What We Care For
Evolutionary Services
Collaboration Campus
Knowledge Garden
ImageMap - turn on images!!!
About us
What's New

Communities of Practice
Learning Is Social. Training Is Irrelevant?

[Communities of Practice] New ideas about social learning may hold the key to understanding how people really learn at work.

Cover Story By DAVID STAMP
David Stamps is associate editor of TRAINING Magazine. (DStamps@lakewoodpub.com)

As the end of a century draws near, the air is sure to thicken with prophecies about the future, including the future of work. The "knowledge worker" will be a favorite topic of management soothsaying; so will "the learning organization."

But for a clear vision of how learning should happen in a business setting, you need only talk to Dede Miller, a customer service representative for Xerox Corp. For two years she was treated to a tantalizing glimpse of the future. Today she finds herself wishing she could go back to it.

To know her story is to understand just how wide a gap still separates learning theory and common training practice - and how hard it will be to apply new approaches to workplace training, even those that make incontestable sense.

In 1994 Xerox began plotting an experiment in operations at its Louisville, TX, customer service center. The object was to see if three separate customer service departments could be integrated into a single unit. From the customers' standpoint, the change was eminently desirable; Xerox's customers had long complained about being shunted from one department to another when they called seeking assistance.

Despite the obvious benefit of reducing customer discontent, consolidation presented some tough questions: Would the advantages of an integrated service department offset the costs involved? People would need to be relocated; databases would have to be reconfigured to provide access to customer information residing in separate service departments.

And some in the company wondered if employees were up to the challenge of learning new jobs within the time constraints of the two-year consolidation experiment. Could account administrators, for instance, learn to do the job of the service reps who scheduled repairs? Could service reps learn the subtle tricks of the sales staff, whose job it was to sell supplies such as paper and toner?

Indeed, learning hurdles appeared to pose the most daunting challenge to Xerox's Integrated Customer Service (ICS) experiment. Back in Leesburg, VA, the corporate training department had painstakingly crafted separate curricula - each involving weeks or months of classroom indoctrination-for each aspect of the various customer service jobs. To combine these curricula and teach new ICS staffers all they'd need to know to perform in an integrated setting would take no fewer than 52 weeks of classroom training for each person. Or so the training professionals calculated.

Researchers at the Institute for Research on Learning proffered a different idea, however. IRL, based in Menlo Park, CA, traffics in a rich mixture of consulting, research, and something akin to corporate espionage. The preferred modus operandi of IRL researchers is to show up one day at a company and then hang around until they fade into the scenery - at which point they pull out their notebooks. Using the ethnographic techniques of anthropologists, IRL agents endeavor to discern how work really gets done in a given setting. As a research organization, IRL uses this information to refine its theories about workplace learning. As a consultancy, it advises clients how they can design learning that reflects the true nature of how employees do their jobs.

Jack and Marilyn Whalen, the IRL researchers contracted by Xerox to advise it on the ICS project, suggested that training need not take a full year; that it could, in fact, be dramatically shortened. How? By moving the service reps out of their isolated cubicles and bringing them together in shared work spaces, where a group of six or seven ICS staffers would be in constant contact with one another. In this communal environment, the workers would teach each other how to do their respective jobs; sales reps would share what they knew about selling, service reps what they knew about service and so on. And one other thing . . . the ICS workers would take customer calls from day one, putting into practice what they learned as soon as they learned it.

The response to this proposal from the corporate training unit back in Leesburg was a long, anguished wail that could be heard all the way to Texas. But Cheryl Thomas, the manager tapped to head up the ICS project, decided to seek a second opinion - actually, 30 second opinions. She asked the employees who'd been selected to be the ICS guinea pigs what they thought of the idea. To the question of whether or not workers could teach each other, the answer she heard was, "Why not? It's what we do already."

What Thomas heard from the trenches, in fact, did little to elevate her estimation of the training department's role in the learning process at Xerox. She was astonished to learn just how removed from the sphere of real work was the prolonged methodology for developing training classes, a process that unfolded along these lines: First came a flyover visit to determine training requirements. Those requirements were handed off to designers back in Leesburg, who designed a curriculum. Months later, training-delivery specialists parachuted onto the scene armed with a detailed script.

But these scripts were next to useless, according to the workers. They complained to Thomas that they heard way too much about tasks they never or seldom performed, and way too little about some of the most crucial aspects of the job. Moreover, there was never an opportunity to practice what they were taught until three or four weeks of classroom training ground to a blessed halt. By the time workers were actually back on the job, they'd forgotten most of what they'd been told. To learn billing and credit procedures, for instance, trainees endured 11 weeks of nonstop classroom lectures before taking their first customer call.

