IT ALL BEGAN ON LABOR DAY WEEKend in 1982. Allan A. Kennedy was sitting in a low beach chair on the shore in front of his cottage on Cape Cod. Next to him was his friend and fellow consultant Tony Merlo. As they relaxed there, watching the sailboats drift across Cape Cod Bay, drinking beer, and listening to a Red Sox game on the radio, Kennedy turned to Merlo and, with the majestic eloquence suited to great undertakings, said: "Gee, Tony, you know, we ought to start some kind of business together."

This identical thought has, of course, passed between countless friends ever since the discovery of profit margins. Coming from most people, it would have fallen into the general category of loose talk. But Kennedy was not most people. For one thing, he was a 13-year veteran of McKinsey & Co., the management consulting firm, and partner in charge of its Boston office. More to the point, he was the co-author of a recently published book that offered a startling new perspective on corporate life -- one that challenged the whole way people thought about business.

The book was entitled Corporate Cultures, a term that was itself new to the language, and it dealt with an aspect of business that, up to then, had been largely ignored. Broadly speaking, that aspect involved the role played by a company's values, symbols, rites, and rituals in determining its overall performance. Citing examples from some of the country's most dynamic companies, Kennedy and co-author Terrence E. Deal showed that these "cultural" factors had a major effect on the attitudes and behavior of a company's employees, and were thus of critical importance to its long-term success.

By any measure, the book was a groundbreaking work, challenging, as it did, the rational, quantitative models of corporate success that were so popular in the 1960s and '70s. But its impact had as much to do with its timing as its content. Published in June 1982, during a period of economic stagnation -- with unemployment at 9.5%, the prime over 16%, and trade deficits soaring to record levels -- Corporate Cultures offered a welcome antidote to the doom and gloom that was abroad in the land. Like In Search of Excellence, which appeared a few months later, it suggested that Japan was not the only nation capable of producing strong, highly motivated companies that could compete effectively in the international arena. America could produce -- in fact, was already producing -- its own.

What the book did not detail, however, was how corporate cultures were actually constructed. The authors could describe a particular culture and demonstrate its effects, but they offered few clues as to how a company might develop a culture in the first place. So the news that Allan Kennedy was going into business was greeted with more than passing interest among the followers of corporate culture. Here was an opportunity to find out how a living, breathing culture could be created, and the creator would be none other than the man who wrote the book.

After an extensive survey of business opportunities, Kennedy and Merlo decided to develop microcomputer software for sales and marketing management. They felt this was their most promising option, given the anticipated growth of the microcomputer market and their own experience as consultants. Acting on that assessment, they resigned from McKinsey and, in February 1983, formally launched Selkirk Associates Inc. with four of their friends.

Kennedy had lofty ambitions for Selkirk. More than a business, he saw it as a kind of laboratory for his theories. He wanted it to function as a society of professional colleagues committed to building a culture and a company that would stress collaboration, openness, decentralization, democratic decisions, respect, and trust. In this society, each individual would be encouraged to devise his or her own entrepreneurial response to the challenges of the business.

For Kennedy, this was not a long-term goal, something that would evolve naturally in the fullness of time. On the contrary, it was a pressing, immediate concern. Accordingly, he focused all his attention on creating such a culture from the start. "I spent lots of time," he says, "trying to think about what kind of values the company ought to stand for and therefore what kind of behavior I expected from people." These thoughts eventually went into a detailed statement of "core beliefs," which he reviewed and amplified with each new employee. In the same vein, Kennedy and his colleagues chose a "guiding principle," namely, a commitment to "making people more productive." They would pursue this ambition, everyone agreed, "through the products and services we offer" and "in the way we conduct our own affairs."

And, in the beginning at least, Selkirk seemed to be everything Kennedy had hoped for. The company set up shop in Boston, in an office that consisted of a large, rectangular room, with three smaller attachments. Each morning, staff members would pile into the main room and sort themselves out by function -- programmers and systems engineers by the windows, administrators in the middle, sales and marketing folk at the other end. In keeping with Kennedy's cultural precepts, there were no private offices or, indeed, any physical demarcations between functions.

It was a familial enterprise, informed with the very qualities Kennedy had laid out in his statement of core beliefs. The work was absorbing, the comradeship inspiring. Most mornings, the staff feasted on doughnuts, which they took to calling "corporate carbos," as a wordplay on "corporate cultures." They began a scrapbook as an impromptu cultural archive. Included among the memorabilia was "The Ravin'," an Edgar Allan Poe takeoff that commemorated Selkirk's first stirrings in earlier temporary headquarters:

Once upon an April morning, disregarding every warning, In a Back Bay storefront, Selkirk software was begun: True, it was without a toilet, but that didn't seem to spoil it.

To strengthen their bonds even further, the staff began to experiment with so-called rites, rituals, and ceremonies -- all important elements of a corporate culture, according to Kennedy's book. Selkirk's office manager, Linda Sharkey, recalls a day, for example, when the whole company went out to Kennedy's place on Cape Cod to celebrate their common purpose with barbecues on the beach. "The sun was shining, and we were all there together," she says. "It was a beautiful day. That's the way it was. We didn't use the terms among ourselves that Allan uses in the book. With us, corporate culture was more by seeing and doing." Sharkey remembers, too, Friday afternoon luncheons of pizza or Chinese food, at which everyone in the company had a chance to talk about his or her accomplishments or problems, or simply hang out.

Kennedy was pleased with all this, as well he might be. "We were," he says, "beginning to develop a real culture."

Then the walls went up.

The problem stemmed from the situation in the big room, where the technical people were laboring feverishly to develop Selkirk's first product, while the salespeople were busy preselling it. The former desperately needed peace and quiet to concentrate on their work; the latter were a boisterous lot, fond of crowing whenever a prospect looked encouraging. In fact, the salespeople crowed so often and so loudly that the technicians complained that they were being driven to distraction. Finally, they confronted Kennedy with the problem. Their solution, which Kennedy agreed to, was to erect five-foot-high movable partitions, separating each functional grouping from the others.

In the memory of Selkirk veterans, "the day the walls went up" lives on as a day of infamy. "It was terrible," says Sharkey. "I was embarrassed."

"It was clearly a symbol of divisiveness," says Kennedy.

"I don't know what would have been the right solution," says Reilly Hayes, Selkirk's 23-year-old technical wizard, "but the wall certainly wasn't. It blocked out the windows for the other end of the room. Someone [in marketing] drew a picture of a window and taped it to the wall. The whole thing created a lot of dissension."

Indeed, the erection of the walls touched off a feud between engineering and marketing that eventually grew into "open organizational warfare," according to Kennedy. "I let the wall stand, and a competitive attitude developed where engineering started sniping at marketing. We had two armed camps that didn't trust each other."

As if that weren't bad enough, other problems were beginning to surface. For one thing, the company was obviously overstaffed, having grown from 12 people in June 1983 to 25 in January 1984, without any product -- or sales -- to show for it. "That was a big mistake," says Kennedy. "We clearly ramped up the organization too fast, particularly given the fact that we were financing ourselves. I mean, for a while, we had a burn rate of around $100,000 per month."

Even more serious, however, was the problem that emerged following the release of the company's initial product, Correspondent, in February 1984. Not that there was anything wrong with the product. It was, in fact, a fine piece of software, and it premiered to glowing reviews. Designed as a selling tool, it combined database management, calendar management, word processing, and mail merge -- functions that could help customers organize their accounts, track and schedule sales calls and follow-ups, and generate correspondence. And it did all that splendidly.

The problem had to do with the price tag, a whopping $12,000 per unit. The Selkirk team members had come up with this rarefied figure, not out of greed, but out of a commitment to customer service -- a goal to which they had pledged themselves as part of their cultural mission. In order to provide such service, they figured, a Selkirk representative might have to spend two or three weeks with each customer, helping to install and customize the product. Trouble was, customers weren't willing to pay for that service, not at $12,000 per unit anyway. After a brief flurry of interest, sales dropped off.

