Motivating with Measures: Accountability, Incentives and the Dark Side

The benefits and risks of using measures for motivation are amplified when employees are made accountable or incentivized.

Measurement and Accountability

Measurement is at the heart of accountability. In the dictionary, accountability has a neutral meaning: an obligation or willingness to accept responsibility for one’s actions. This is the denotative, or literal, meaning. In a work setting, the denotative meaning of accountability is a goal that defines who will do what by when.

Accountability in this sense is the basis of management by objectives  (MBO).  While MBOs were popularized in the 1950s, they remain a central element of most organizations’ annual performance appraisals. 

While some objectives are task-based, the best objectives are measurement-based. We have found that the most effective examples of accountability-based motivation use SMART goals—goals that are Specific, Measurable, Agreed upon, Realistic, and Time-bound. Describing expectations in terms of measures at the beginning of a project motivates performance. 

Accountability is a word often loaded with connotations. How would you feel if you were told in a business meeting that you will be held accountable? