company. Forward recognized that organizational commitment was governed by a norm of reciprocity. If Chaparral were to gain its employees’ commitment, it had to show commitment to its employees. They had to feel that Chaparral had their interests in mind and that they belonged (Luthans, 1991).

Chaparral implemented this principle in treating all employees equally, investing in their personal development, and letting them share in the company's profits and wealth. For example, every employee was placed on salary rather than hourly wages, a manifestation and symbol of egalitarianism and respect for workers that contributed to organizational commitment. Bonuses for everyone were linked to company profits. There was also a profit-sharing scheme and a program through which most employees were stockholders (Leonard-Barton, 1995).

Not surprisingly, given the organization's multiple mission, many policies that expressed commitment to the workforce also expressed a commitment to learning. When Chaparral was first set up, management decided not to look for workers with industry experience. Rather, they sought employees who enjoyed hard work and had a high potential for learning so that they could be socialized into the unique learning culture of Chaparral (Leonard-Barton, 1992). Job applicants underwent a highly selective and rigorous screening process that included extensive interviews with Chaparral employees at different levels. Pay was structured in a way that rewarded learning and skill development as well as performance (Leonard-Barton, 1992).

Chaparral invested heavily in employee development and training that focused both on helping employees self-actualize and improve performance. Eighty-five percent of Chaparral's employees were enrolled in courses with direct or fairly direct relevance to their jobs. The company had an extensive apprenticeship program for every employee, which included courses on generic skills and company-specific knowledge as well as on-the-job training for workers. In order to keep supervisors excited about their work, they were encouraged to take time off from their jobs for paid “sabbaticals” involving work-related projects. Chaparral used these absences to give senior operators opportunities to fill in for their supervisors. As a result, they gained valuable supervisory training and experience (Leonard-Barton, 1992).

 THE CONTEXTUAL FACET

The final stage in our analysis of Chaparral Steel looks at the context in which organizational learning emerged. The contextual facet is different from all of the other facets because it focuses on factors over which management has limited or no control. Managers can set up OLMs as well as set policies and exercise leadership that influence the psychological climate and shape cultural norms. However, the contextual facets—error criticality, environmental uncertainty, task structure, and proximity to the core mission—are relative givens. They represent both opportunities and constraints to which managers must respond and adapt in promoting organizational learning. Thus, identifying and understanding the contextual facet is particularly important for managers wishing to apply the principles of organizational learning from Chaparral Steel, or any organization, to their own.

The most obvious contextual factor in understanding Chaparral's success was the fact that it was a greenfield site. Although Forward and his associates did not think in terms of the learning organization, they carefully designed and built the organization from the bottom up with the intention of integrating production and learning. They did not have to contend with and change the existing physical plant, organizational structure, and culture to promote learning. Managers attempting to apply the lessons of Chaparral in an existing organization would need to consider a very different strategy for getting there.

The case of Chaparral Steel is consistent with the contention that a relatively high degree of environmental uncertainty creates a demand for organizational learning. For Chaparral, the major source of uncertainty was defining and achieving competitive advantage in a changing industry:

We realized that this business has traditionally been labor-intensive, capital-intensive, and energy-intensive, and from the beginning, we knew that we had to design all of those things out of it. That was the only way we could be competitive with Third World countries. (Luthans, 1991, p. 65)

Chaparral Steel was designed from inception to be an international low-cost supplier of steel products, but that meant competing with Third World companies with access to a much lower-wage labor force. From the beginning, Chaparral had to invent a new strategy for succeeding in an environment where most American producers had failed.

Essentially learning became a strategic imperative and part of Chaparral Steel's core mission:

We are always trying to push back the technological frontier…. Maybe our largest challenge is to cut the time it takes to get technology out of the lab and into operations. The kind of lags that many industries experience would simply kill us. (Kantrow, 1986, p. 99)

To be competitive, Chaparral aimed at attaining 1.5 to 3 worker hours per ton by enlisting accessibility to the most advanced technology and to a highly committed innovative workforce (Luthans, 1991). To engage workers’ commitment and encourage their innovativeness, Forward formulated a dual-purpose mission that called for the simultaneous manufacturing of cutting-edge products that require the invention of “test-the-limit” designs and the continuous improvement of production processes (to maintain cost advantage). Because both objectives require learning and experimentation, learning fused with task performance and wove itself into the organization's mission.

The question for other organizations that wish to learn from Chaparral's experience is whether the nature and degree of environmental uncertainty demands the same learning imperative. Learning is only one strategy for achieving competitive advantage. The fact that all organizations face increasingly intense competition in the age of globalization does not necessarily mean that learning needs to be defined as part of the organization's core mission. Conversely, managers who wish to promote learning need to take into account the proximity of their areas of responsibility to the core mission of the organization. The multi-facet model suggests that units or processes that are ancillary to the core mission may not be able to attract the resources and commitment to learning evident in Chaparral and necessary for online/internal OLMs.

Superficially at least, error criticality does not seem to be a major driving force for learning in a steel manufacturing organization. Mistakes in steel manufacturing are costly but usually not fatal. Nevertheless, Forward did express the belief that, given the intense competition, any false move could prove fatal to the organization:

We have to go like hell all of the time. If the price of what we sell goes up too high… all of sudden lots of folks will be jumping in. …We constantly chip away the ground we stand on. (Leonard-Barton, 1992, p. 24)

As evident from Forward's attitude toward risk, he believed that the greatest danger in the current situation is not taking risks, not pushing into the unknown for the purposes of learning.

The task structure is an important but often overlooked factor in shaping the potential for organizational learning. Chaparral was a minimill, which meant that it focused entirely on the production and sales of steel. In this respect, it was different from traditional U.S. steel producers that brought together mining, transport, and manufacturing of a wide variety of products, all under one organizational roof. Being a minimill made it possible for Chaparral to maintain its small size and to quite literally bring all its operation under one physical and organizational roof. As the case illustrates, these factors were extremely important in creating conditions for developing successful online/internal agent OLMs and for minimizing bureaucratic constraints.

The fact that Chaparral manufactured a small number of very specialized steel products also facilitated organizational learning. It made it relatively easy for the company to sharply focus the relevant knowledge it needed, to match people with this knowledge, and to bring them together. Creating the necessary configurations of knowledge through online/internal OLMs would have been much more difficult if Chaparral were producing a wide range of products requiring different kinds of knowledge. Another factor that facilitated learning was the fact that Chaparral was producing a commodity and not a finished product made up of many different components with their own specialized technologies. It would have been much harder, though not impossible, to use online/internal OLMs for producing innovations in automobiles or even automobile manufacturing methods because automobiles involve so many different parts, materials, systems, and technologies.

Production technologies and work processes themselves also influence the ways in which learning can occur. Organizational learning structures are shaped and constrained by the structure of work, its pacing, its distribution across time and space, the numbers and types of people involved, the kinds of information involved in the work, and a variety of other factors. Leonard-Barton hinted at these differences in her description of learning at Chaparral:

Just as continuous processing has great advantages for manufacturing over most batch processing, so the unimpeded flow of information aids learning more than fragmented, batch-processed information. (Leonard-Barton, 1992, p. 29)

The analogy between production processes and information processing can be turned around to suggest that information flow will be different in continuous processing environments than in large batch-processing environments. Woodward's (1965) classic study showed that relative to mass production, effective continuous process environments have shorter spans of control, more managers per worker, more skilled workers, lower formalized procedures, less centralization, higher verbal communication, less written communication, and a more organic structure. These features highlight the fact that, by nature, continuous process organizations emphasize problem solving, quick reaction times, and flexibility. The only major exception is that, according to Woodward's model, continuous process organizations require more levels of hierarchy, whereas Chaparral's productive learning flattened the organization.

It can be clearly seen that there is a kind of fit between these characteristics and Chaparral's extensive use of online/internal agent OLMs, as well as its learning policies and culture. Forward suggested that Chaparral itself is actually not so unique in this regard:

We're not the only ones—there are others like us. Nucor does many of the same things, but it has a slightly different personality. … There are a number of quality minimills. We are all a bit different, but we all have to run like hell. (Kantrow, 1986, p. 97)

The point is not that continuous process manufacturing organizations are more likely to be learning organizations but that successful organizations adapt configurations to their particular tasks. Given similar strategic situations and similar technologies, organizations may develop very similar ways of learning.

 CONCLUSION

There is no doubt that Chaparral was a very effective and highly impressive learning organization. However, the multi-facet analysis shows that managers ought to be wary of blind imitation. Organizational learning can take many different forms, as reflected in the variety of OLMs. The effectiveness and intensity of learning at Chaparral was achieved through carefully crafted online/internal OLMs designed to fit the strategic situation, the task structure, and other factors. The idea of fit does not suggest a simple contingency model because there are simply too many factors involved to make such a model credible or useful in practice.

Perhaps the most important lesson from Chaparral itself was actually that managers need to take an experimental approach to organizational learning. On one hand, they need to systematically think through the kinds of OLMs and policies that are necessary, feasible, and appropriate, given the existing strategy, task, and culture. On the other hand, they need to take each step in implementing these OLMs and policies as an opportunity to see what works, what can be reinforced, and what needs to be done differently. Although it makes eminently good sense to see the learning organization as a holistic system, the way to get there, at least for most managers, is one step at a time.

11

Demystifying the Learning Organization I

The Case of Hewlett-Packard

When Peter Senge (1990) reframed the issue of organizational learning as the “art and practice of the learning organization,” he transformed an esoteric academic concept into a popular vision of a brave new world of organization. This framing of the issue, however, created an implicit distinction between companies that learn and those that do not. Quite naturally, forward-looking, self-respecting CEOs wanted their companies to be identified as the former (Gerhardi, 1999). This distinction also implied that becoming a “learning organization” requires a deep and far-reaching transformation. Thus, before anyone could clearly define this distinction in operational terms, many managers aspired to fundamentally change their organizations either through an application of Senge's five disciplines or some other method. As we pointed out in the introductory chapter, this gap between the vision and practice persists and contributes to mystification.

Demystifying the learning organization means stepping back from “either-or” or transformational thinking. As we have pointed out in earlier chapters, organizational learning can be understood by observing the structural mechanisms (OLMs) that make individual learning organizational and at the contextual, policy, psychological, and cultural factors that influence the effectiveness of the learning processes that take place within these structures. Structural organizational learning mechanisms exist in some form in most, if not all, organizations. If they did not, organizations could not survive for long in a changing world. In this respect, talking about “the” learning organization creates a false dichotomy and distracts attention from the everyday managerial actions that promote learning, although not necessarily transforming the organization as a whole.

The claim that all organizations learn does not mean that all organizational learning is equal. The literature that has emerged over the past decade clearly indicates that some organizations, or organizational units, excel at knowledge creation and management. The question, then, is what actions managers can take to enable their organizations to learn more, faster, and better. In the next two chapters, we will take close look at exemplary organizations, Hewlett-Packard Consulting (this chapter) and Chaparral Steel ( Chapter 12 ), in order to illustrate that they invest heavily in both OLMs and in policies that promote learning under a given set of strategic and task conditions. These observations suggest an observable, operative definition of the learning organization that can guide management action.

If any organization deserves to be an exemplar of organizational learning and knowledge management, it is Hewlett-Packard (Davenport & Prusak, 1998; Parskey & Martiny, 2000). Founded in Dave Packard's Palo Alto garage in 1939, HP was ranked 11th in the Fortune 500 in 2005 with $73.061 billion in revenues, more than 140,000 employees in 178 countries, and thousands of products in IT infrastructure, global services, business and home computing, imaging, printing, and Web-related applications. In an industry known for its turbulence and the meteoric rise and fall of companies, HP has maintained consistent growth and leadership. HP is also well known for an organizational culture that promotes technological and managerial innovation, employee commitment, and social responsibility.

This chapter examines two case studies of organizational learning at Hewlett-Packard. Although both cases illustrate productive learning, one of these initiatives was discontinued whereas the other spread and became institutionalized. By comparing the two cases, we will attempt to understand both the design of the learning organization and the factors that lead to more sustainable learning processes. The example of the successful program will show that organizational learning can be promoted in a gradual, but deliberate, process that does necessarily involve transformation or heroic measures.

 HP'S WORK INNOVATION NETWORK (WIN)

Zell (1997, 2001) tells the story of HP's “Work Innovation Network” (WIN). In 1990, a group of internal consultants from HP's Strategic Change Group began to teach HP managers how to improve the productivity of their units by organizing them as self-managing teams. The initial change strategy was to illustrate the benefits of self-managing teams to the wider organization by conducting intensive pilot projects with a small number of sites throughout the organization. These initial efforts led to reports of substantial improvements in quality, productivity, and responsiveness, which spread quickly throughout the company and generated great interest in work redesign. Within a short time, this group found itself unable to keep up with the demand for consulting services (Zell, 2001). In response, the consultants created WIN, through which participating managers met once each quarter for 2 to 3 days in order to learn about work redesign and autonomous teams.

WIN was based on a set of explicit principles for countering common barriers to the diffusion of innovation in organizations. The first principle was avoiding forced, top-down change by enabling managers to “pull” information and innovations that they needed on a justin-time basis. Thus, participation in WIN was strictly voluntary and restricted to managers who expressed a willingness to actually experiment with self-managing teams. WIN itself was self-managed—the content and format of each meeting was planned by consultants and a rotating group of participating managers who continuously sensed the needs of the network.

WIN meetings were a forum in which managers could discuss their own redesign efforts among themselves as well as with the consultants. Networking was facilitated through seating arrangements and informal meetings in which participants could choose both their partners and the topics of discussion. Once participants began to implement redesign, they presented detailed cases at forum meetings. Peer consulting was used as a method for reflecting on experience and for developing skill in giving and receiving feedback. “Networking” allowed managers to maximize face-to-face communication among units, to facilitate the dissemination of information and knowledge, and to form alliances to push through changes that did not receive initial top-down support.

Creating a “safe environment” (that is, psychological safety) was considered extremely important for encouraging open and honest dialogue in which members felt comfortable sharing both their successes and failures. As WIN grew, some members began to feel that presentations became less personal and truthful:

A WIN member circulated an e-mail urging a return to presenting the whole picture of redesign efforts. “We're veering away from honesty,” the memo said. “We need to return to personal disclosure about presenters’ sources of pride and disappointment, because that's what the audience learns from most” (Zell, 2001, p. 81).

Maintaining a safe environment was also the reasoning behind the rejection of a suggestion to videotape the peer-consulting sessions.

Finally, WIN consultants introduced “action research” as a method for ensuring high-quality and relevant solutions, as well as systematic evaluation of both processes and outcomes. Participating managers were required to document their learning in “white papers” that described their objectives, methodologies, impact on organizational performance, and lessons learned. These papers were then distributed and discussed by the participants at each WIN meeting so that the learning could be disseminated, helping others deal with similar obstacles (Zell, 2001, p. 81).

According to participants’ feedback, WIN was a very successful project. Eighty percent said they had received full return on their investment; more than two thirds said it gave them moral support and good ideas on how to implement self-managing teams. The WIN network was considered a more valuable source of information than HP's internal consultants and other sources of information. Anecdotal evidence provided some support for the effectiveness of the transition of self-managing teams, which was adopted in one or more units of a third of HP's twenty divisions.

In 1995, WIN's existence was threatened by a companywide effort to cut costs that banned all nonessential travel and meetings of over 20 people. WIN's leaders argued for treating WIN as a special case so as “not to send the wrong message” to the company. However, these efforts failed to persuade management, and the program was discontinued.

 THE HPC KNOWLEDGE MANAGEMENT INITIATIVE

HPnsulting (HPC) is the 4,000-employee-strong service wing of Hewlett-Packard. Its mission is to “ensure that customers get the greatest return from the products and services they buy from HP” (Parskey & Martiny, 2000, p. 27). The HPC “Knowledge Initiative” was started to help HPC consultants to “tap into the knowledge of the broader HP organization to solve [clients’] business issues” (Parskey & Martiny, 2000, p. 28). Our analysis of the initiative is based on two insider accounts written by the managers of HPC's Global Knowledge Management Marketing and Global Knowledge Services (Martiny, 1998; Parskey & Martiny, 2000).

In 1995, a team of consultants interviewed HPC's clients on the effectiveness of the organization in using HPC knowledge in their work. The survey results, as reported to HPC management, revealed that clients believed that the value and depth of HPC knowledge is highly dependent on the consultants assigned to their project. These results were disturbing because HPC felt that it had to deliver ever increasing new and innovative services in a consistent, high-quality manner, regardless of the consultants assigned to the engagement. In order to meet this challenge, the survey concluded, HPC “must rapidly leverage our experience from one project to the next and from one part of the world to another” (Parskey & Martiny, 2000, p. 28).

The results of the survey overcame doubts among HPC consultants about the need for increased knowledge sharing and reuse in meeting customer demand. More important, HPC's new vice president and general manager realized that organizational learning and knowledge management were critical to the success of the business. In order to respond to this need, the general manager launched the “Knowledge Management Initiative” and remained actively involved in it throughout its operation.

HPC's senior management team set forth a vision in which HPC consultants “feel and act as if they have the knowledge of the entire organization at their fingertips when they consult with customers.” (Parskey & Martiny, 2000, p. 30) Achieving this state required making this knowledge explicit and available. It also meant that consultants would have to know how to locate the specific information and knowledge they needed. Most important, it required willingness on their part to share their knowledge and to learn from one another's experience. The vision also linked these behaviors to an organizational reward system recognizing “those consultants that share and those that leverage others’ knowledge and experience as the most valuable members of the HP team” (Parskey & Martiny, 2000, p. 30).

The project began with the establishment of the Organizational Learning/Knowledge Management (OL/KM) Team to lead the initiative. This team specified three measurable definitions of success for the initiative: learning from the success and failure of different projects, sharing knowledge among projects, and minimizing redundant “innovation” (reinventing the wheel). In addition, three mechanisms were established for achieving these objectives. The first mechanism was a “postproject review” process to facilitate learning from successes and failures of completed projects. The guidelines for running the reviews promoted openness and trust by setting a number of ground rules: All members of the project team would participate; lessons learned would be based on consensus; and the emphasis would be on learning rather than blaming or finding fault.

A second OLM, periodic “project snapshots,” were instituted for ongoing efforts. This mechanism consisted of team meetings devoted entirely to identifying and documenting knowledge that had been generated in the course of the project so far and that could be used by other teams. The third mechanism, “learning communities,” was defined as “informal groups of people that cross organizational boundaries and meet… face to face or through conference calls…to discuss best practices, issues, or skills that the group wants to learn about” (Parskey & Martiny, 2000, p. 39). The third objective, minimization of reinventing the wheel, was facilitated by developing generic versions (“genericizing”) of tools that were developed by consultants to assess clients’ requirements in connection with particular HP products.

Specific criteria were set for selecting the specific business issues for the pilot studies that would launch the initiative (Martiny, 1998). These criteria included high visibility and strategic importance for the success of HP, high readiness based on a business need for knowledge sharing, and a strongly committed leader. HP's SAP (a comprehensive organizational resource management package) practice in North America met these criteria and was of special interest because a previous knowledge management effort, based on the use of Lotus Notes database of lessons learned, had failed due to low consultant participation. Thus, the pilot project, which was dubbed “OWL– Orchestrating Wisdom and Learning,” was expressly not technology driven. Rather it was designed to “develop a committed core consultant group that would identify, share, and leverage knowledge for the benefit of others in the organization… and then become advocates for the value of knowledge management” (Parskey & Martiny, 2000, p. 35).

The OL/KM Team that was assigned to lead the initiative began by knowledge mapping. A group of experienced consultants was asked to break down into key components or steps the consultant's core activity of selling or delivering a solution. On the basis of an analysis of specific examples for each component, they identified what the consultant needs to know, where this knowledge comes from, who “owns” it, what skills are required, what tools or templates exist or should becreated, and what barriers exist to effective execution. Based on this information, the team developed a new solution delivery approach. It also developed a 5-day workshop to acquaint consultants with this approach, as well as the benefits of organizational learning/knowledge management.

The workshop was designed not simply to train the consultants but to involve them in the expansion of the KM initiative as well. Consultants were required to develop OL/KM implementation plans tailored to their business units’ specific needs for creating, sharing, and leveraging knowledge. Learning communities formed and project teams shared their experiences in delivering new solutions to clients. New solutions were documented and made available to other consultants. Through these OLMs, the new delivery approach was spread throughout the entire North American SAP consulting business. It was credited with the decrease in clients’ implementation cycles from 18 to 6 months (Parskey & Martiny, 2000, pp. 37–39).

At the same time that the Knowledge Management (KM) Initiative was started, HPC was deploying a project management methodology worldwide. The OL/KM Team made a strategic decision to use this program as a vehicle for integrating its tools into the core project management practices of HPC. Owing to its phased nature, project work was an ideal ground for project snapshots and postproject reviews. For example, the transition from the selling phase to the delivery of the solution represented a natural point for conducting project snapshots. To smooth the integration of project snapshots with project management, the OL/KM Team developed a project snapshot training module that was included in HPC consultants’ training in project management.

Learning communities also became widespread in HPC. Martiny (1998) provides a detailed description of what went on when learning communities met:

During a learning community session, a consultant spoke up about a technical assessment he would be conducting the following day at a client site, and asked other participants for any experience or assessment tools they had used in similar situations. A participant piped in that he had recently completed the development of an assessment tool and reviewed how he delivered it successfully in a client engagement. He then sent the assessment tool electronically to the consultant in need. The second consultant evolved the assessment tool for his application, leveraging both the first consultant's experience and the tool to deliver the assessment to his client. At the next Learning Community session, the second consultant shared his newly gained experience including what worked well, what he had learned, and what he would do differently next time. He made the enhanced assessment tool available to the Learning Community and both he and the first consultant offered to be contacts for additional information or mentoring. (pp. 75–76)

According to Martiny (1998), learning communities were credited with reducing delivery time while improving quality through the sharing of best practices. They also increased productivity by facilitating the sharing of tacit knowledge and by helping to standardize proposal and presentation materials.

The solution development groups, which developed new products (“solutions” in HP jargon), found that knowledge mapping was useful for identifying what knowledge was needed—and sometimes already available—for developing or delivering new solutions. They held 2-day knowledge mapping workshops involving a wide range of consultants from around the world. These workshops also served as a catalyst for creating new relationships and forming “learning communities” among participating consultants (Parskey & Martiny, 2000, p. 41).

As project reviews and snapshots and learning communities took hold throughout HPC, the organization added new knowledge management roles and developed an electronic infrastructure and an incentive system to support them. “Knowledge service managers” were assigned responsibility for identifying local knowledge needs and priorities, as well as for providing and implementing appropriate solutions in the highly decentralized HPC organization. “Knowledge architect managers” and “solution development knowledge managers” developed tools and methodologies to enhance OL/KM in HPC. “Learning community lead people” were responsible for the conduct and smooth functioning of learning communities within their subject domains.

The role of “knowledge consultant” was created to overcome the difficulty of promoting and coordinating organizational learning in an organization that practices and is ideologically committed to decentralization. Knowledge consultants facilitated knowledge flow around HPC and linked the central OL/KM Team with the different field units. They reported directly to the OL/KM Team and maintained extensive personal networks and good understanding of local cultures. For example, knowledge consultants recommended that the focus on learning communities and project snapshots is less necessary for Japanese than for Western consultants, because the Japanese tend to be naturally adept at sharing tacit knowledge, which is difficult to verbalize or communicate directly. The outcome was that in Japan the focus shifted from learning communities to supporting the management of “explicitknowledge” by building a repository for sharing documents (Parskey & Martiny, 2000, pp. 46–47).

In order to support OL/KM, HPC developed a Web portal that allowed consultants worldwide access to a database of documents stored electronically. This portal was supervised by the “HPC Standards Board,” which set guidelines for standardizing the elements accessible through the portal, knowledge structures, service portfolio development, and collateral creation. HPC management also instituted “Knowledge Master Awards,” which recognized and rewarded individuals who exemplified the highest standards of knowledge mastery. These individuals received an all-expense-paid vacation or cash award, but the main idea was to make them role models for the rest of the organization. They received HPC-wide recognition, and their stories were widely publicized. These stories emphasized improved business results through leveraging and reusing of knowledge, as well as through fundamental values such as expertise, teamwork, and willingness to contribute to the overall good of the organization (Parskey & Martiny, 2000, p. 461).

As of the year 2000, HPC's OL/KM initiative was regarded as a success story. To sustain its success, the OL/KM Team intended to concentrate on three areas that, based on its experience, it identified as key success factors: cultivating leadership involvement and support, reinforcing OL/KM performance by developing OL/KM evaluation metrics, and developing a method for gauging the progress of entire units in terms of OL/KM. The common theme underlying all efforts was the integration of the OL/KM projects with HPC's core activities. Cultivating leadership involved the identification of OL/KM sponsors who would participate in prioritizing knowledge initiatives, setting clear expectations, and serving as role models. In addition, it meant incorporating knowledge-based measures into senior managers’ roles and adding training in OL/KM to management development programs. Another important part of cultivating leadership was keeping sponsors involved and informed of progress, successes, and challenges in the wider HPC OL/KM initiative (Parskey & Martiny, 2000, p. 48).

The second area for reinforcing OL/KM performance was evaluation. Commitment to sharing knowledge was now considered to be an important behavior for HPC consultants. In order to promote this behavior and to use it as criteria for reward and promotion, HPC had to develop ways of measuring commitment to sharing knowledge for every role in HPC. There was also a need to educate senior management on the existence of these measures and to encourage their use in communicating expectations and monitoring progress.

HPC used a “maturity” of knowledge model as a method for gauging the progress of entire units in terms of OL/KM. Maturity models provided managers with “concrete tangible ways of… comparing explicit competencies at the program and project level relative to a standard” (Kam & Thomas, 2002, p. 4). HPC's knowledge maturity model divided the knowledge management capacities into five levels: ad hoc, repeatable, defined, managed, and optimized. The characteristics of each level were clearly defined so that they could be used as criteria for assessing the state of knowledge management in a particular unit.

 WHY DID WIN DIE AND THE KM INITIATIVE LIVE?

The case of both WIN and the KM Initiative illustrate how the creation of OLMs and the appropriate cultural norms can produce productive organizational learning. However, there is a critical difference between the outcomes of the two cases: WIN was allowed to simply die, whereas the KM Initiative was systematically expanded and institutionalized throughout the organization. The immediate explanation for this difference is that the latter received top management support, whereas the former did not. However, this explanation only points to deeper questions: What led management to support one initiative rather than the other? How did management support manifest itself?

In this chapter, we will use the multi-facet model as an analytical tool to help us answer these questions. We will systematically apply the model to these two cases in order to understand what may have led one project to become institutionalized whereas the other was only a fleeting, temporary OLM (see  Table 11.1 ). Furthermore, the analysis will help us see more clearly what it means for learning to beorganizational. In order to get there, the analysis will address the following questions: What was learned? Who learned and to what extent was the learning organizational? How did organizational learning actually take place? How did each of the structural, cultural, psychological, policy, and contextual facets influence learning?

What Was Learned?

Strictly speaking, this first question is not derived from the multi-facet model. However, before analyzing organizational learning, it is important to clarify whether some significant change took place that could be classified as learning. In the case of WIN, the learning involved the introduction of new methods for organizing (self-managing teams) and for overcoming barriers to innovation in change. These methods were based on a well-defined set of concepts and values. The methods were not aimed at any specific set of issues but could be applied generically throughout the organization. The learning in the KM Initiative focused on adding value to the customer by building capacity to systematically create new knowledge from existing, but scattered, knowledge.

Table 11.1 Comparing WIN and the KM Initiative at Hewlett-Packard

Who Learned?

This question is important for distinguishing individual and organizational learning. Through WIN, managers in different HP units developed new skills and capabilities that they then applied, changing the way work was designed. These changes in work design were credited with improving performance on achieving business objectives. The organization learned in the sense that the innovations advocated by the WIN initiators was disseminated and implemented widely in the organization. Butorganizational learning was limited because few, if any, mechanisms were developed for systematically disseminating this knowledge or storing this knowledge in organizational memory. For the most part, dissemination and storage took place informally rather than through formal policies and operating procedures.

In the KM Initiative, on the other hand, both the HPC consultants and the organization learned extensively. The consultants developed a new appreciation for the importance of OL/KM, as well as a complex set of skills and methods for producing it. The organization learned in the sense that these values and methods were institutionalized through an extensive and relatively permanent set of organizational learning mechanisms and policies intentionally designed to reinforce them.

How Did Organizational Learning Actually Take Place?

From the perspective of the multi-facet model, the answer to this question can be found by looking at the organizational learning mechanisms (OLMs) employed by each project. Through OLMs, organizational members identified problems or opportunities, collected relevant information, processed this information, generated new knowledge, and disseminated that knowledge to where it was needed. There is a very clear contrast between the two cases in terms of the OLMs involved.

WIN functioned as a single, though complex, off-line OLM. It was a bit of a hybrid in terms of agency. The introductory workshops were off-line/external agent OLMs in which consultants passed their knowledge on to managers who felt that it would meet their needs. However, in more advanced stages of the WIN process, it became more of an internal agency OLM as managers presented their own cases, provided feedback, and consulted with each other on practice issues. To the extent that networking led to ongoing, informal information sharing and peer consultation among participants, then WIN also constituted an online/internal OLM.

The KM Initiative, on the other hand, involved a wide variety of OLMs. Postproject reviews were off-line/internal OLMs. Project snapshots also involved internal agents but occurred in much closer proximity to task performance. The learning communities, in which consultants shared information informally or through conference calls, were online/internal agent OLMs. Another OLM, the knowledge-mapping team, was set up in order to develop a method for service delivery using organizational learning and knowledge management.

The KM Initiative was particularly noteworthy, however, for the way in which one “generation” of OLMs spawned a whole new set of OLMs to meet various learning and information needs that arose as the project developed. The initial OLM, which could be called the knowledge-mapping team, was basically an off-line, external agent OLM. It developed both a new service delivery method to be used by the entire organization and a workshop for dissemination. The knowledge-mapping team was a temporary OLM and ceased to exist once the new knowledge was created.

The dissemination workshops were themselves off-line OLMs involving both external and internal agents (trainers and participants, respectively). The notable thing about these workshops was that this second-generation OLM functioned as a mechanism for disseminating a third generation of OLM for knowledge production and sharing. Participants in the workshops not only learned a new method for service delivery but also developed plans for tailoring postproject reviews, project snapshots, and learning communities to the specific needs and features of their units. Thus, one generation of OLM gave birth to another. In addition, linking new methods with the design of tailor-made OLMs constituted a combination of top-down and bottom-up learning strategies.

The existence of a wide variety of OLMs does not necessarily ensure that they will function efficiently or maximize the potential learning. A tremendous waste of energy could easily have resulted from multiple OLMs working on the same issues in parallel, by the “not invented here” syndrome, and by the tendency to “reinvent the wheel.” This challenge was met by the creation of the roles of knowledge service manager, knowledge architect manager, solution development knowledge manager, and knowledge consultant. Each of these roles functioned as an off-line/external OLM for linking local OLMs, facilitating information sharing and coordination, and ensuring that local learning was disseminated throughout the organization.

In addition to creating a wide, diverse, and tightly linked network of OLMs, the KM Initiative also created a number of technologies, or tools, for supporting the organizational learning process. These technologies included knowledge mapping, the Web portal, genericizing tools, the Knowledge Master Award, and methods for measuring learning at both individual and unit level. In the WIN case, the only reported technology for supporting learning was action research.

Comparing WIN with the KM Initiative reveals a major difference in the OLMs created for the purpose of learning. WIN was based on one large and complex OLM. It clearly aimed at stimulating learning, and the networking among participants was intended to spawn new initiatives, but WIN did not invest directly in the development of new OLMs. The KM Initiative, on the other hand, made extensive use of a wide variety of OLMs and support technologies.

How Did Cultural Norms and the Psychological Climate Influence the Learning?

As pointed out in  Chapter 3 , OLMs are a necessary, but insufficient, condition for productive organizational learning. Productive learning requires behaviors driven by cultural norms of inquiry, issue orientation, transparency, integrity, and accountability (see  Chapter 3 ). The enactment of these norms depends largely on a psychological climate of safety and commitment to the organization (see  Chapter 4 ).

The available materials on the two cases provide very little descriptive material that would enable us to infer or compare the specific values and the psychological climate that characterized these two projects. Hewlett-Packard is widely recognized for a strong and unique culture. One of the explicit principles of WIN, for example, was creating a “safe environment” (psychological safety) that would encourage the participants to expose their errors as well as their successes (“transparency” and “integrity”). These norms were strongly enough felt for at least one participant in WIN to publicly object when he perceived that presentations were becoming less truthful and personal. There is no concrete evidence about the participants’ commitment to the organization. Commitment could be inferred from the fact that they voluntarily participated in a program with no material reward, but there could be other equally plausible explanations.

Similarly, in the KM Initiative, the importance of creating a climate of psychological safety was espoused in the guidelines for the postproject reviews, which called for an emphasis on learning as opposed to blaming or fault finding. However, these guidelines rightfully belong to the leadership and policy facet (see  Chapter 6 ) and do not say anything about the organizational members themselves. There is little evidence as to whether the HPC consultants experienced psychological safety, how deeply committed they felt to the organization, and whether their behavior in the various OLMs was guided by learning values or not.

How Did Organizational Policy Influence Learning?

The policy facet focuses on the formal and informal steps taken by management to promote organizational learning. The multi-facet model points to three policies that are particularly important for learning: commitment to learning, commitment to the workforce, and tolerance for error.

A look at the policy facet reveals a major difference between the KM Initiative and WIN. The KM Initiative was shaped and supported by a wide range of policies aimed at giving concrete expression to the organization's commitment to learning. For example, management set clear learning goals for the project. Thus, training was linked to specific learning goals and action planning to achieve those goals. Management linked the KM Initiative into a broader change effort by integrating OLM components into changes in project management methodology.