"We just blew it," says Kennedy. "We were arrogant about the market. We were trying to tell the market something it wasn't interested in hearing. We took an arbitrary cultural goal and tried to make it into a strategy, rather than saying we're a market-driven company and we've got to find out what the market wants and supply it." Unfortunately, six months went by before Kennedy and his colleagues figured all this out and began to reduce Correspondent's price accordingly.

By then, however, Selkirk's entire sales effort was in shambles, a victim of its commitment to employee autonomy. Sales targets were seldom realized. Indeed, they were scarcely even set. At weekly meetings, salespeople would do little more than review account activity. "If a salesman said each week for three weeks in a row that he expected to close a certain account, and it never happened," says Merlo, "well, we didn't do anything about it. In any other company, he would probably have been put on probation." As it was, each of the participants entered the results of the meeting in a red-and-black ledger book and struck out once again to wander haphazardly through uncharted territory. "The mistake we made," reflects Merlo, "was using real money in a real company to test hypotheses about what sales goals should be."

Finally, in June 1984, Kennedy took action, laying off 6 people. In July, Correspondent's price was dropped to $4,000 per unit, but sales remained sluggish. In September, Kennedy laid off 5 more people, bringing the size of the staff back to 12.

One of those laid off was the chief engineer, a close friend of Kennedy's, but a man whose departure brought an immediate ceasefire between the warring factions. That night, the remaining staff members took down the walls and stacked them neatly in the kitchenette, where they repose to this day. "We felt," says Sharkey, "like we had our little family back together again."

With morale finally rebounding, Selkirk again cut Correspondent's price in the early fall, to $1,500. This time, sales responded, and, in November, the company enjoyed its first month in the black.

But Selkirk was not yet out of the woods. What remained was for Kennedy to figure out the significance of what had happened, and to draw the appropriate conclusions. Clearly, his experiment had not turned out as he had planned. His insistence on a company without walls had led to organizational warfare. His goal of providing extraordinary service had led to a crucial pricing error. His ideal of employee autonomy had led to confusion in the sales force. In the end, he was forced to fire more than half of his staff, slash prices by 87%, and start over again. What did it all mean?

Merlo had one answer. "We're talking about an experiment in corporate culture failing because the business environment did not support it," he says. "The notion of corporate culture got in the way of tough-minded business decisions." He also faults the emphasis on autonomy. "I don't think we had the right to be organized the way we were. I think we should have had more discipline."

Kennedy himself soon came around to a similar view. "Look in [the statement of core beliefs] and tell me what you find about the importance of performance, about measuring performance or about the idea that people must be held accountable for their performance," he says. "That stuff should have been there. I'm not discounting the importance of corporate culture, but you have to worry about the business at the same time, or you simply won't have one. Then you obviously won't need a culture. Where the two come together, I think, is in the cultural norms for performance, what kind of performance is expected of people. And that's a linkage that wasn't explicit in my mind three years ago. But it is now." He adds that, if the manuscript of Corporate Cultures were before him today, he would include a section on performance standards, measurement systems, and accountability sanctions.

On that point, he might get an argument from his co-author, Terrence Deal, a professor at Vanderbilt University and a member of Selkirk's board of directors since its inception. Deal does not disagree about the importance of discipline and performance standards, but he questions the wisdom of trying to impose them from above. The most effective performance standards, he notes, are the ones that employees recognize and accept as the product of their own commitment, and these can emerge only from the employees' experience. "One of the things that we know pretty handsomely," says Deal, "is that it's the informal performance standards that really drive a company."

In fact, Kennedy may have gotten into trouble not by doing too little, but by doing too much. Rather than letting Selkirk's culture evolve organically, he tried to impose a set of predetermined cultural values on the company, thereby retarding the growth of its own informal value system. He pursued culture as an end in itself, ignoring his own caveat, set down in his book, that "the business environment is the single greatest influence in shaping a corporate culture." Instead, he tried to shape the culture in a vacuum, without synchronizing it with the company's business goals.

In so doing, Kennedy reduced corporate culture to a formula, a collection of generic "principles." It was a cardinal error, if not an uncommon one. "There are a lot of people," says Deal, "who take our book literally and try to design a culture much as if they're trying to design an organization chart. My experience across the board has been that, as soon as people make it into a formula, they start making mistakes." By following the "formula," Kennedy wound up imposing his own set of rules on Selkirk -- although not enough of them, and not the right kind, he now says.The irony is that a real corporate culture allows a company to manage itself without formal rules, and to manage itself better than a company that has them.

Deal makes another point. Kennedy, he observes, might be less concerned with performance today if he had not hired so many friends at the beginning. Friends are nice to have around, but it's often hard to discipline them, or subject them to a company's normal sanctions. Over the long run, Deal says, their presence at Selkirk probably undermined the development of informal performance standards.

Kennedy himself may have played a role in that, too. He estimates that, over the past year, he has spent only one day a week at Selkirk. The rest of the time he has been on the road as a consultant, using his fees to help finance the company. In all, he has sunk some $1 million of his own money into Selkirk, without which the company might not have survived. But it has come at a price."Nobody had to pay attention to things like expenses, because there was a perception of an infinite sink of money," Kennedy says.

The danger of that perception finally came home to him last summer, when three of Selkirk's four salespeople elected to take vacations during the same month. The result was that sales for the month all but vanished. Kennedy had had enough. "I told the people here that either you sustain the company as a self-financing entity, or I will let it go under. I'm unwilling to put more money on the table."

And yet, in the end, it was hard to avoid the conclusion that a large part of Selkirk's continuing problem was Allan Kennedy himself -- a thought that did not escape him. "I've got a lot to learn about running a business successfully," he says, "about doing it myself, I mean. I think I know everything about management except how to manage. I can give world-class advice on managing, but -- when it comes right down to it -- I take too long and fall into all the traps that I see with the managers I advise."

Whatever his shortcomings as a manager, there is one thing Kennedy can't be faulted for, and that is lack of courage. Having drawn the inevitable conclusion, he went out looking for someone who could help him do a better job of managing the company. For several months, he negotiated with the former president of a Boston-based high-tech firm, but the two of them were unable to come to terms. Instead, Kennedy has made changes at Selkirk that he hopes will achieve the same effect. In the new structure, Merlo is taking charge of the microcomputer end of the business, while Betsy Meade -- a former West Coast sales representative -- has responsibility for a new minicomputer version of Correspondent, to be marketed in conjunction with Prime Computer Corp. As for Kennedy, he will concern himself with external company relations, product-development strategies, and, of course, corporate culture.

Kennedy is full of optimism these days. He points out that, despite its checkered history, Selkirk has emerged with a durable product and an installed base of about 1,000 units. In addition, the company will soon be bolstered with the proceeds from a $250,000 private placement. Meanwhile, he says, some of the company's previous problems have been dealt with, thanks to the introduction of a reliable order-fulfillment process, the decision to put sales reps on a straight commission payment schedule, and the establishment of specific sales targets for at least the next two quarters. "I think we have much more focused responsibility," he says, "and much more tangible measures of success for people in their jobs."

Overall, Kennedy looks on the past three years as a learning experience."There are times when I think I should charge up most of the zigs and the zags to sheer rank incompetence," he admits."But then there are other times when I look back and say, 'Nobody's that smart, and you can't do everything right.' In life, you have to be willing to try things. And if something doesn't work, you have to be willing to say, "Well, that was a dumb idea,' and then try something else." Now, he believes, he has a chance to do just that.

In the meantime, he is in the process of writing another book. He already has a proposal circulating among publishers. In his idle moments, he occasionally amuses himself by inventing titles. One of those titles speaks volumes about where he has been: Kicking Ass and Taking Names.