Worst of all, none of the talking heads had ever set foot in a customer service setting. "They tell us how work gets done based on what someone else has told them," the workers told Thomas. "When we finally get back to the job, co-workers have to explain to us how things really get done."

This common phenomenon has been discussed for years in training circles, but always with the problem attributed to "change resistance" by bullheaded workers and supervisors. As far as Thomas was concerned, the workers' complaints signaled a defect in the training, not in the trainees or their peers.

Over the initial objections of the training department (some trainers assigned to the ICS project later became staunch advocates of the new approach), IRLS recommendations were put into place. Not only did the ICS workers prove themselves adept at teaching their jobs to each other, by their own accounts they were exhilarated by the challenge of doing something new and different. "It was the best sort of team-building I'd ever seen," says Rick Hawkins, who came to the ICS project from account administration. "It forced us to rely on each other daily." In learning new skills for the ICS pilot, Hawkins estimates he spent no more than eight hours away from the job, listening to co-workers in an informal classroom setting; the rest he picked up while on the work floor.

Dede Miller describes the experience as nonstop learning. "We shared information with each other all the time. Even when we weren't asking for help, we were learning because we could hear other people in our work group on the phone with customers, and we'd pick up tips about how they handled certain kinds of calls."

Today Miller sits alone in a cubicle surrounded by other workers, each in their separate space, each doing exactly the same job. Though the ICS pilot proved that workers can learn by doing and by teaching each other, the business case for ICS was less compelling. At a time when money was tight at Xerox, the expense involved in integrating service departments was not insignificant. Following the retirement of Curt Stiles, the operations vice president who had been one of ICS's most ardent champions, a decision was made not to extend the project beyond its original two-year pilot phase, which ended last July. (Xerox has, however, taken steps to instill the project's learning precepts into other parts of the company - including, of all places, the corporate training department. More on that later.)

"None of us wanted to go back to our old jobs," sighs Miller. "For the first time ever we were able to solve all of a customer's problems without having to pass them off to someone else. That made us and the customer both very happy."

Which makes it all the more frustrating now when Miller has to transfer a customer to another department. "I know I could solve their problem if I could just get access to the information I had back in ICS," she says. "I wish people could have understood why we were so passionate about making it a success and sticking with it. I guess we were just too far ahead of the curve on this one."

In fairness to Xerox, the document company has been no slouch when it comes to experimenting with new ideas about workplace learning. Its own Palo Alto Research Center (PARC) has been a breeding ground for some of the very concepts that the folks at IRL and a handful of other innovative consulting firms are trying to apply. Still, Miller is right about being ahead of the curve; these emergent ideas run so contrary to traditional approaches to worker training that even forward-looking companies like Xerox can expect to encounter slow going when they attempt to put them into practice.

At the core of the new thinking is the notion that work and learning are social activities (see "Practice, Learning, Meaning, Identity"). As people work together, they not only learn from doing, they develop a shared sense of what has to happen to get the job done. They develop a common way of thinking and talking about their work. Eventually they come to share a sort of mutual identity - a single understanding of who they are and what their relationship to the larger organization is. It is in these groups where some of the most valuable and most innovative work-related learning occurs.

A few years ago a formal-sounding name was coined to describe such groups: "communities of practice." But these naturally occurring communities are anything but formal; they are so informal as to often be nearly invisible. If you want to find out how work really gets done in a company, say adherents of the communities-of-practice persuasion, don't look at the organizational chart; look for the hidden associations among workers.

Though communities of practice exist by the thousands - perhaps millions - in businesses large and small, our work culture does little to encourage them. In the name of privacy and productivity, workers are isolated in cubicles and encouraged to communicate via e-mail (the word interactive has come to mean one person watching a computer screen). In the name of consistency, companies pen elaborate instructions for every possible process and procedure.

But follow the corporate anthropologists down onto the work floor or into the office labyrinth, and what you'll find are workers craning their necks like giant birds to peer over a cube wall and carry on a conversation with their neighbor. And if you listen in on the conversation . . . well, yes, they might be talking about the game-winning goal that Mario Lemieux scored against the Blackhawks last night, but it might also be that they are trying to figure out how to circumvent the official processes and procedures in a way that will allow them to solve an unusual problem that's cropped up with an important customer.

