Managing without Managers
by Ricardo Semler, Harvard Business Review (Sep-Oct 1989)
In Brazil, where paternalism and the family business fiefdom still flourish, I am president of a manufacturing company that treats its 800 employees like responsible adults. Most of them –including factory workers – set their own working hours. All have access to the company books. The vast majority vote on many important corporate decisions. Everyone gets paid by the month, regardless of job description, and more than 150 of our management people set their own salaries and bonuses.
This may sound like an unconventional way to run a business, but it seems to work. Close to financial disaster in 1980, Semco is now one of Brazil’s fastest growing companies, with a profit margin in 1988 of 10% on sales of $37 million. Our five factories produce a range of sophisticated products, including marine pumps, digital scanners, commercial dishwashers, truck filters, and mixing equipment for everything from bubble gum to rocket fuel. Our customers include Alcoa, Saab, and General Motors. We’ve built a number of cookie factories for Nabisco, Nestle, and United Biscuits. Our multinational competitors include AMF, Worthington Industries, Mitsubishi Heavy Industries, and Carrier.
Management associations, labor unions, and the press have repeatedly named us the best company in Brazil to work for. In fact, we no longer advertise jobs. Word of mouth generates up to 300 applications for every available position. The top five managers – we call them counselors – include a former human resources director of Ford Brazil, a 15 year veteran Chrysler executive, and a man who left his job as president of a larger company to come to Semco.
When I joined the company in 1980, 27 years after my father founded it, Semco had about 100 employees, manufactured hydraulic pumps for ships, generated about $4 million in revenues, and teetered on the brink of catastrophe. All through 1981 and 1982, we ran from bank to bank looking for loans, and we fought persistent, well-founded rumours that the company was in danger of going under. We often stayed through the night reading files and searching the desk drawers of venerable executives for clues about contracts long since privately made and privately forgotten.
Most managers and outside board members agreed on two immediate needs: to professionalize and to diversify. In fact, both of these measures had been discussed for years but had never progressed beyond wishful thinking.
For two years, holding on by our fingertips, we sought licenses to manufacture other companies’ products in Brazil. We travelled constantly. I remember one day being in Oslo for breakfast, New York for lunch, Cincinnati for dinner, and San Francisco for the night. The obstacles were great. Our company lacked an international reputation – and so did our country. Brazil’s political eccentricities and draconian business regulations scared many companies away.
Still, good luck and a relentless program of beating the corporate bushes on four continents finally paid off. By 1982, we had signed seven license agreements. Our marine division –once the entire company – was now down to 60% of total sales. Moreover, the managers and directors were all professionals with no connection to the family.
With Semco back on its feet, we entered an acquisitions phase that cost millions of dollars in expenditures and millions more in losses over the next two or three years. All this growth was financed by banks at interest rates that were generally 30% above the rate of inflation, which ranged from 40% to 900% annually; There was no long-term money in Brazil at that time, so all those loans had maximum terms of 90 days. We didn’t get one cent in government financing or from incentive agencies either, and we never paid out a dime in graft or bribes.
How did we do it and survive? Hard work, of course. And good luck – fundamental to all business success. But most important, I think, were the drastic changes we made in our concept of management. Without those changes, not even hard work and good luck could have pulled us through.
Semco has three fundamental values on which we base some 30 management programs. These values – democracy, profit sharing, and information – work in a complicated circle, each dependent on the other two. If we eliminated one, the others would be meaningless. Our corporate structure, employee freedoms, union relations, factory size limitations – all are products of our commitment to these principles.
It’s never easy to transplant management programs from one company to another. In South America, it’s axiomatic that our structure and style cannot be duplicated. Semco is either too small, too big, too far away, too young, too old, or too obnoxious.
We may also be too specialized. We do cellular manufacturing of technologically sophisticated products, and we work at the high end on quality and price. So our critics maybe right. Perhaps nothing we’ve done can be a blueprint for anyone else. Still, in an industrial world whose methods show obvious signs of exhaustion, the merit of sharing experience is to encourage experiment and to plant the seeds of conceptual change. So what the hell.