Most important, perhaps, learning was integrated into the organization's performance management system. Indeed, the organization invested effort in developing a new and sophisticated set of criteria and measurement tools for evaluating learning and learning-oriented behaviors at both the individual and unit level. A conscious effort was made to make managers aware of the existence, and the importance, of this new set of performance measurement tools. Finally, rewards and promotion linked directly to the learning exhibited through these tools.

In the case of WIN, on the other hand, no specific policies were developed or implemented to support learning or to disseminate beyond the program itself. To a certain extent, the lack of policy supports was an implicit, if not fully intentional, part of WIN's learning strategy. From the beginning, it defined itself as an alternative to change imposed from above. WIN adhered to such a radically bottom-up approach that it was described as “a bit of a guerilla movement” (Zell, 2001, p. 84). In retrospect, its advocates lamented not trying to cultivate top-level sponsors, which they saw as the root cause of WIN's demise when HP encountered financial difficulties.

Although we can say very little about the psychological climate and the cultural norms actually enacted by participants in both cases, we can identify aspects of HP policy that support learning. “Commitment to the work force” comes through most explicitly in official HP policy. For example, “employee commitment” is one of the explicit corporate objectives (HP history and facts, n.d.):

To help HP employees share in the company's success that they make possible; to provide people with employment opportunities based on performance; to create with them a safe, exciting and inclusive work environment that values their diversity and recognizes individual contributions; and to help them gain a sense of satisfaction and accomplishment from their work.

This objective reflects Hewlett-Packard's strong and unique culture, known as “the HP way.” This culture, which influences both formal policy and informal management behavior, was originally instilled by its two inspirational founders, Bill Hewlett and Dave Packard and assiduously maintained by succeeding corporate leadership (HP history and facts, n.d.; Rogers, 1995a).

The “HP way,” as it was understood at the time that these two cases took place, has been described as follows:

When former employees reminisce about the HP Way, they toss around words like “integrity,” “trust,” and “team.” They tell stories about how Hewlett and Packard regularly roamed the halls, talking with employees about their projects; how employees put on annual skits where they ribbed their bosses, including Bill and Dave; how co-workers were reassigned to new jobs rather than fired; how the company for a time implemented a shortened work week for all employees so certain individuals would not lose their jobs. “Everybody respected everybody else. Management was part of the team,” Cottrell (former head of HP's European division) said. “Something you'll notice is that even retirees talk in terms of ‘we.’” Following the founders’ lead, managers were expected to know their employees and share with them what was going on in the company. (Dong, 2002)

The HP way clearly reflected a commitment to learning in both formal policy and informal management behavior. It mainly emphasized “learning-in” the organization; that is, providing individuals with opportunities for developing their task-related knowledge and skills. However, it also placed importance on information sharing (Dong, 2002; Kotter & Heskett, 1992, pp. 60–61).

It is more difficult to find direct evidence about policies that support a “tolerance for error” in the service of learning. As mentioned earlier, the KM Initiative emphasizing the importance of creating a climate of psychological safety was espoused in the guidelines for the postproject reviews, which called for an emphasis on learning as opposed to blaming or fault finding. HP History and Facts (n.d.) stresses “the importance of developing leaders who coach, relay good news and bad, and give feedback that works” and who “demonstrate self-awareness and a willingness to accept feedback and continuously develop.” These espoused values are intended to encourage managers to be open about their own errors as well of the errors of others and, in doing so, to be a role model for their subordinates.

The evidence clearly indicates that overall HP culture is oriented toward fostering the psychological climate and values conducive to learning. The very emergence of both WIN and the KM Initiative programs, aimed at organizational learning and knowledge management, are reflections of this culture. However, this corporate culture was translated into concrete management policies mainly in the case of the KM Initiative. The contextual facet may help us understand the reasons why one initiative received more management attention than the other.

How Did the Context Influence the Learning?

The contextual facet focuses on factors that are not within management's direct control. These factors include error criticality, environmental uncertainty, task structure, proximity to core mission, and committed leadership. These factors create pressures to engage in learning but also posit conditions that shape and/or constrain learning. Although management may have no control, or no direct control over these factors, it must take them into account and adopt strategies for dealing with them.

Perhaps the biggest difference between WIN and the KM Initiative involved their proximity to the core mission of the organization. The learning in WIN was aimed at transmitting a generic kind of knowledge involving the use of self-managing teams and assimilation of innovation. As important as these tasks were, they were not directly related to the delivery of products or services. WIN was offered to managers and units throughout the organization with no particular focus on a specific task or problem. Essentially, the participants in WIN could apply this method to any issue, regardless of its strategic importance or proximity to the core mission of the organization.

The KM Initiative, on the other hand, was tightly focused on improving HPC's core mission. Indeed, delivering knowledge to customer was the core mission. The ability of the consultants to learn and improve their knowledge was critical for effective performance prior to the KM Initiative. The KM Initiative added an organizational component to this ongoing individual learning process.

A combination of environmental uncertainty and error criticality also created conditions that made the KM Initiative a strategic necessity. The learning was actually stimulated by environmental pressures—that is, customer perceptions that the quality of HPC services depended more on the individual consultant than on the organization as a whole. Management realized that if this perception was not addressed, it could lead to difficulty satisfying customers in the long term. Furthermore, because organizational effectiveness was closely associated with the individual performance of HPC consultants, any failures or errors on their part could be very costly. The success or failure of the entire business rested on continually building this knowledge base and making sure that it was as fully and equally distributed as possible.

There is no evidence to indicate that WIN, on the other hand, was a response to environmental uncertainty or error criticality. Rather, it was driven by the belief of a group of organizational members that they possessed knowledge that could be useful to others. And, as the case demonstrates, they were right. There were no apparent environmental or strategic pressures driving WIN. Given this context, the program succeeded admirably in meeting the needs of internal customers but was treated as expendable by management.

The structure of the HPC consultants’ primary task, delivering knowledge to customers, also easily lent itself to organizational learning and knowledge management. It was relatively easy to bring together the relevant information for knowledge mapping and sharing because this information was held by the consultants themselves. The key was creating OLMs that bring together the right people at the right time. HPC's project structure lent itself to convening both interim and postproject reviews. Another challenge for the KM Initiative was making the tacit knowledge explicit and making sure that there were sufficient means and incentives for widely sharing explicit knowledge. The web of OLMs and technologies that HPC developed for accomplishing fit easily into the structure of the consulting task. In fact, the organizational learning and knowledge management became an integral part of HPC's project management methodology.

The issue of task structure was not really salient in the WIN case because the program focused on transmitting a generic set of management techniques rather than generating learning around a specific task or problem. It was implemented as an independent series of workshops without integration into a particular task arena. Participants were able to choose the specific issues or task areas they wanted to engage through WIN, but they had to design their learning according to the structure and timing offered by WIN. The learning that took place through WIN appeared to integrate well into the participants’ work—at least for those who stayed with it. However, the WIN workshops were distanced from the places of actual task performance and involved significant travel, a factor that led to the program's eventual demise.

How Did Leadership Influence Organizational Learning?

In one sense, WIN was an example of grassroots, informal leadership. In fact, the WIN advocates, based on their ethic of self-management and bottom-up change, did not seek sponsorship or support from upper management. As a result, management did not identify with the program and could not be convinced that its contribution was important enough to override cost-cutting policies.

In the KM Initiative, on the other hand, the general manager of HPC recognized the strategic importance of organizational learning and knowledge management of the program from the very beginning. He made his commitment to learning clear and played a central role in starting, supporting, and institutionalizing the KM Initiative. Furthermore, top management emphasized the importance of fostering leadership at lower levels. Leadership in learning and knowledge management became an important criterion for promotion and the evaluation of manager performance.

 DEMYSTIFYING THE LEARNING ORGANIZATION

The Work Innovation Network and the Knowledge Management Initiative are both impressive examples of organizational learning in Hewlett-Packard. The very existence of these efforts would seem to justify calling HP a “learning organization” in which “people continually expand their capacity to create the results that they truly desire, where new and expansive patterns of thinking are nurtured, where collective aspiration is set free, and where people are continually learning how to learn together” (Senge, 1990, p. 3). However, looking at the learning organization in this way actually obscures the important differences between these two cases.

The difference between WIN and the KM Initiative was not that one represented better or more extensive learning. The difference was that the learning in the KM Initiative was simply more organizational. It was more organizational in the sense that it involved a more extensive system of tightly linked organizational learning mechanisms. Both WIN and the KM Initiative emerged from same organizational cultural context, but the latter was more organizational in the sense that it was grounded in a set of formal organizational policies. Finally, the KM Initiative was more organizational in the sense that it was initiated and supported by the formal leadership of the organization. As a result of these factors, the KM Initiative became institutionalized, increasing the likelihood that learning would be sustained and integrated into the everyday functioning of the organization.

By showing how learning can be made more organizational, the multi-facet model provides a framework for demystifying the learning organization. This framework can guide managers who aspire to building a learning organization by specifying the key factors that they must attend to. It suggests the potential effect of each factor and the steps that can be taken with it.

OLMs represent the basic tools that managers have at their disposal for making learning organizational and structuring it to address specific demands for knowledge. The policy facet points to additional tools that are largely within management's direct control. If designed and deployed appropriately, policies foster the psychological climate and cultural norms conducive to learning. The contextual facet, on the other hand, highlights the demands and constraints that impinge on any organizational learning effort. As illustrated in the analysis of these two cases, the degree to which a manager attends to these contextual factors may have a critical influence on the long-term viability of an organizational learning effort.

One advantage of the multi-facet model is that it provides an alternative to the either-or framing suggested by the learning organization concept. Managers may be inspired by visions of major transformation, but very few of them sit at the level where they can effect such systemwide changes. However, managers can use the multi-facet model to understand whatthey can do to be agents of organizational learning (Friedman, 2002).

The multi-facet model provides managers with a guide for taking realistic and practical steps toward fostering learning from any position in the organization. It suggests concrete actions that can be taken to clarify the demands for learning, identifying and adapting to the constraints, and marshalling the resources within their control. Our experience is that serious inquiry into these questions almost always leads to the discovery of some space for enhancing organizational learning.

The multi-facet model also implies that, given its particular contextual conditions, organizations develop their own unique configurations of learning mechanisms and policies in order to create the structures, psychological climate, and cultural conditions conducive to learning. From a managerial perspective, promoting organizational learning is a process of design: that is, consciously shaping learning-oriented mechanisms, policies, and practices in light of the demands, materials, constraints, and resources at hand (Shani & Docherty, 2003). This design approach is based on the assumption is that there is no single method or formula for enhancing organizational learning. Because of the wide variety and complexity of organizations, learning has to be adapted to an organization's specific needs, characteristics, and circumstances.

10

High-Quality Organizational Learning

Why Do Some After-Action Reviews Work and Others Don't?

A brief report that appeared in the Israeli daily Maariv on 2 July, 1997, raises a question about the quality of learning from after-action reviews:

On 31 June 1996, a foot soldier carrying a wireless set was electrocuted when the set's antennae made contact with an overhead high tension electric cable. This was the sixth accident of this kind in the Israel Defense Forces. A committee appointed to investigate the accident discovered that specific instructions issued based on lessons learned from the previous accidents had not been followed. (p. 4)

The report clearly documents a failure in organizational learning. Five accidents were followed by after-action reviews, yet the same accident occurred again, and with a lethal outcome, on a sixth occasion. The Israel Defense Force is not the only organization that fails, occasionally, to learn from its experience. Why or when do such failures occur, and what can be done to avoid them?

An obvious cause of failures of learning is a low-quality learning process. As we pointed out in  Chapter 2 , the mere existence of organizational learning mechanisms (OLMs) is insufficient to ensure productive learning. In  Chapter 3 , we presented five cultural norms that increase the likelihood of productive learning. In this chapter, we offer a more inclusive list of factors that influence the probability that lessons learned from after-action reviews will be remembered, thereby lowering the likelihood that past errors will be repeated.

The factors were identified in a study that compared two episodes of low-quality and two episodes of high-quality after-action reviews in an elite unit of the Israel Defense Forces. The unit operates a variety of OLMs, the most important of which are after-action reviews. Every combat or training mission is followed by an after-action review, and preparation for operational missions includes a review of relevant lessons learned from similar missions. After-action reviews in the unit vary in quality. Some are considered exemplary by the officers and soldiers who participated in them; others are regarded as a waste of time. The principal criterion on which these judgments are based is the effectiveness of assimilation, namely, the long-term implementation of lessons learned. The episode at the beginning of the chapter is an example of failure according to this criterion.

Using assimilation as a criterion for determining the success of after-action reviews makes sense because long-term implementation of lessons learned prevents repetition of accidents and errors. There are two qualifiers for this criterion. First, the lessons learned have to be valid, that is, effective solutions to the causes of the problem. In addition, they have to be reviewed periodically to test their continued relevance to the organization's mode of operation, as well as to changing environmental conditions.

Aserious limitation of the assimilation criterion is that it requires the passage of time in order to know if a given learning process was successful. To correct this limitation, we decided to study the ways in which the processes that produced effective or ineffective assimilation differed systematically. This way we hoped to complement the outcome criterion of assimilation with criteria that pertain to the after-action reviewprocess.

Owing to its unorthodox and dangerous combat methods, the elite unit under study experiences training accidents. The study concerned four reviews of such accidents. The first step was to ask some 30 officers to identify after-action reviews that represented high- and low-quality reviews and, based on their evaluations, select two high- and two low-quality reviews on which there was consensus regarding their quality. The reviews spanned a period of several years and were sequenced as a low-quality review, two high-quality reviews, and a low-quality review. We interviewed every available officer who participated in each review and analyzed all the documentation that was available on them.

The accidents required more complex after-action reviews than the postflight reviews discussed in  Chapter 8  because of the complexity of the exercises, the number of persons involved, the severity of the outcome, the uncertainties regarding their causes, and the number of required changes. To convey a sense of the two types of reviews, we present an outline of one example of each.

 LOW-QUALITY AFTER-ACTION REVIEW

The review concerned the death of a soldier during a dangerous combat exercise. The review proceeded in four phases: information gathering, drawing of lessons learned, implementation, and assimilation.

Information Gathering

As soon as he was informed that a soldier was missing, the exercise commander convened a debriefing on the ground to assess the situation and begin a search. The unit's commander conducted a more systematic review after the soldier was found dead. Although the information-gathering phase is intended only to ascertain the facts of the accident, the commander drew some key lessons learned during this phase. The corps to which the unit belonged appointed an external investigative committee that four hours later began interviewing officers and soldiers who participated in the exercise. Another external investigation was launched by the military police, as required by the Israel Defense Forces regulations. The phase of information gathering continued for 2 days.

Drawing of Lessons Learned

We could not obtain reliable details on this phase either from the interviewees or from the documents (which we take to indicate its shallowness). The usual format requires the commander of the exercise, the officers in charge, and other participants to provide information on the objectives, planning, and execution of the exercise from their different perspectives and answer the questions of the unit's commander and other officers who conduct the review. At the conclusion of this process, the unit's commander decides on lessons learned regarding training procedures, combat doctrine, and logistics.

Implementation

Three committees were appointed by the unit's commander to translate the lessons leaned into changes in training procedures, combat doctrine, and logistics. The changes were summarized in a document that was presented in a conference attended by the unit's high- and middle-rank officers. This phase continued for 2 months. The overall duration of the review was markedly short for an accident that involved the death of a soldier.

Assimilation

To ensure assimilation, the lessons learned were refreshed in several “safety symposia” in the period following the after-action review. The after-action review of a similar accident some 2 years later (the high-quality process below) revealed that some key relevant lessons learned were never implemented.

 HIGH-QUALITY AFTER-ACTION REVIEW

This episode occurred under the successor of the commander who succeeded the officer who commanded the unit during the low-quality review. The two accidents were similar—a soldier who disappeared during an exercise of the same combat technique was later found dead. The two reviews consisted of the same four phases.

Information Gathering

The deputy commander arrived on the scene as soon as the soldier was declared missing. He debriefed the officers and soldiers to assess the situation and direct the search. Several additional debriefings were held later, both on the grounds and in the commander's office. No conclusions were drawn at this stage. The external investigations by the Military Police and the committee appointed by the corps commander began their work shortly after the accident. This phase continued for 3 days.

Drawing of Lessons Learned

An initial after-action review was held in the unit 4 days after the accident. Its report did not identify the causes of the accident but proposed three hypotheses, one of which was considered the most plausible. In addition, it offered 22 recommendations. Two of these were the same recommendations of the external committee in Episode 1, which had not been assimilated. The commander accepted most of the recommendations (this phase continued for 5 days).

Implementation

Eight days after the accident, the commander appointed five committees to study in depth the conclusions and lessons learned of the initial review relating to different subjects (e.g., training and standard operating procedures). Each committee was headed by the most senior officer knowledgeable on its subject. The committees were instructed to conduct extensive investigations and not to limit themselves to narrow issues relevant to the accident. The five committees were coordinated by a steering committee headed by a senior officer. The committees prioritized their agendas and dealt with subjects relevant to routine operations of the unit first. Proposed lessons learned were tested by the training and operational subunits and the results were fed back to the committees, which changed them if necessary. Detailed instructions for implementation were issued to the relevant subunits and units of the corps. The set of final lessons learned and the review process were documented in detail. This process continued for 3 months.

Assimilation

During the next 2 years, the unit held several symposia to refresh the lessons learned, and their implementation was closely inspected. An investigation 3 years after the accident showed that all the lessons learned were operative.

Why did the two cases produce such different assimilations? An examination of their outlines ( Table 10.1 ) reveals that although they followed the same four phases, the, low- and high-quality reviews differed in their breadth and duration. Finer grained analysis produced seven more specific attributes that could account for their different outcomes:

· The commander's leadership style

· The extent to which the review was systematic

· The review's (low vs. high) place in the unit's agenda

· The productiveness of the internal inquiry

· The effectiveness of the external committee

· The plausibility of the lessons learned

· The effectiveness of the implementation and assimilation processes

Table 10.1 After-Action Review Process in the High- and Low-Quality Cases

These attributes can be used both to guide effective reviews and to evaluate their quality. We now discuss each of them in detail.

Leadership Style

Chapter 6  proposed that leadership is a key influence on the probability that organizational learning will take place and that it will be productive. The comparison between the low- and high-quality reviews showed that this factor is especially important when the conditions for learning are difficult. After-action reviews of fatal accidents take place under trying circumstances. The unit is pressured by higher echelons, the media, and the families of the dead and injured soldiers. There are strong feelings of sorrow, guilt, and fear among those who were directly involved and unit members whose friends have died or are likely to be punished. Everyone worries that results of the after-action review may affect him personally. Two behaviors distinguished between commanders of the high- and low-quality reviews. The first, support, was specific to their circumstances. The other, receptiveness, is relevant to any process of organizational learning.

Support. The commanders of the unit in the high-quality reviews supported their subordinates by being sensitive to their emotional needs and shouldering responsibility for the outcomes of the review. Both were explained by a commander:

My principal concern was to preserve the unit. Fatal accidents can cause units to fall apart, and to preserve them you must take care of their people. The officer in charge of the exercise was a reservist who expected to return home after a short period in service. Instead he faced the prospect of going to jail, possibly for a long period. Such experiences destroy people. At the outset of the review, I told the participants that they should relax because I was responsible for whatever happened in the unit, for better or worse. I tried to impress on them that I trusted them and that they were not alone.

In contrast, the commander during a low-quality review was visibly stressed and left an impression that his only concern was to save his own skin:

In four meetings, he and other officers screamed at four different junior officers…. What disturbed me most was that our commander was willing to take responsibility for success but not for failures. It is impossible to draw [valid] lessons learned when you know that if, god forbid, something happens, you will be left on your own.

Receptivity. The purpose of after-action reviews is to introduce required changes in current practices and procedures. This purpose is not likely to be achieved if the commander is not receptive to the suggestions and opinions of others. The commander may reject or not implement them outright, or signal through his reluctance to listen that making suggestions, to begin with, is a waste of time. Overruling the objections of others, the commander of a low-quality review decided on lessons learned that were subsequently not followed in practice and rescinded by his successor. The same commander rejected suggestions involving new techniques or novel ideas because “We tried it before. It will not work.” As a result, “Instead of going forward with new methods and technologies, the unit went backward. A week after he left, his successor began to adopt the suggestions that were rejected.”

Systematic Design of the Review Process

Other researchers of after-action reviews offered the following guidelines for conducting after-action reviews:

· Invite the right people.

· Appoint a facilitator.

· Revisit the objectives.

· Go through the project (i.e., action) step by step.

· Ask what went well? Find out why these aspects went well and express the learning as advice for the future.

· Ask what could have gone better.

· Ensure that participants leave the meeting with their feelings acknowledged.

· Record the meeting.

These guidelines are appropriate for relatively simple, one-shot reviews, such as the postflight reviews described in Chapter 8 .

The criterion of systematic design is intended for complex reviews such as those discussed in the present chapter. It specifies the assignment of committees to investigate a broad range of issues, manning them with persons that have relevant professional and organizational authority, providing them with sufficient resources, appointing a steering committee to coordinate their work, and monitoring their progress periodically through meetings that critique and integrate their work. High-quality reviews were designed according to this prescription:

Work was divided among several committees. One analyzed the training program. Other committees analyzed operational plans (to find if the training program was compatible with them), the unit's combat doctrine, and so on and so forth. The committees were headed by senior officers and composed of senior and middle-rank officers with the best expertise on their subject matter in the unit. Each committee had a detailed plan with specific goals, subgoals, and a timetable. All the committees met every 2 days to present their findings, discuss them, and decide on how to proceed.

Compare this with the testimony of a participant in a low-quality review who said that “I do not remember a real process. The most serious debriefing was conducted on the grounds about half an hour after the accident, and that was it.”

This quote points to another important principle of systematic design: Separate information gathering from making conclusions and the drawing of lessons leaned. The separation is important for conducting a thorough information search, basing lessons learned on all potentially available information, and preventing commanders from committing themselves to premature conclusions that they will have difficulty reversing. The quote shows that the commander of the low-quality review violated this principle. Lessons leaned were drawn at an early stage of the after-action review, the stage which should have been devoted to gathering information. In contrast, the commander of the high-quality review took deliberate steps to separate the two phases. At the beginning of a preliminary debriefing, he instructed participants that the objective was to gather as much information as possible soon after the accident and that he would make no conclusions based on this debriefing. An added benefit of his clarification was that “it helped people to talk freely without worrying about the consequences of what they were saying.”

The Place of the Review in the Unit's Agenda

In  Chapter 5 , we claimed that when the learning effort is closer or more relevant to the core mission, there is a higher likelihood that it will receive resources and attention, catch on, and succeed. Safeguarding soldiers’ lives is high on any fighting unit's agenda. Nevertheless, the high- and low-quality reviews differed in the extent to which participants remembered them as central or marginal in the unit's ongoing operation. The difference can be attributed to the attention that the different commanders paid to the review process. One high-quality review was actually headed by the commander himself. He was deeply involved in all its activities and saw to it that the reviewand its outcomes were presented to the unit's senior officers, who were therefore familiar with its details. Another indication of a review's place in the unit's agenda was the time that was devoted to it relative to other tasks. A participant in a high-quality review remembered the following:

We thought that the review was of high quality and so felt obliged, professionally and morally, to participate. All of us who participated in the review except for one officer were simultaneously engaged in other missions—the unit was particularly busy at the time with operational missions as well as maintaining its training operations. I was involved in the planning of an operation that took a long time. Although I did not downplay the importance of the review, I finally had to spend more time on that [other] operation.

The officer's efforts to participate in the review in spite of competing operational duties is a clear indication of its high place in the unit's agenda. Significantly, he and other officers felt morally obliged to participate because they perceived the review as a serious investigation into the death of a comrade.

Another indication of the review's place in the unit's agenda is participants’ ability to remember its details. Participants in high-quality reviews remembered them in detail; those who participated in low-quality reviews had a hard time doing that:

In my opinion, the process was a complete failure. I arrived to the unit 2 or 3 weeks after the accident, and they were already busy with the implementation of lessons learned. The accident could have happened a year rather than 2 months before I arrived. The review did not receive any special attention.

The place of the review in the unit's agenda sends a message about the importance of its lessons learned, which clearly affects the probability that these will be remembered and implemented.

The Productiveness of the Internal Inquiry

This criterion pertains to the learning process itself: To what extent do the processes of gathering and analyzing information by the unit's members enable them to draw valid conclusions and lessons learned? According to the multi-facet model, a positive answer to this question depends on two factors: (1) the extent to which the culture in the after-action reviews conforms to the learning norms and (2) the extent to which participants feel psychologically safe during the review process. We did not find evidence for accountability in either low-or high-quality reviews, and participants in both types of reviews exhibited integrity. The latter was most clearly observed in the reports of the external committees in both types of reviews, which included candid admissions of errors by officers who testified to the committees. For example, the officer who supervised one of the exercises admitted that he had not reviewed the safety regulations before the exercise (as required) because “We do this exercise routinely.” Another supervising officer admitted that he neglected to check the area in which the soldier was found dead. Similar to the pilots and navigators in Chapter 8 , members of this unit attributed the norm of integrity that was ingrained in it to the long process of socialization of its members:

Right from basic training in the unit, we are told to report after any exercise on what we did [correctly] or failed to do with an emphasis on the latter…. Reporting truthfully is a norm that is part of the culture of the unit. Compared to other organizations with which I am familiar, people here are honest, and you can obtain from them something that approximates the [objective] truth.