Learning and innovation happen on the job every day. That much we know. The question is, can the ideas bubbling up around communities of practice be used to map a strategy for moving away from the artificial "training curricula" we have today to "learning curricula" that reflect the way work - and learning - really happens?

Lest all of this be dismissed as some malevolent attack by outsiders on the training profession, it should be noted that the profession's own most revered gurus have for years given voice to ideas and criticisms virtually identical to many of the points the communities-of-practice newcomers are trying to make:

>From Malcolm Knowles came the lesson that much of the expertise one seeks to develop in a group of adult learners usually resides in some or all of the group members themselves.

>From Thomas Gilbert came the admonition that the way to find out what's worth teaching is to study master performers, and that the only way to discover the behaviors that really distinguish master performers from ordinary ones is to observe the masters directly while they work. (No, he didn't call this anthropology.)

>From Bob Mager came the demand that training be lean and elegant, focused relentlessly on the skills people genuinely need to acquire in order to perform the job, and not on extraneous material that some trainer or manager believes would be "good for them."

As for the nagging reminder that the way people actually acquire skills is by practicing them, not by being lectured at . . . ever hear that one at a training conference?

But such bedrock training and performance theories have always been honored more in the breach than in the observance. As the ICS project at Xerox illustrated, these ideas routinely fail to translate into practice within actual corporate training departments.

Discussion around communities of practice will undoubtedly get muddied once trainers and management consultants weigh into the debate. For now, however, it's worth keeping in mind that the core principles of this nascent perspective are profoundly simple, and they reflect something many of us know in our bones to be true:

  • Learning is social.
  • Learning happens on the job.

Though classroom instruction may sometimes be necessary, we should be leery of any training that fails to keep learning as close to practice as possible.

If the principles underlying the communities-of-practice thinking are simple, what can or should be done with them is another matter altogether. Even if the only thing you want to do is to observe the communities in place within your company, you first have to identify them, and that can be a challenge in its own right. Communities of practice may exist within departments, but they are just as likely to cross departmental boundaries. Don't be terribly disappointed if, when you find them, they don't coincide with the self-directed work teams you rolled out with great fanfare a couple of years ago.

Bosses can be surprisingly blind to communities of practice flourishing under their noses, observes Peter Hillen, a partner with Congruity Corp., a consulting firm in Los Altos, CA. "A manager will say something like 'I see you spending a lot of time with the guys in the sales department. I hope that's not taking time away from your work.'

"What the manager doesn't realize," says Hillen, "is that the guys in the sales department are helping him do his work."

Some companies confuse communities of practice with competencies and go looking for them in hopes of cataloging skill sets and maybe even enshrining those skills into some sort of corporate knowledge base. It is true enough that knowledge is the coin of the realm within communities of practice; moreover, one has to be able to give as well as take knowledge in order to remain a member in good standing. But the knowledge that gets passed around in these communities is not limited to the sort of explicit information that can be cataloged or computerized or bullet-pointed in a training curriculum. Quite often it takes the form of implicit, or tacit, knowledge.

In a December 1993 article in the Journal of Management Inquiry, researchers Scott Cook and Dvora Yanow described how tacit knowledge is employed at three small workshops in the Boston area to produce some of the finest flutes in the world. Since the physical dimensions and tolerances of the flutes have never been explicitly spelled out, craftsmen rely on imprecise statements like "It doesn't look right," or "It doesn't feel right."

"Yet the extremely precise standards of the instruments, on which the flute's ultimate style and quality depend, have been maintained through these sorts of individual and mutual judgments of hand and eye," the authors noted.

If that sounds too removed from the work world that most of us know, there are examples closer to home. IRL's researchers have observed workers at Xerox who reportedly can tell by its feel in their fingers which type of toner will work best with certain types of paper . . . but only by its feel. Congruity's Hillen claims to know of a company in Silicon Valley where a $6 million piece of equipment used to make semiconductors works less well when certain members of the manufacturing team are absent, though neither the company nor the team members can identify the exact reason why.

Of course, the $6 million question surrounding communities of practice turns on whether or not these informal groups can be created - or, to use a favorite management term, implemented.

Virtually everyone who has studied them agrees that communities of practice cannot be created out of the blue by management fiat; they form of their own accord, whether management tries to encourage them or hinder them.

That consensus of expert opinion has not deterred some companies from attempting to decree them into existence, however. Etienne Wenger, a senior research scientist at IRL, is not sure what to make of the occasional e-mail he receives informing him that the sender has "just implemented communities of practice." Apparently Wenger is supposed to be pleased by the news, since it was he who coined the term communities of practice in a 1991 book Situated Learning (Cambridge University Press), co-authored with Jean Lave.

So far, Wenger has managed to maintain an air of forbearance. "They haven't implemented anything," he says. "Communities of practice are already there; they've existed as long as humans have been working together."

A community of practice is not a magic solution that companies should try to create at all costs, says Wenger. In fact, communities of practice can occasionally be the cause of problems instead of solutions. Outside work life, support groups such as Alcoholics Anonymous are an example of a community of practice. But strictly speaking, a street gang also meets the criteria, observes Wenger.

To concentrate on communities as the end product leads you in the wrong direction, says Wenger. "Communities are a way of thinking about how work gets done, a language to talk about something we've always known - that people learn on the job and they learn from working together. The idea is to take that understanding and develop learning that reflects the practices in those communities."

That said, it's probably just a matter of time before we see two-day seminars on building communities of practice, in which participants are sent home with three-ring binders and detailed blueprints. "These sorts of grassroots movements have a history of getting co-opted," concedes Wenger.

For now, the notion that communities of practice can be willed into existence finds its principal adherents in the buzz surrounding the Internet and corporate intranets, where collaborative groupware supposedly will enable knowledge workers to "hotlink" their bright ideas from cubicle to cubicle.

"One of the great myths of the Information Age is the idea that technology will create collaboration," says Brook Manville, the former chief information officer at McKinsey & Co. who now directs that firm's research in the field of knowledge management. "For years the IT people have rolled out technology thinking that communities will cluster around it," observes Manville.

It doesn't happen that way. Common work issues and a desire to learn from one another are the drivers behind these communities, not technology.

To put it another way, "No one has yet invented a technology that replaces a pitcher of beer," says Congruity's Hillen. Hillen has nothing against technology; his firm works with Silicon Valley companies such as National Semiconductor to help them identify and foster the growth of some naturally occurring communities of practice. "The Web and the Internet can facilitate collaboration, but they don't create it," says Hillen. "You need people together in a room where they can see each other before you get the kind of rapport that leads to real collaboration."

An unproven (and perhaps unprovable) conjecture circulating through the communities-of-practice fraternity these days speaks to that very notion: that face-to-face contact is a condition for true collaboration. The theory purports to explain why, at a time when so much fevered activity centers around distance learning, videoconferencing and the Internet, the number of business miles traveled each year continues to rise. "The speculation is that once people have had an opportunity to interact - so to speak - from a distance, it's only a matter of time before they want to meet in person," says Susan Stucky, IRL's associate director. "I just wish there was a way to prove the idea," she adds.

While it's generally agreed that communities of practice cannot be commanded by management dictum or jump-started by technology, it is very easy to destroy them by meddling - even when the meddling is a well-intentioned effort to nurture them.

One of the worst things management can do, says Hillen, is to "reward" an informal work group by giving it a formal job to do. "Open management's eyes to the fact that a certain group of workers has developed collaborative links to others in a company, and the response is likely to be 'Hey, here's a resource we didn't know we had,' says Hillen. "The community of practice needs to do the work it thinks it needs to do, not the work some guy in a suit tells it to do."

A few companies have found that the best way to foster embryonic communities is to stand back and let them grow. One of the best-documented cases of hands-off management occurred at National Semiconductor a couple of years ago, when word drifted up to management that product groups across the company were seeking out the advice of a certain informal group of design engineers who had gained a reputation for doing superior reviews of new chip designs. Management was astute enough not to try to force the group to write up its rules or to create a computerized library of chip designs that others could copy. The only way to spread this specialized knowledge throughout the company was to let other engineers interact with it, to become members of the community if they could.

IRL's Wenger likes to point out that all communities of practice are local. That doesn't mean there aren't communities whose members are geographically dispersed; it means that every community of practice takes a parochial view of the organization. That's why communities can be the source of problems as well as solutions. Their local viewpoint may keep them from understanding the needs of others in the company. "No practice has the full picture," says Wenger, "not even the practice of management."

Communities are not isolated; they must interact with others. And it is in their interactions across community boundaries where troubles often flare up. The worst problems occur when some external force (read management) tries to force communication across boundaries.

McKinsey's Manville describes a classic, and not uncommon, scenario: A company observes that one of its plants seems to outperform others. To ensure that a newly built plant replicates this same efficiency, a team of industrial engineers is deployed to codify every process to the nth degree. The problem is, the industrial engineers may not be fluent in the language of the workers who actually make the plant such a model of success.