Participatory Hot Air
The first of Semco’s three values is democracy, or employee involvement. Clearly, workers who control their working conditions are going to be happier than workers who don’t. Just as clearly, there is no contest between the company that buys the grudging compliance of its work force and the company that enjoys the enterprising participation of its employees.
But about 90% of the time, participatory management is just hot air. Not that intentions aren’t good. It’s just that implementing employee involvement is so complex, so difficult, and not uncommonly, so frustrating that it is easier to talk about than to do.
We found four big obstacles to effective participatory management: size, hierarchy, lack of motivation, and ignorance. In an immense production unit, people feel tiny, nameless, and in capable of exerting influence on the way work is done or on the final profit made. This sense of helplessness is underlined by managers who, jealous of their power and prerogatives, refuse to let subordinates make any decisions for themselves – sometimes even about going to the bathroom. But even if size and hierarchy can be overcome, why should workers care about productivity and company profits? Moreover, even if you can get them to care, how can they tell when they’re doing the right thing?
As Antony Jay pointed out back in the 1950s in Corporation Man, human beings weren’t designed to work in big groups. Until recently, our ancestors were hunters and gatherers. For more than five million years, they refined their ability to work in groups of no more than about a dozen people. Then along comes the industrial revolution, and suddenly workers are trying to function efficiently in factories that employ hundreds and even thousands. Organizing those hundreds into teams of about ten members each may help some but there is still a limit to how many small teams can work well together. At Semco, we’ve found the most effective production unit to consist of about 150 people. The exact number is open to argument, but it’s clear that several thousand people in one facility makes individual involvement an illusion.
When we made the decision to keep our units small, we immediately focused on one facility that had more than 300 people. The unit manufactured commercial food-service equipment – slicers, scales, meat grinders, mixers – and used an MRP II system hooked up to an IBM mainframe with dozens of terminals all over the plant. Paperwork often took two days to make its way from one end of the factory to the other. Excess inventories, late delivery, and quality problems were common. We had tried various worker participation programs, quality circles, kanban systems, and motivation schemes, all of which got off to great starts but lost their momentum within months. The whole thing was just too damn big and complex; there were too many managers in too many layers holding too many meetings. So we decided to break up the facility into three separate plants.
To begin with, we kept all three in the same building but separated everything we could –entrances, receiving docks, inventories, telephones, as well as certain auxiliary functions like personnel, management information systems, and internal controls. We also scrapped the mainframe in favor of three independent, PC-based systems.
The first effect of the break-up was a rise in costs due to duplication of effort and a loss in economies of scale. UnfortunateIy, balance sheets chalk up items like these as liabilities, all with dollar figures attached, and there’s nothing at first to list on the asset side but airy stuff like “heightened involvement” and “a sense of belonging.” Yet the longer term results exceed our expectations.
Within a year, sales doubled, inventories fell from 136 days to 46, we unveiled eight new products that had been stalled in R&D for two years, and overall quality improved to the point that a one-third rejection rate on federally inspected scales dropped to less than 1%. Increased productivity let us reduce the workforce by 32% through attrition and retirement incentives.
I don’t claim that size reduction alone accomplished all this, just that size reduction is essential for putting employees in touch with one another so they can co-ordinate their work. The kind of distance we want to eliminate comes from having too many people in one place, but it also comes from having a pyramidal hierarchy.
Pyramids and Circles
The organizational pyramid is the cause of much corporate evil, because the tip is too far from the base. Pyramids emphasize power, promote insecurity; distort communications, hobble interaction, and make it very difficult for the people who plan and the people who execute to move in the same direction. So Semco designed an organizational circle. Its greatest advantage is to reduce management levels to three – one corporate level and two operating levels at the manufacturing units.