Our analysis did show consistent differences between the high-and low-quality reviews in terms of inquiry, transparency, and psychological safety:

Inquiry. The following quote from a participant in a low-quality review demonstrates lacking inquiry:

Although my roommate OKed the exercise, no one came to talk with him…. The debriefing on the ground did not find anything unusual. The commander decided on some minor changes and that was it. I do not remember any follow-up discussions. During the internal review, we were hardly asked two and a half questions on the accident, and most attention was paid to the injury of a soldier during the search.

Lacking inquiry is shown in the above quote through the narrow and unfocused information search, the fact that no attempt was made to go beyond the obvious (“anything unusual”), and the making of quick decisions without serious analysis. Not surprisingly, the status quo was left essentially unchanged, in spite of the fact that a soldier had been killed. Compare this with the manifestation of inquiry in a participant's description of how interim conclusions served to broaden and deepen the analysis in high-quality reviews:

During an after-action review, you try to map all the factors that may have determined the outcome. You begin from the present and go backward looking at the training course, the operational process, beliefs, behaviors, and then—based on common sense, and debriefs, many debriefings—you focus and come to a conclusion. Now this is your understanding of the accident, but different people may come to different understandings from different perspectives. That is why it is important to take into account the information and conclusions of other committees and the external investigation. At the end, you come to a reasoned conclusion based on your analysis and that of others in relation to training, safety, combat doctrine, and so on.

Transparency. Manifestations of transparency were usually associated with those of integrity, inasmuch as both entailed disclosure of error. Here is how one participant in a high-quality review described how the two values were manifested in the learning process:

The unit is generally characterized by openness [i.e., transparency] and truthfulness [i.e., integrity]. I do not think that anyone tried to hide something or to force something on others. Everyone said his mind. Other people could be angry or disagree with him, but he was heard. I attribute this to the spirit of the unit, which is influenced by the commander as well as the way we debrief right from basic training, particularly the way in which errors are dealt with in them.

The following quote underscores the subtle distinction between integrity and transparency. Integrity is admitting error to yourself or others when it is discovered (as opposed to becoming defensive), whereas transparency is allowing others to observe your actions so that they provide you with valid feedback on them:

There was openness at least to hear what people had to say. There were different committees, and people came to say what they knew and thought. In the final analysis, the intention is not to cover up or falsify what has happened. People truly wish that accidents will not be repeated. Listening to people with different perspectives allows you to come to different conclusions.

Psychological safety. Psychological safety is important for productive learning. In its absence, participants become occupied with defending themselves instead of trying to understand what went wrong and how to improve for the future. Commanders’ sensitivity and support played a decisive role in the presence or absence of psychological safety in the reviews that we studied. One officer described how the commander's leadership style and lack of sensitivity lowered his psychological safety this way:

The atmosphere during the review was uncomfortable. It had an effect on me and must have affected the others. An officer was late, and the commander let him know that in his customary threatening style. A soldier was seriously injured when he fired a flare during the search. As the subject came up in the review, I remember saying that I did not order to fire the flare. I must have been anxious because I would not have mentioned it otherwise. One officer ran to call a medic instead of helping the injured soldier. My impression during the incident was that he was shocked, but in the review he said that he was acting deliberately. In my opinion, he was trying to make his actions look better.

The Effectiveness of the External Committee

The rationale for appointing external committees is clear: Enlist relevant expertise from outside the unit and correct potential biases in the work of internal reviews with a review by persons who can study it objectively. To be effective, external committees should possess expertise and objectivity and be perceived to posses them. The credibility of the external committee in a low-quality review was damaged irreparably because it included an officer who was on leave from the unit but scheduled to return as its commander. The credibility of another was destroyed because the accepted opinion in the unit was that its members lacked the required expertise to understand the exercise. Neither committee managed to affect the lessons learned that were decided on by the respective commanders. In contrast, participants in one high-quality review were impressed, not just with the objectivity and expertise of the external committee but by the time and effort that they put into their work. Ninety percent of the recommendations of this committee were accepted by the unit's commander.

The Plausibility of Lessons Learned

There are two methods to ensure that lessons learned will be implemented: Apply external force and generate internal commitment. Applying external force relies on monitoring the implementation processes and punishing those who are caught disobeying them. Generating internal commitment depends on convincing the unit's members that the review has unearthed the causes of the accident and that the lessons learned are plausible, that they provide effective and applicable remedies to these causes. The plausibility of lessons learned was often judged by how they were drawn. Thus, a participant in a low-quality review reported that its lessons learned were perceived as “panicky” because the commander decided on them at an early stage without waiting for the conclusion of the full review. In contrast, a participant in a high-quality review thought that “the review produced the correct lessons learned because it considered all the aspects of the accident and elicited a wide range of opinions from different persons.”

Judgments of plausibility were sometimes determined by the compatibility of the lesson learned with the unit's culture. A key lesson learned in the low-quality accident was hotly contested as impractical. The underlying issue was a tension between safety and risk taking that was ignored in the review. The informal norm in the unit was that model fighters take risks to accomplish their missions. Ignoring this norm, the change in procedure decided on in the review limited soldiers’ freedom of movement to increase their safety. The change was contested as “impractical” and was never implemented in spite of the risks involved. A high-quality review recognized the conflict between formal procedures that stress safety and the unit's culture and introduced new safety regulations in conjunction with a campaign that educated soldiers on the risks involved and that promoted the idea that model fighters avoid unnecessary risks.

The Effectiveness of the Implementation and Assimilation Processes

Implementation is often mentioned as the bottom-line criterion for high-quality organizational learning. Because of our interest in the prevention of repeated errors in the long term, we extend this criterion to include assimilation. The effectiveness of implementation and assimilation is partly dependent, of course, on satisfying the previous six criteria of the review process. As we'll show, these form a causal chain that extends from the commander's leadership style to successful assimilation. In addition, the low- and high-quality reviews differed in terms of specific actions that were taken in the latter and neglected in the former to ensure that lessons learned were implemented and assimilated. These included (a) detailed documentation of the lessons learned; (b) dissemination of the lessons learned to all relevant units with explanations of their underlying rationale, changes in mode of operation that they entailed, and risks that were involved in disobeying them; (c) close supervision and monitoring of implementation of the lessons learned; and (d) regular “refresher” conferences, in which the above are repeatedly illustrated with “war stories” on the accidents (such as the story of the original accident) that the lessons learned were intended to prevent.

As we have just mentioned, these six criteria form a causal chain with the commander's leadership style to determine the review's success or failure. The commander affected all the factors that differentiated between the low- and high-quality reviews, except for the external committee. He determined the design of the review process, and his leadership style influenced the productiveness of the learning process. Proper design had two outcomes: First, systematic data collection and analysis increased the probability that the causes of the accident would be identified correctly and that the correct lessons learned would be drawn. Second, involving as many unit members in the review as possible by assigning them to various subcommittees and testing the plausibility of preliminary lessons learned improved understanding of the rationale of new procedures and regulations and commitment to their implementation. The commander also determined the place of the review in the unit's agenda, which, in turn, determined the amount of time that they could devote to the process and sent a message to the unit at large regarding the importance of the review and the implementation of its lessons learned.

The receptivity of the commander and the latitude that he left to the committees to pursue their investigations were key factors in promoting the value of inquiry. The commander's support shielded members from fear and promoted integrity, transparency, and issue orientation.

The seven factors identified in this chapter are not limited to determining the quality of after-action reviews in military units. The extent to which the review is designed and carried out systematically, the leadership style, receptiveness and support of higher echelons, the productiveness of the learning process, the expertise and objectivity of external resources, the plausibility of the lessons learned, and the effectiveness of implementation will affect the success of assimilating lessons learned in every type of organization. For mangers who wish to maximize assimilation, the chapter has a general lesson beyond ensuring that the seven criteria are achieved: To succeed, they must see that the lessons learned win both the hearts and minds of the organization members. It is not sufficient that lessons learned be valid in some objective sense or in the opinions of experts. Unless the organization members judge them to be valid and important, they will not make the necessary effort to change their current patterns of behavior. That is why it is important to involve as many members as possible in the process, to let those who are directly involved participate in a meaningful way, and to persuade the remaining members of the organization of the importance of lessons learned through extensive communication and the symbolic value of serious support and resources.

_________________

AUTHOR’S NOTE: This chapter was written in cooperation with Neta HorinNaot and is based on Norin-Haot.