The codified processes then get handed off to instructional designers whose job it is to build a training curriculum around them. Now you've compounded the translation problems because trainers speak a language that is altogether different from that of either the workers or the engineers. In the end, management is left wondering why the new plant doesn't match the productivity of the one it was supposed to replicate.

"The smart thing would be to send workers from the first plant to train the new workers," says Manville. "Knowledge transfer will be much smoother between workers who speak a common language." On top of that, you might even be sowing the seeds of a future community; workers at the new plant now know someone they can call up and talk to when questions arise.

One needs only to look at Xerox's ICS project to see another example of communication that failed to transcend community boundaries. After hearing from Xerox's customer service employees about the perceived shortcomings of the corporate training department, Cheryl Thomas tried to make it as clear as possible to Xerox trainers that the old approach to instructional design would have to change.

At first, trainers resisted. "That's not how we work," they told her. Eventually they agreed to create a curriculum that interleaved classroom training with on-the-job practice. But this first effort was an utter failure. The trainers took notes and made observations and returned to their Leesburg laboratory. When they unveiled the new curriculum, it was still top-heavy with classroom instruction, even though the lectures were now punctuated with on-the-job practice sessions. The employees complained to Thomas that nothing had changed; it was the same old approach, with only slight modifications, as far as they could see.

"We had to throw away the first attempt and start all over again," says Thomas. This time, she told the instructional designers they wouldn't be going back to the lab; they were going to camp out on the job with the ICS staff. They were going to observe how the work got done. And they would develop the curriculum pieces - small pieces - for those discrete parts of the learning that, the workers felt, couldn't take place solely on the job and would need some sort of formal discussion.

That proved a bigger paradigm shift than some of the training designers could make, and those designers were taken off the project. "We were lucky," says Thomas. "Of the six designers, three got the message. In fact, they ended up saying, 'You know, I always felt that it made more sense to do it this way.' "

"It's not that we, as trainers, haven't always known that it makes sense to do it this way," says Carol Ivory, a program manager in Xerox's training department. "We know that people don't learn everything they need to know in the classroom." But managers want something that is neat and clean and easy to schedule, so that's how training gets designed - around set chunks of classroom time. Tell the average manager at Xerox that workers will be cutting into their work time to train each other and you'll get a lot of resistance, says Ivory.

That's not to say there wasn't some resistance from trainers to that very same idea, especially from classroom presenters, who felt most threatened by the new approach, Ivory concedes. "As trainers, we think we need to control every aspect of the trainee's activity. It was enlightening to see the ICS workers take off on their own and come up with some very creative ways of teaching each other."

Though the ICS project was not extended beyond its initial pilot, one unexpected development grew out of it. Shortly after the project ended, management asked Thomas to spearhead an effort to push some of the ICS approaches to instructional design into the corporate training department. The goal would be to revamp the design process to involve workers and managers more directly, and to create training products that used more on-the-job practice and less classroom time.

"I'd never seen training anywhere on my career path," says Thomas. "All I knew about instructional design was what I'd seen come out of it, and I didn't like it." Did management realize what they were asking? Thomas wondered. Did they realize how big a change this would entail? Three months later, Xerox came to Thomas with a strategy document and promises of management support. "If you can recognize and leverage the idea of naturally existing communities of practice, training gets better because you're not fighting human nature," says Ralph Volpe, manager of Xerox's Education and Learning Unit in Rochester, NY. "Learning becomes more efficient and less expensive, and people like it better when you involve them in the development."

Or, as Thomas puts it: "You only need to have seen the difference between the way employees operated in the ICS project compared with how they did their jobs before to recognize that this is the direction we need to be headed."

Back in Louisville, TX, Dede Miller takes consolation from the news that the company is headed in the direction of the sort of nonstop learning she so enjoyed in the ICS experiment. Miller herself seems to be headed in the opposite direction. Beginning last month she was scheduled to be enrolled in the credit and billing curriculum for customer account administration. That, you may recall, is the 11-week marathon classroom event.

Though pleased to be gaining new skills, Miller can't disguise how little she relishes the prospect of 11 weeks in a classroom. "I like the show-me-how approach," she says. "Let me get my hands on the PC and let me learn. Three months in a classroom isn't my personal learning style."

Is it anyone's?

Community of Practice
Definitions | Issues | Best Practices | Recommended Reading | Resources for Further Learning

Back to the top

© Copyright, 2001, Community Intelligence Labs

Last updated on 02/26/00
CoIL webmaster