Maverick! Ricardo Semler’s 10 democratic changes at Semco

by | 31 Jan 2014 | Business & Leadership

In my previous blogs, I have given some background to the story of Ricardo Semler, who took over the Brazilian manufacturing company, Semco, from his father and nearly ruined his health by behaving as the kind of hard-driving autocratic business owner that he assumed he ought to be.

Semler decided to change his life and his company. He started asking his colleagues to decide how they wanted to run the company.  A form of ‘democracy’ took hold – but I have argued that it was not democracy as we know it in its usual political form. This was a form of genuine, total democracy – which is, technically speaking, a lot more like anarchy or libertarianism.

I have argued in my earlier blogs Ricardo Semler, Maverick: organisational democracy or anarchy? and Anarchy in the workplace: in praise of self-organisation that this kind of organisational ‘anarchy’ is what we need for the organisations of the future.

To judge for yourself whether the most significant changes that Semler made at Semco or, to be more accurate, the changes that Semco decided to make to itself, are strictly speaking ‘democratic’, or something else, you need to know what was done.  Here’s a brief account.

 1.       Small, symbolic changes that make a big difference

Semler ended random searches aimed to prevent petty theft of stock. He took away ‘the big time clock at the main gate’ and introduced many smaller time clocks; employees were ‘requested’ not to clock-in their colleagues.

He eliminated dress codes. Manual workers, who were given work overalls, were invited to choose the colour of these overalls (petroleum blue: smart, but good at disguising the most typical workplace stains). Executive parking spaces were done away with.

‘’Democracy,’ writes Semler, ‘begins with little things like neckties, time clocks, parking spaces and petroleum blue uniforms.’

2.       Workers’ committees take on a managerial role

Committees were created in Semco’s four business units to represent all employees except management. Their mandate was to represent workers’ interests; committee members were given time off for their committee duties.

The committees asked for improvements to common areas and plant cafeterias. Then they started to question various business expenses. Then they began to get involved in areas that were traditionally the preserve of management.

The worker committees span off groups that began to explore their own productivity.

‘They . . . set production goals,’ writes Semler, ‘and suggested major changes in products.’

Later on, committees began to switch from production line to batch production, with individual workers developing a range of skills. The committees began to make financial/staffing decisions.

In hard times committees took decisions to lower wages or increase hours to save jobs. If layoffs were inevitable, the committees took part in the selection of staff who were laid off, considering history, loyalty, the ability to find new job and family responsibilities. The dismissal of an employee with more than three years’ service or who was older than fifty years old needed special approval.

 ‘Together we tried to be socially just,’ writes Semler. The various processes slowed things down, but, he says: ‘perhaps that was an unavoidable price for corporate democracy.’

3.       A new leadership network

Semler set up weekly meetings ‘of all those in leadership positions.’ The marketing manager, for example, would present their spending plans to the other members of the leadership team to get them approved – something that had previously been the prerogative of the top management team.  If the new leadership team cannot agree, a vote is taken.

Agendas for the weekly leadership began to get shorter as people got into the habit of making their own decisions.

The new system made leaders demonstrate exactly how they added value to the process: since the leadership team now set budgets, management roles that were clearly ineffective risked being eliminated.

‘This group puts together the budget,’ Semler records himself saying to the marketing manager, while explaining how the new system will work, ‘and they’ll only include you in it if they think you’re a good investment’.

The group began to take real control of operational decisions.

 ‘Managers have been consulting employees for centuries,’ says Semler. ‘It’s only when bosses give up decision-making and let their employees govern themselves that the possibility exists for a business jointly managed by workers and executives.’

4.       A bonfire of bureaucracies

Semco began to ditch its procedure manuals (and its organisation chart) and never replaced them.

Semler writes: ‘In their quest for law, order, stability and predictability, corporations make rules All those rules cause employees to forget that a company needs to be creative and adaptive to survive. Rules slow it down.’

Semco stopped policing items such as travel expenses, for example.

‘[A company] doesn’t need to know if the taxi ride being claimed for a manager is for business. Or if another manager couldn’t have stayed on a hotel with three rather than four stars.’

It let executives make their own arrangements for servicing and maintaining their company cars.

A whole range of day to day decisions about expenditure, argues Semler, should be ‘ruled by wisdom that varies from factory to factory and worker to worker’.