It consists of three concentric circles. One tiny, central circle contains the five people who integrate the company’s movements. These are the counselors I mentioned before. I’m one of them, and except for a couple of legal documents that call me president, counselor is the only title I use. A second, larger circle contains the heads of the eight divisions – we call them partners. Finally, a third, huge circle holds all the other employees. Most of them are the people we call associates; they do the research, design, sales, and manufacturing work and have no one reporting to them on a regular basis. But some of them are the permanent and temporary team and task leaders we call co-ordinators. Counselors, partners, co-ordinators, and associates. Four titles. Three management layers.
The linchpins of the system are the co-ordinators, a group that includes everyone formerly called foreman, supervisor, manager, head, or chief. The only people who report to co-ordinators are associates. No co-ordinator reports to another co-ordinator – that feature of the system is what ensures the reduction in management layers.
Like anyone else, we value leadership, but it’s not .the only thing we value. In marine pumps for example, we have an applications engineer who can look at the layout of a ship and then focus on one particular pump and say, “That pump will fail if you take this thing north of the Arctic Circle.” He makes a lot more money than the person who manages his unit. We can change the manager, but this guy knows what kind of pump will work in the Arctic, and that’s worth more. Associates often make higher salaries than co-ordinators and partners, and they can increase their status and compensation without entering the “management” line.
Managers and the status and money they enjoy – in a word, hierarchy – are the single biggest obstacle to participatory management. We had to get the managers out of the way of democratic decision making, and our circular system does that pretty well.
But we go further. We don’t hire or promote people until they’ve been interviewed and accepted by all their future subordinates. Twice a year, subordinates evaluate managers. Also twice a year, everyone in the company anonymously fills out a questionnaire about company credibility and top management competence. Among other things, we ask our employees what it would take to make them quit or go on strike.
We insist on making important decision’s collegially, and certain decisions are made by a company-wide vote. Several years ago, for example, we needed a bigger plant for our marine division, which makes pumps, compressors, and ship propellers. Real estate agents looked for months and found nothing, So we asked the employees themselves to help, and over the first weekend they found three factories for sale, all of them nearby. We closed up shop for a day, piled everyone into buses, and drove out to inspect the three buildings. Then the workers voted – and they chose a plant the counselors didn’t really want. It was an interesting situation – one that tested our commitment to participatory management.
The building stands across the street from a Caterpillar plant that’s one of the most frequently struck factories in Brazil. With two tough unions of our own, we weren’t looking forward to front-row seats for every labor dispute that crime along. But we accepted the employee decision, because we believe that in the long run, letting people participate in the decisions that affect their lives will have a positive effect on employee motivation and morale.
We bough the building and moved in. The workers designed the layout for a flexible manufacturing system, and they hired one of Brazil’s foremost artists to paint the whole thing, inside and out, including the machinery. That plant really belongs to its employees. I feel like a guest every time I walk in.
I don’t mind, The division’s productivity, in dollars per year per employee, has jumped from $14,200 in 1984 – the year we moved – to $37,500 in 1988 and for 1989 the goal is $50,000. Over the same period, market share went from 54% to 62%.
Employees also outvoted me on the acquisition of a company that I’m still sure we should have bought. But they felt we weren’t ready to digest it, and I lost the vote. In a case like that, the credibility of our management system is at stake. Employee involvement must be real, even when it makes management uneasy. Anyway, what is the future of an acquisition if the people who have to operate it don’t believe it’s workable?
We have other ways of combating hierarchy too. Most of our programs are based on the notion of giving employees control over their own lives. In a word, we hire adults, and then we treat them like adults.
Think about that. Outside the factory, workers are men and women who elect governments, serve in the army, lead community projects, raise and educate families, and make decisions every day about the future. Friends solicit their advice. Salespeople court them. Children and grandchildren look up to them for their wisdom and experience. But the moment they walk into the factory, the company transforms them into adolescents. They have to wear badges and name tags, arrive at a certain time, stand in line to punch the clock or eat their lunch, get permission to go to the bathroom, give lengthy explanations every time they’re five minutes late, and follow instructions without asking a lot of questions.
