Managing Safety Dilemmas

Paul Balmert

Paul Balmert is a graduate of Cornell University’s School of Industrial and Labor Relations, and his career in chemical manufacturing spans 30 years. In 2000, Mr. Balmert formed Balmert Consulting (, a consulting practice that is principally focused on improving operations execution, including improving the management effectiveness in leading and managing safety performance. Mr. Balmert is the author of the best-selling book Alive and Well at the End of the Day; The Supervisor’s Guide to Managing Safety in Operations. He may be reached at

V. Scott Pignolet

October 1, 2005

dilemma: de-‘le-ma, noun—an argument presenting two equally conclusive alternatives

In the July issue of Insulation Outlook, we introduced the idea that managing performance with the goal of sending people

home safe at the end of the day frequently involves balancing dilemmas: two assumptions or conditions that are both true and

in opposition. We looked at the accountability dilemma—where what a supervisor controls is different from what s/he is

responsible for—and the risk dilemma—where acceptable risk differs from acceptable consequences. Both situations present a

real quandary.

In this second installment of our three-part series, we examine two more dilemmas that can send someone rushing off to the

medicine cabinet for migraine relief: the leader dilemma and the measurement dilemma.

The Leader Dilemma

The Hewlett-Packard (HP) board of directors’ decision to part company with CEO Carly Fiorina put the spotlight squarely on a

dilemma every high-profile business leader faces.

It might appear that this dilemma is all about the gap between what a leader is held accountable for and what that leader can

control or influence. That is the accountability dilemma, as described in the first article of this series on managing safety


The leader dilemma is an entirely different animal, and a very troubling one at that. It is first and foremost about

leadership at the top of an organization. That should come as good news to just about every reader of this article: Most

people work and live a long way from the corporate suite.

But no one should get too comfortable. There is something in the leader dilemma for leaders at any level who want to send

their team members home safe at the end of every day. This gets to the heart of what it takes to be the leader and what

leadership styles produce the best results: Who makes the best leaders?

Let’s start with a question: Which of the following leadership prototypes makes the best business leader?

  • Type A: the high-profile visionary; a superb communicator who is a tough and unrelenting driver of change
  • Type B: the self-effacing, limelight-avoiding tactician who tends to focus on continuous improvement rather than radical


These descriptions are oversimplifications of individual leadership styles, designed to make a point. Few leaders exactly

match either description, and there are plenty of good leaders who fit neither.

On the other hand, anyone who has been around an organization for a while has probably seen enough of both leadership

behavior types to recognize them and is likely to have an opinion on which style is the most effective.

So, which type of leader is best?

The modern theory of management teaches leadership techniques right out of the Type A description: Figure out the vision,

sell it to the organization and set about orchestrating the grand strategy to accomplish it. Stay relentlessly on task, but

leave the details to others. This is what “good” leaders are supposed to do (and is exactly why Fiorina was hired).

Type A leaders are not always the easiest bosses to work for, however, and that can play a big role in determining who makes

the better leader.

Working From the Bottom Up: Going From Good to Great

One reliable way to determine the most successful leadership style is to start at the bottom line, find businesses that get

the best results and examine the type of leaders who produced those results. Jim Collins did exactly that, as detailed in his

book Good to Great. The results are surprising.

Collins was a professor at Stanford University’s Graduate School of Business and is the author of Built to Last, a book that

documents the approach taken by businesses that “got it right from the start.” Successful companies that he researched

include General Electric Co., Johnson & Johnson, Wal-Mart and American Express. Ironically, HP was also one of those

built-to-last companies.

Built to Last was a bestseller, and Collins went on the lecture tour, giving speeches that drew crowds. However, audience

members often articulated what you or I would have wanted to say had we been there: “Nice to hear about a company that got it

right from the start, but that’s not my company.”

Collins heard that comment so many times that he finally decided to stop arguing and start looking: Which companies went from

average to great, and how did they do it?

When he started examining good-to-great companies, he set the standard high. “Good” was defined as average returns to

shareholders over a period of 15 years; “great” described shareholder returns that were three times the market average for 15

years. Thirty years of business performance ruled out any one leader’s good or bad effects.

Surprise, Surprise

So who showed up on the list of companies in Good to Great? First surprise: Not high-flying, high-tech companies. The winners

of the 1,473 companies examined were 11 rather standard names: The Kroger Co., Abbott Laboratories, Circuit City® Stores

Inc., Fannie Mae, The Gillette Company, Kimberly-Clark Corporation, Nucor Corporation, Phillip Morris, Pitney Bowes, Walgreen

Company and Wells Fargo.

The second surprise lies in the answer to the question, Was the secret of their success strategy technology, products, IT

and/or service? All played a role?but a supporting one. The foremost role was leadership—more specifically, leader

behavior. It turns out, there is truth to the quote by Peter Drucker, “Companies don’t compete; managers compete.”

What kind of leader behavior was associated with business results that went from good to great, Type A or Type B? The answer

even shocked the author. The leader behavior described as Type B was found as the single-most critical factor in achieving

the business results Collins described as “great.” Companies in the great group were led by a succession of leaders described

as “self-effacing, quiet, reserved, even shy—these leaders are a paradoxical blend of personal humility and

professional will.”

Sure, these leaders could communicate and had a sense of direction, but like the New England Patriots, they were the

consummate no-namers—passionate about getting sustained results but happy to give others the credit, with no need for

the limelight.

The Dilemma, Please

So where is the dilemma? Here’s one: How did these relatively low-profile performers find their way to the top of their

organizations in the first place? Most of those in charge of selection—not to mention those on Wall Street—are

far more inclined to choose Type A leaders, those who make waves and get noticed. More significantly, their leader behavior

styles fit the perceived norm of what leaders are supposed to do.

Were it not for Collins’ bottom-up approach, we probably wouldn’t know who the Type B leaders are. The last place one would

ever expect to see them is on the cover of Business Week. It’s unlikely that even an M.B.A. student could name more some of

the 11 CEOs.

Here’s another dilemma: How do these types get results? The Type B leadership style seems counterintuitive. After all,

leaders are the people making great speeches and heading charges up the hill, right?

Collins’ research suggests a different effect of the Type A leadership style, where the leader becomes the focal point of

everyone in the outfit. The focus is on them, potentially at the expense of performance.

A Different View

Maybe it’s time to rethink what leaders actually do. It turns out that the way they lead plays a pivotal role in sustained

business success and, there is every reason to believe, in sustained safety performance improvement. It makes sense, and that

is the message for everyone who shares the goal of leading people to work and back home safely.

Collins makes the point that the good-to-great leaders are fanatically driven to produce sustained results for the good of

the organization, no matter what it takes and no matter who receives credit. Harry Truman once said, “You can accomplish

anything in life, provided that you don’t mind who gets the credit.”

This type of leadership is needed at every level of an organization where leaders hold the safety of those who work for them

in their hands. The workers must be the focus of leadership efforts, with the goal of keeping them safe to pursue all the

truly important things in their lives.

Keep your eye on that target, and don’t think you have to be a Type A leader to get great results!


At a 1972 executive council meeting at HP—in response to an industry award nomination—Bill Hewlett remarked,

“Look, we’ve grown because the industry grew. We were lucky enough to be sitting on the nose when the rocket took off. We

don’t deserve a damn bit of credit.”

After a moment of silence Dave Packard responded, “Well, Bill, at least we didn’t louse it up completely.”

The Measurement Dilemma

No good deed goes unpunished.

Consider the law of unintended consequences: What starts off as a solution to one problem may wind up creating a far bigger,

unpredictable one.

This maxim explains all sorts of problems. For example, when NASA was worried about leaky O-rings on solid rocket boosters 20

years ago, seal-test pressure was increased to ensure that there would be no leaks. Another example: Two hundred years ago,

gypsy moths were imported to the United States as a potential replacement for silk worms. When the plan failed, a few moths

were released, and they ultimately created a swarm of insects capable of deforesting New England.

Solutions to improve safety performance are also susceptible to this law. In measuring safety performance, the law of

unintended consequences created the measurement dilemma.

Manufacturing Measurement, Then and Now

The sophisticated measurement processes in place in manufacturing operations worldwide are one critical factor in the

revolution of manufacturing productivity and quality seen in the last 25 years.

Those from the baby boomer generation know it wasn’t always that way. In the 1960s, measurement in operations was pretty

crude. Those working in production may have known how much was made, but they probably knew little more. If a product did not

meet specification, that was the quality inspector’s problem, not the workers’.

By the 1980s, that model was crumbling. Demanding customers simply weren’t buying that kind of product quality anymore, and

there were plenty of manufacturers willing to supply better quality products. Thus, industry changed.

Quality and productivity gurus with names like Deming, Juran and Crosby stressed “doing it right the first time.” This

approach required far better information and performance measurement than previously seen. Everyone got into the business of

performance measurement.

Unlike many management fads, this was a change that has not gone away. The world of manufacturing measurement was forever

changed and became subject to the pressure of continuous improvement: Over time, Two Sigma became Six Sigma.

Meeting the Numbers

Anyone who has spent a day in management knows that measuring performance is only one part of the job. Rewards also play a

significant role: Set the goal, measure the performance and then recognize and reward performance. This model has served

management well in the quest for quality.

Paying for performance also has been the source of unintended consequences. Troubles at Worldcom as well as a few other

well-publicized corporate accounting scandals have shown how “meeting the numbers” can go awry. Managers did whatever it took

to meet the numbers that Wall Street expected. Their personal fortunes were at stake.

“The probability that a performance measure will be corrupted is proportional to its use in determining compensation.” This

is Darly’s law, and Darly was right on the money. However, measurement as it relates to improving product quality has one big

factor on its side: It is virtually impossible to make the numbers into something that they’re not. Customers keep the system


As tempting as it might be to try to make product quality look better, the data are what they are. The process began by

defining quality as “conformance to requirements.” Darly defined the requirements, and the customers are the scorekeepers in

this game.

When it comes to product quality, “meeting the numbers” means meeting the numbers.

Measuring Safety Performance

In managing and measuring safety performance, “meeting the numbers” has long loomed large.

An old story illustrates just how far a company might go to keep its safety record: It seems a poor fellow working for one of

the industry leaders in safety had the misfortune of falling off the top of a smokestack. By the time he hit the ground, he

was fired. That’s one way to meet the numbers!

Measurement processes for safety run directly in the face of both Darly’s law and the law of unintended consequences. Here is

an example of each that has undermined the measurement system:

  • Believing that increasing the number of peer-to-peer safety observations would improve safety performance, a small reward

    was offered to employees making two or more safety observations each month. The number of observations steadily increased,

    but safety performance did not. As a matter of fact, those filling out safety observation cards had an injury rate higher

    than their nonparticipating peers.

  • Convinced that better results from safety audits ultimately would lead to improved safety performance, the internal

    safety audit process was stepped up. Performance rewards and consequences were established based on audit findings. Getting

    “audit-ready” took on a life of its own. Audit grades improved, and that led to even tougher audits. Meanwhile, safety

    performance headed in an independent direction. Over time, audit scores and safety performance no longer aligned.

It is not difficult to see the effect of Darly’s law and the law of unintended consequences in each case. It appears that no

good deed goes unpunished.

Things don’t always go this way, though. There are plenty of success stories on the impact of observation and audit programs.

However, there are enough failures to suggest that the problem is real.

Defining The Measurement Dilemma

Here, then, is the measurement dilemma: Measurement is a vital part of managing and improving safety performance—”If you

can’t measure it, you can’t manage it”—but the process of measuring can alter behavior.

A change in behavior can be beneficial; however, it also may be simply a case of management getting the desired data. In the

latter case, the value of the measurement is diminished or nullified.

Unlike product quality, when it comes to measuring safety performance, there is no customer to help keep the system honest.

The dilemma is more pronounced when behavior is being measured and is particularly acute when the factor being measured is

used as a “leading indicator.”

If that is the case, and if management does not appreciate what is really happening, this can lead to a false reading about

performance improvement.

Results and Consequences

How can you step up measurement without creating unintended consequences? Here are three ideas: Beware of certainty, use

multiple measures and remember Darly’s law.

Naming the beast is the first part of taming it. In physics, this is called the uncertainty principle: You can’t be sure

about everything you observe. For safety, this is the measurement dilemma. The best way to manage the problem is to beware of

certainty: Don’t fool yourself into thinking that the data present an accurate picture of what is really going on.

Multiple measures are better than one. In performance measurement, this is known as the “balanced scorecard” approach: Do not

depend on just one measure of performance or leading indicator as the means of understanding what is happening.

If the factors don’t match up, that should be the tip-off to dig deeper into the numbers to see what is really going on. For

example, audit scores are up, but injury performance is flat or down. Why?

Incentives can be a great way to encourage and recognize the right behavior, but they can also undermine the measurement

system. Ignore Darly’s law at your own risk. If incentives are used, the Darly effect should be considered.

All of this means that leaders have to rely on other means to calibrate underlying behavior and performance, and on other

methods to drive change and improvement. For example, if an incentive for safety observations is to be used, it could be

paired with an independent audit of behavior. If audit performance becomes important, audit scores should be correlated with

bottom-line safety results.

The Final Word

The measurement dilemma should not discourage leaders from applying the techniques of performance measurement to the process

of managing safety. Leaders should just be sure to recognize the various factors that come into play: For this dilemma,

recognition is the name of the game.