The willing workers are then asked to do their

The Red Beads: work adhering to the strict procedures. They dip
their paddles and then have the results recorded
They have more to do with which is passed up to management. When a
worker produces a higher number of red beads
safety than you might think than the other workers Deming chastises him. If
a worker produces a lower number Deming
By Thomas A. Smith heaps praise on them and tells the workers who
made a higher
For over 45 years Dr. W. Edwards Deming number of red beads
would teach some major points of his to do better or the
management theory doing what he called “a plant will be closed.
stupid experiment” but he promise it would be After each willing
one “you would never forget.” He carried his worker’s results are
Red Bead experiment out over four days. It was read aloud Deming
a simple yet powerful demonstration of just how praises or chastises
perverse the American management system is them based on
and how it prevents continual improvement by whether they have made more or fewer red
overemphasizing maintenance of the status quo. beads. At the end of each day’s work he talks to
the workers and reminds them they must
He sets up the audience for a role playing improve. At the end of the fourth day he
experience by asking for volunteers to explains he can’t see any improvement and the
participate. He calls for six willing workers and plant will have to be closed. He thanks the
two inspectors. He explains how in his willing workers and tells them to pick up their
imaginary company they do everything the checks on their way out.
wrong way by keeping over staffed and adds a Deming then reviews and reflects on the results.
chief inspector. He then adds a recorder whose Below are figures from an actual Red Bead
job is to make independent counts. Deming Experiment1:
played the role of the foremen since they have
Name Day Day Day Day All 4
nobody available who knows they job and then 1 2 3 4
explains to the willing workers their job is to Dick 14 10 9 10 43
make only white beads. The customer will Pat 17 5 8 5 35
accept only white beads. Bob 8 8 9 6 31
Steve 12 11 12 8 43
Deming then explains “the rules.” The first rule Horst 12 11 12 8 43
Dave 9 11 7 10 37
is, everyone must put forth their best effort. The All 6 71 51 50 48 220
workers agree to this. Then he tells them they Avg (x- 11.8 10.2 9.6 9.2 9.2
have procedures they must follow and shows bar)
them how to handle a paddle they will use to dip
into a container filled with both red and white Deming starts to analyze the results. He points
beads. He provides every worker with the exact out that some worker’s results are above
same instructions of how to dip their paddle into average and some are below. On the first day
a container that is filled with 3,200 white beads Steve was The Man of the Day. Much better
and 800 red beads. He selects one willing than the rest of the workers. On the first day Pat
worker to be “average.” He constantly reminds had 17 red beads but improved to be number
the workers their job is to “make white beads one with only 5 on day two. Deming says she
only.” If they can’t make only white beads went from the worst to the best in just one day.
management will close the plant down. He tells He wonders what happened to the other
the workers he made a red bead just to show workers. He feigns dismay with their effort. He
them what they look like. just can’t understand what went wrong.

© Mocal, Inc. 2010

He goes on and the message becomes clear. number of red beads x-bar been? Some people
Even with the same identical tasks, tools, hazards a guess of 10 since the daily production
instruction and talent production outcomes will is 50 and 20% of 50 equals 10. Deming tells
vary. He has done this experiment hundreds of them they are wrong. He prods them for an
times and the results are always the same. That answer and to explain their reason for their
is, the outcomes of each worker varies even answer. He tells them x-bar has to settle down to
though the imaginary factory is always set up something but none of their answers are correct.
the same way. Some of the workers are above He pushes the audience to think harder and
and some are below average. This happens finally someone comes up with the idea the red
every single time. There is nothing you can do beads may be different than the white. Deming
about it. The system they work in is not points out they are different, they are painted
something they control. Put different people in red so they are somewhat larger because of the
the system and the results are always the same. coating. They feel different to the paddle. He
has used different paddles over the years and
He then displays a control chart that establishes each paddle has produced different x-bar.
the limits of variation. The chart is shown
below:

Day 1 Day 2 Day 3 Day 4

Deming then explains there are no patterns in Paddle Number 1 gave 11.2. Paddle number 2
the data. It is a nearly perfect constant cause gave 9.6 and Paddle number 3 is at 9.2. Every
system displayed on the chart. This system paddle has variation in it. It may be too small to
allows you to compute the limits of variation be detected by just a quick look at it but it is still
into the future. A pattern may be 7 or 8 points in there.
a row. No such thing. Nobody exceeded the
upper control limit. Came close but it didn’t If you have moderately good statistical control
happen. With this system you can be confident then x-bar will settle down to something
no one will exceed the upper or lower control eventually that you can use to make a valid
limits. You have statistical stability in the sense prediction. To change x-bar you will have to fix
what happens in the future will be pretty close things that are in the system.
to what has happened in these twenty-four
points. There is no absolute guarantee since The moral of the red bead experiment is the
empirical evidence is never complete. following:

Then Deming lays a different scenario on the  Variation exists in every process. You
audience. Given the fact there are 80% or 3,200 can never eliminate all of it.
white beads and 20% of 800 red beds what  When you plan something you must
would have been their prediction for the average include making a prediction about how

© Mocal, Inc. 2010 2

 things will happen. Past performance as they call them. It started with Frederick
guarantees nothing in the future. Taylor in the 1890’s and has been with us
 The work system has things in it that are ever since. So what does all of this and the
beyond the workers control. It is the red beads have to do with safety?
system, not the skill of the workers that
determines how they will perform. Let’s start with the fact safety management
 Management are the only ones that can takes its lead from command and control. To
change the system. It is best for this day foremen are told employee injuries
management to enlist the help of the are mostly the result of the unsafe actions or
workers to do this. at-risk behaviors of the individual workers.
 Some workers are above and some are In this world the most basic assumption
below average. There’s nothing you can management makes about safety is workers
do about it and the difference may not can control their performance and their
mean anything. performance is why most accidents happen.
They either ignore or deny variation exists
The relationship of the Red Bead experiment
in workers and treat them as though they are
to safety management bionic machines.

They are also ignorant about variation in all

The Red Bead experiment is a good example the common causes in the system that result
of how bad management leads to bad results. in accidents.1 These managers believe
Most people who watch the experiment holding workers accountable for their own
understand the foremen (Dr. Deming) is accidents is good management. In other
committing some really stupid mistakes. words if employees would only make an
Ironically they are exactly what American honest effort to “do things right the first
managers do in various degrees every day. time” their safety would be guaranteed. The
He starts with employing extra layers of sign of a good manager is one that isn’t
managers whose only job is to watch what afraid to hold people accountable for their
the workers do and report the results. Then action. Thus management must keep
he holds the workers accountable for their motivating workers to stay on guard and be
mistakes, making red beads, for which they alert because for some unknown reason
have no control. The audience can see the workers are lax when it comes to their own
red beads are built into the system and safety. The Red Bead Experiment shows the
beyond the control of the workers. But the folly of this line of thinking.
foreman cannot. His solution is to force the
workers to follow rigid procedures with no The foreman’s role could be easily adapted
way of letting them make any kinds of from lecturing workers on the perils of
suggestions for improvement. In short, he is making red beads to warning them about
the perfect command and control manager. having an accident. Dr. Deming exhorts the
workers to do better work. Do not make any
Deming conducted this experiment for over red beads or the company will go out of
forty-five years and many of his associates business. We employ the same technique to
continue using it today. The feedback from
current audiences shows this style of
management still reigns. It’s as though deep 1. William Shewhart who introduced Dr. Deming
down inside, somewhere in the DNA of to the thinking of statistical process control
American managers is a genetic code that defined common causes as those causes that
instructs them, no compels them to follow are inherently part of the process (or system)
the theory of command and control. The hour after hour, day after day and affect
truly believe they have no other option when
everyone working in the process.
it comes to managing people or subordinates

© Mocal, Inc. 2010 3

 prevent accidents. Management tells understand management structures actually
workers they are responsible for their own create behaviors. Becoming a systems
safety and will be held accountable for any thinker as opposed to a command and
accident or safety violation. This allows control manager, you come to realize why it
management to ignore any deficiencies that is necessary to look beyond individual
may exist in the system. Management uses mistakes or misfortunes to truly understand
“positive re-enforcement” to help workers to safety problems. We need to look beyond
stop committing unsafe actions or at-risk single events. We have to examine the
behaviors as though they are consciously structures or common causes in the system
and willingly doing so. that shape individual actions and generate
the conditions where events (accidents) are
They can’t see how this approach parallels more likely to happen. Stated another way,
what Dr. Deming is doing wrong when he if employees are guilty of committing unsafe
admonishes the workers for making red actions it is more than likely they do so
beads. Managers ask “Can’t workers control because of the system. Why? Because
their own actions?” as though it is a people do not knowingly try to inflict pain
rhetorical question. They don’t understand upon themselves.
employee injuries are the same as the red
beads in the experiment. The workers did This isn’t to say people aren’t responsible
not put the red beads in the container. Most for their own actions or some accidents are
accidents like defects are built into the indeed the result of special causes the
system just like the red beads. That is to say employees can control. But the fact is
ultimately businesses are set up to produce employees don’t design the work system -
things for customers. But all processes have management does. Therefore management
by-products that include waste, scrap and must own up to its responsibility and
defects. The worst kinds of defects in any accountability for the outcomes of the
work system are employee accidents. system which includes accidents. People
who stay focused on event explanations,
Dr. Deming was the first management such as unsafe actions, are always going to
expert to understand the majority of work from a reactive mode and never get to
accidents occur due to variation of common this higher level of understanding. They will
causes in the system. This thinking is also mistake symptoms, in this case unsafe
contrary to traditional command and control actions, for causes.
theory where managers believe most
employee accidents are caused by the unsafe Four Levels of Safety
actions and at-risk behaviors of workers.
They believe employees do things on
purpose and involuntarily in spite of all the Basically there are four levels of safety
effort management puts forth to control management each one having structures that
them and that is what causes safety drive the thinking and actions of managers
problems. The assumption being everything and workers. The first is called the
else in the system is OK. Although the red superstitious level where managers pay very
beads have been introduced arbitrarily in Dr. little attention to safety. As far as they are
Deming’s experiment they really represent concerned accidents are a result of fate,
variation that could come from anywhere in chance, luck, or magic. At this level
the system. They just happen to be managers believe all it takes for workers to
discovered in the final inspection process. be safe is a little common sense. The second
level is the referred to as the Neo or New
Peter Senge said “When placed in the same Taylorism level. This is a little more
system, people however different, tend to sophisticated than the first level but not too
produce similar results.” 2 It’s important to much so. At this level managers focus on
events and end up taking action after an
© Mocal, Inc. 2010 4
 accident happens. They are content to make any kind of valid determination is with
believe this is being proactive. Their safety the aid of statistical thinking which
credo should be: There’s never enough time ultimately is what Dr. Deming is telling us
to do things right but always plenty of time to do. Look for the patterns of things that
to do things over. Most companies operate at cause the red beads (employee accidents) in
this level. The third level is the beginning of your operations. Then have teams use the
what is called a learning organization. At Check, Plan, Do, Study and Act cycle to
this level managers are starting to study the system and find ways to prevent
understand systems and the thinking of SPC. them in the first place. That’s what continual
They try to identify common and special improvement is all about.
causes but their knowledge is limited so they
alternate between Level 3 and Level 2 •Employees are customers of safety
sending mixed messages to workers. And management
finally there is Level 4 where managers Level 4 - Continual •Fix the system
Improvement •Empowered teams apply CPDSA cycle to
have attained what Dr. Deming referred to safety problems
•Profound knowledge
as profound knowledge and look at safety
through a different lens. At this level
• The thinking of SPC is used to determing
managers and workers work together on Level 3 - Learning common and special causes
teams to constantly fix the system so Organization • Systems thinking
• Teamwork
accidents are kept to an absolute
minimum.
•MBO - Compliance & Certification is the
goal
Deming estimated that as much as 99% of Level 2 - Neo- •Chaos management - Firefighting routines
Taylorism & heroic efforts solve safety problems
causes of accidents stem from common •Event focused - Every single
causes in the system and only 1% from incident/accident is investigated
carelessness. These kinds of accidents will
•Command & Control - Taylorism The Boss =
not be eliminated until the system is Level 1- The Customer
corrected. 3 The only place the system is Superstitious •Unsafe actions cause most (85%) of
accidents
being worked on is Level 4. In his Management •Common sense is all it takes to be safe
experiment the red beads represent defects
created by the system. But they could just as The Four Levels of Safety Management
well be accidents.

Focusing on fixing a defect or changing an

unsafe behavior would be similar to a doctor To learn more about how your
company can manage your safety
prescribing an aspirin for a headache caused
program for continual
by a brain tumor. Instead we should be improvement contact Thomas A.
looking at the common causes in the work Smith at Mocal, Inc. Mr. Smith
system to determine how their variation works with management and
results in employee accidents. In the Red hourly employees to help them
Bead experiment participants start to realize learn about new theory of management to obtain
x-bar of the red beads is dependent on team skills and work on culture change. His book;
variations in the paddle and the beads. System Accidents: Why Americans Are Injured At
Things not easily identified at first like the Work And What Can Be Done About It has received
diameter and roundness of each of the 50 high praise and can be obtained at Amazon.com. He
holes in the paddles and the beads can be reached at [email protected] or his
company website at www.mocalinc.com or (248)
themselves.
391-1818.
We should first ask, was the accident a
result of the common causes or was it
because of a special cause? The only way to

© Mocal, Inc. 2010 5

End notes:
1
Walton, Mary, The Deming Management Method,
Perigee Books, 1986, p. 46
2
Senge, Peter, The Fifth Discipline, The Art & Practice of
The Learning Organization, Doubleday, 1990, p. 42
3
Deming, W. Edwards, Out of the Crisis, MIT, 1995, p.
479.

© Mocal, Inc. 2010 6