One of my first moves when I took control of Semco was to abolish norms, manuals, rules, and regulations. Everyone knows you can’t run a large organization without regulations, but everyone also knows that most regulations are poppycock. They rarely solve problems. On the contrary; there is usually some obscure comer of the rule book that justifies the worst silliness people can think up. Common sense is a riskier tactic because it requires personal responsibility.
It’s also true that common sense requires just a touch of civil disobedience every time someone calls attention to something that’s not working. We had to free the Thoreaus and the Tom Paines in the factory and come to terms with that fact that civil disobedience was not an early sign of revolution but a clear indication of common sense at work.
So we replaced all the nitpicking regulations with the rule of common sense and put our employees in the demanding position of using their own judgement.
We have no dress code, for example. The idea that personal appearance is important in a job – any job – is baloney. We’ve all heard that salespeople, receptionists, and service reps are the company’s calling cards, but in fact how utterly silly is that. A company that needs business suits to prove its seriousness probably lacks more meaningful proof. And what customer has ever canceled an order because the receptionist was wearing jeans instead of a dress? Women and men look best when they feel good. IBM is not a great company because its salespeople dress to the special standard that Thomas Watson set. It’s a great company that also happens to have this quirk.
We also scrapped the complex company rules about travel expenses – what sorts of accommodations people were entitled to, whether we’d pay for a theater ticket, whether a free call home meant five minutes or ten. We used to spend a lot of time discussing stuff like that. Now we base everything on common sense. Some people stay in four-star hotels and some live like spartans. Some people spend $200 a day while others get by on $125. Or so I suppose. No one checks expenses, so there is no way of knowing. The point is, we don’t care. If we can’t trust people with our money and their judgement, we sure as hell shouldn’t be sending them overseas to do business in our name.
We have done away with security searches, storeroom padlocks, and audits of the petty cash accounts of veteran employees. Not that we wouldn’t prosecute a genuinely criminal violation of our trust. We just refuse to humiliate 97% of the work force to get our hands on the occasional thief or two-bit embezzler.
We encourage – we practically insist on – job rotation every two to five years to prevent boredom. We try hard to provide job security, and for people over 50 or who’ve been with the company for more than three years, dismissal procedures are extra complicated.
On the more experimental side, we have a program for entry-level management trainees called “Lost in Space” whereby we hire a couple of people every year who have no job description at all. A “godfather” looks after them and for one year they can do anything they like as long as they try at least 12 different areas or units.
By the same logic that governs our other employee programs, we have also eliminated time clocks. People come and go according to their own schedules – even on the factory floor. I admit this idea is hard to swallow, most manufacturers are not ready for factory-floor flextime. But our reasoning was simple.
First, we use cellular manufacturing systems. At our food-processing equipment plant, for example, one cell makes only slicers, another makes scales, another makes mixers, and so forth. Each cell is self-contained, so products – and their problems – are segregated from each other.
Second, we assumed that all our employees were trustworthy adults. We couldn’t believe they would come to work day after day and sit on their hands because no one else was there. Pretty soon, we figured, they would start co-ordinating their work hours with their co-workers.
And that’s exactly what happened, only more so. For example, one man wanted to start at 7am, but because the forklift operator didn’t come until 8, he couldn’t get his parts. So a general discussion arose, and the upshot was that now everyone knows how to operate a forklift. In fact, most people can now do several jobs. The union has never objected because the initiative came from the workers themselves. It was their idea.
Moreover, the people on the factory floor set the schedule, and if they say that this month they will build 48 commercial dishwashers, then we can go play tennis, because 48 is what they’ll build.
In one case, one group decided to make 220 meat slicers. By the end of the month, it had finished the slicers as scheduled – except that even after repeated phone calls, the supplier still hadn’t produced the motors. So two employees drove over and talked to the supplier and managed to get delivery at the end of that day, the 31st. Then they stayed all night, the whole work force, and finished the lot at 4:45 the next morning.
When we introduced flexible hours, we decided to hold regular follow-up meetings to track problems and decide how to deal with abuses and production interruptions. That was years ago, and we haven’t yet held the first meeting.
Hunting the Woolly Mammoth
What makes our people behave this way? As Antony Jay points out, corporate man is a very recent animal. At Semco, we try to respect the hunter that dominated the first 99.9% of the history of our species. If you had to kill a mammoth or do without supper, there was no time to draw up an organization chart, assign tasks, or delegate authority. Basically, the person who saw the mammoth from farthest away was the Official Sighter, the one who ran fastest was the Head Runner, whoever threw the most accurate spear was the Grand Marksman, and the person all others respected most and listened to was the Chief. That’s all there was to it. Distributing little charts to produce an appearance of order would have been a waste of time. It still is.
What I’m saying is, put ten people together, don’t appoint a leader, and you can be sure that one will emerge. So will a sighter, a runner, and whatever else the group needs. We form the groups, but they find their own leaders. That’s not a lack of structure that’s just a lack of structure imposed from above.
But getting back to that mammoth, why was it that all the members of the group were so eager to do their share of the work – sighting, running, spearing, chiefing – and to stand aside when someone else could do it better? Because they all got to eat the thing once it was killed and cooked. What mattered was results, not status.
Corporate profit is today’s mammoth meat. And though there is a widespread view that profit sharing is some kind of socialist infection, it seems to me that few motivational tools are more capitalist. Everyone agrees that profits should belong to those who risk their capital, that entrepreneurial behavior deserves reward, that the creation of wealth should enrich the creator. Well, depending on how you define capital and risk, all these truisms can apply as much to workers as to shareholders.
Still, many profit-sharing programs are failures, and we think we know why. Profit sharing won’t motivate employees if they see it as just another management gimmick, if the company makes it difficult for them to see how their own work is related to profits and to understand how those profits are divided.
In Semco’s case, each division has a separate profit-sharing program. Twice a year, we calculate 23% of after-tax profit on each division income statement and give a check to three employees who’ve been elected by the workers in their division. These three invest the money until the unit can meet and decide – by simple majority vote – what they want to do with it. In most units, that’s turned out to be an equal distribution. If a unit has 150 workers, the total is divided by 150 and handed out. It’s that simple. The guy who sweeps the floor gets just as much as the division partner.
One division chose to use the money as a fund to lend out for housing construction. It was a pretty close vote, and the workers may change their minds next year. In the meantime, some of them have already received loans and have begun to build themselves houses. In any case, the employees do what they want with the money. The counselors stay out of it.
Semco’s experience has convinced me that profit-sharing has an excellent chance of working when it crowns a broad program of employee participation. When the profit-sharing criteria are so clear and simple that the least gifted employee can understand them, and perhaps most important, when employees have monthly access to the company’s vital statistics – costs, overhead, sales, payroll, taxes, profits.
Lots of things contribute to a successful profit-sharing program: low employee tumover, competitive pay, absence of paternalism, refusal to give consolation prizes when profits are down, frequent (quarterly or semi-annual) profit distribution, and plenty of opportunity for employees to question the management decisions that affect future profits. But nothing matters more than those vital statistics – short, frank, frequent reports on how the company is doing. Complete transparency. No hocus-pocus, no hanky-panky, no simplifications.
On the contrary, all Semco employees attend classes to learn how to read and understand the numbers, and it’s one of their unions that teach the course. Every month, each employee gets a balance sheet, a profit-and-loss analysis, and a cash-flow statement for his or her division. The reports contain about 70 line items (more, incidentally than we use to run the company, but we don’t want anyone to think we’re withholding information).
Many of our executives were alarmed by the decision to share monthly financial results with all employees. They were afraid workers would want to know everything like how much we pay executives. When we held the first large meeting to discuss these financial reports with the factory committees and the leaders of the metalworkers’ union, the first question we got was, “How much do division managers make?” We told them. They gasped. Ever since, the factory workers have called them “maharaja.”
But so what? If executives are embarrassed by their salaries, that probably means they aren’t earning them. Confidential payrolls are for those who cannot look themselves in the mirror and say with conviction, “I live in a capitalist system that remunerates on a geometric scale. I spent years in school, I have years of experience, I am capable and dedicated and intelligent. I deserve what I get.”
I believe that the courage to show the real numbers will always have positive consequences over the long term. On the other hand, we can show only the numbers we bother to put together, and there aren’t as many as there used to be. In my view, only the big numbers matter. But Semco’s accounting people keep telling me that since the only way to get big numbers is to add up the small ones, producing a budget or report that includes every tiny detail would require no extra effort. This is an expensive fallacy, and a difficult one to eradicate.
A few years ago, the U.S. president of Allis-Chalmers paid Semco a visit. At the end of his factory tour, he leafed through our monthly reports and budgets. At the time, we had our numbers ready on the fifth working day of every month in super-organized folders, and were those numbers comprehensive! On page 67, chart 112.6, for example, you could see how much coffee the workers in Light Manufacturing III had consumed the month before. The man said he was surprised to find such efficiency in a Brazilian company. In fact, he was so impressed that he asked his Brazilian subsidiary, an organization many times our size, to install a similar system there.
For months, we strolled around like peacocks, telling anyone who cared to listen that our budget system was state-of-the-art and that the president of a Big American Company had ordered his people to copy it. But soon we began to realize two things. First, our expenses were always too high, and they never came down because the accounting department was full of overpaid clerks who did nothing but compile them. Second, there were so damn many numbers insidethe folder that almost none of our managers read them. In fact, we knew less about the company then, with all that information, than we do now without it.
Today we have a simple accounting system providing limited but relevant information that we can grasp and act on quickly. We pared 400 cost centers down to 50. We beheaded hundreds of classifications and dozens of accounting lines. Finally, we can see the company through the haze.
As for Allis-Chalmers, I don’t know whether it ever adopted our old system in all its terrible completeness, but l hope not. A few years later, it began to suffer severe financial difficulties and eventually lost so much market share and money that it was broken up and sold. I’d hate to think it was our fault.
In preparing budgets, we believe that the flexibility to change the budget continually is much more important than the detailed consistency of the initial numbers. We also believe in the importance of comparing expectations with results. Naturally, we compare monthly reports with the budget. But we go one step further. At month’s end, the co-ordinators in each area make guesses about unit receipts, profit margins, and expenses. When the official numbers come out a few days later, top managers compare them with the guesses to judge how well the co-ordinators understand their areas.
What matters in budgets as well as in reports is that the numbers be few and important and that people treat them with something approaching passion. The three monthly reports, with their 70 line items, tell us how to run the company; tell our managers how well they know their units, and tell our employees if there’s going to be a profit. Everyone works on the basis of the same information, and everyone looks forward to its appearance with what I’d call fervent curiosity.
And that’s all there is to it. Participation gives people control of their work, profit sharing gives them a reason to do it better, information tells them what’s working and what isn’t.
Letting Them Do Whatever the HeIl They Want
So we don’t have systems or staff functions or analysts or anything like that. What we have are people who either sell or make, and there’s nothing in between. Is there a marketing department? Not on your life. Marketing is everybody’s problem. Everybody knows the price of the product. Everybody knows the cost. Everybody has the monthly balance sheet that says exactly what each of them makes, how much bronze is costing us, how much overtime we paid, all of it. And the employees know that 23% of the after-tax profit is theirs.
We are very, very rigorous about the numbers. We want them in on the fourth day of the month so we can get them back out on the fifth. And because we’re so strict with the financial controls, we can be extremely lax about everything else. Employees can paint the walls any color they like. They can come to work whenever they decide. They can wear whatever clothing makes them comfortable. They can do whatever the hell they want. It’s up to them to see the connection between productivity and profit and to act on it.