‘To do otherwise only gives those tough guy controllers the comfortable feeling that the company is organized, and provides jobs for dozens of disturbed souls who should be retrained for some useful purpose.’

Semco phased out clerical staff, asking secretaries to think about what they would like to do long term as their jobs are slowly phased out (marketing, sales, engineering …?)

5.       Divide and Prosper

Semco divided its plants into smaller units. Workers in these units then chose to organise themselves into ‘cells’: teams of workers who assembled groups of different machines and ‘fashion[ed] a product from beginning to end giving them accountability for the product’s quality and an the enormous satisfaction that comes with a complete task.’

Because workers now had, in effect, a ‘basket of jobs’, the appropriate salary for this array of skills was agreed. The cells took over quality control, and also hiring and firing.

Semler agrees that not being able to order in bulk was more expensive for each unit. But then the smaller units carried less inventory (‘some of our units turn over their complete inventory 17 times a year, as against an industry average of slightly more than three such rotations.’) Products might take longer to produce, but delivery times still fell. There was no longer any need for expensive and time-consuming quality control departments.

‘In times of robust economic growth we have found our divided plants make more money than they did when they were larger. And we have also found that smaller plants bounce back from bad times or a crisis much faster than larger ones. From all of this I have come to believe that economy of scale is one of the most overrated concepts in business.’

6.       Profit (and information) sharing

Discussions about profit sharing quickly led to enquiries about the salaries of top executives. Nobody was forced to reveal their salary, but ‘enough salaries were known to give an extremely accurate idea of pay at all levels of the company.’

‘The truth may not be pretty, or easily explained,’ says Semler, ‘but it is always better to be out in the open.’

A simple formula was agreed on for the allocation of profit for each autonomous unit: 40% would go towards taxes; 25% for dividends; 12% for reinvestment and the remaining 23% would be shared.

There were no consolation prizes if there was no profit and the units were free to allocate the money in whatever way they thought was most appropriate. In the event, it was decided to split the money evenly, giving the same profit share to every employee regardless of salary. But the teams could have chosen to distribute profit by age, or seniority, or to make the money available as low-cost loans to enable team members to get onto the property ladder.

‘Few ideas are as capitalist as profit-sharing, which rewards with a part of the company’s earnings the people who help generate this blessed surplus. What is a bonus scheme, after all, but a form of profit-sharing? … The truth is that profit-sharing doesn’t create employee involvement, it requires it.’

7.       Appraise your boss

Semco allowed its workforce a say in the appraisal and appointment of senior executives.

Mangers were evaluated by their teams twice a year by means of anonymous multiple choice questionnaires exploring technical ability, competence, leadership and other aspects of the leader’s performance. The questions were weighted to reflect the significance of each aspect and the results were made public. Most managers scored around 80-85%; a score of less than 70% would create pressure for change.

New managers face a group interview with the team as part of the appointment process and were graded against the various competencies. A score of over 70% allowed the appointment to proceed. Factions form (factory workers, engineers, administrators) but, interestingly, people‘s instinct for compromise and fair play seems to come to the fore. Anarchy does not rule. Or, as I would like to have it, anarchy does rule, and is remarkably sensible and efficient.

8.       Rounding the pyramid

Semler introduced a system of ‘concentric circles’ in an attempt to break down the transitional hierarchical pyramid of control.


The company’s six most senior executives, the innermost circle, were dubbed Counsellors; their role was to ‘stimulate decisions’. Each Counsellor owns a meaningful 1% of company, jointly controlling a significant 5% bloc.  The role of Chief Executive was rotated every six months among Counsellors.


The leaders of business units, of which there were between seven and ten, were dubbed Partners; their role was to ‘run the company


Everybody else – the all-encompassing largest circle – was an Associate.


Scattered throughout largest circle were Coordinators: ‘Leaders of departments or specific activities, guiding teams of five to twenty Associates in their areas.’

Coordinators can trade jobs with other Coordinators, or move back to being Associates.  A specialist software engineer Associate might earn more than Engineering Coordinator. No Coordinator or Associate can report to another. The hierarchy of the typical organisation is successfully flattened.

Semler writes:  ‘My circles would free people from hierarchical tyranny; they would act as leaders when they wanted and command whatever respect their competence earned them.’

Associates make all decisions they feel comfortable with, and if necessary consult a Coordinator, who raises issues at weekly Monday meetings with Partners. On Tuesdays, representatives from each business unit (not necessarily a Partner) discuss issues that affect all business units (such as a company-wide wage increase) or are significant (such as a large investment in new equipment).

 ‘Just three circles, four job categories, and two meetings,’ writes Semler. ‘That’s it . . .  Implementing the new system meant ripping apart the pyramid, clearing away whole levels of management, eliminating a host of titles, and breaking established chains of command.’

In implementing these changes, Semco lost perhaps 50 leaders, from Executive VP to foremen.

‘Their tasks were redistributed, but not formally, because we no longer had specific job descriptions, only those four titles. We just said, ‘This person isn’t here anymore …’ and let the people involved figure out what to do about it …’

‘Sadly, we lost a lot of talent in the transition. But it was a price we were willing to pay.’

 9.       Set your own salary and bonus

Staff were invited to suggest their own salaries, based on what they would be able to earn outside; what others with similar skills and responsibilities earned at Semco, what friends with similar backgrounds earn, and what they needed to live. Everyone came to know what everyone else was paid. It was agreed that top salaries would be no higher than ten times entry level pay.

Semco set its operational budget every six months, based on the clever perception that a one-year budget allows people to imagine that the second half of the year will be miraculously better than the first half.  In these budget sessions, if times were hard, high salaries stood out as a target for cuts. When setting their own salaries, people began to assess their own worth quite realistically.

Managers were invited to set their own goals, and then decide to what extent they had achieved them, thereby deciding the appropriate level of their own bonus.

10.   Spin off new satellite suppliers

Semco began a ‘satellite programme’ of farming out many functions by encouraging employees to set up as entrepreneurial sub-contractors.

Semler writes ‘Theoretically, farming out work means a loss of the profits of verticalisation.’ But then he reminds us that Henry Ford’s attempts at owning every aspect of the car manufacturing business (‘vertical organisation’) had ended in tears: at one time the Ford Motor Company owned its own forests, rubber plantations, iron mines and cargo ships. But it wasn’t very good at running those businesses.

The company retained the things that it did best: operations it would be expensive or impossible to contract out; for example, a special 5 metre lathe and specialist knowledge areas, such as design engineers and computer-aided design. Less specialised areas were spun-off: sheet metal work; drafting in ink; legal work; software design. Semco leased machines to its former workers at favourable costs and helped new businesses to set prices, control costs, manage inventory and deal with paperwork.

The company shrunk from 830 employees in 9 business units at 5 locations in 1987 to under 300 employees in 6 business units on 2 sites in 1991. But some 200 workers were by then working in satellite companies supplying Semco, or employed as full-time consultants. The new satellite suppliers are allowed to use parking facilities, desks, computers and phones at Semco sites.

Semler writes: ‘… we’ve shrunk Semco to a core of sales, engineering, design, materials handling, purchasing and assembly people …’ in pursuit of the ‘…vision of a smaller, more fluid, more flexible, less defined company …’

It’s not democracy – it’s better than that

Well, that’s what Semco did. Every change was instigated by common agreement. There were a great many other changes that were deemed to be unsuccessful, or unwelcome, that were abandoned. The 10 changes listed here are merely the changes that proved to be the most successful, and therefore enduring. Perhaps, by now, Semco has decided to change some of these arrangements again.

It’s a fluid, self-organising system. It has been inspired by Semler himself, but it now no longer needs him.

Which is just what Semler wanted. Not just because he wanted to run his company ‘in a simpler way, a more natural way’, and because he wanted to stop killing himself with overwork, but also because he thought he might make more money.

‘I stood to make at least as much money in partnership with a motivated workforce as I would as the sole beneficiary of the fruits of less inspired workers.’

Semler, let me remind you, is still the owner. Semco is not a cooperative; it’s a self-organised capitalist venture.

I’m interested in cooperatives, trust-owned companies and other enlightened ways of running a successful venture for the benefit of the many, not the few.

But it’s perfectly possible, as Semler shows, to own a company and make a lot of money while also giving people control over their own working lives in a way that delivers job satisfaction, peace of mind, and a satisfying sense of communal effort towards a shared purpose.

I don’t think its democracy, technically speaking. I think it’s better than that.

You might like to read this post to find out whether your own workplace is stuck on the industrial era or fit for the Age of Ideas: Is your organisation stuck in the industrial era?


Share this article on social channels: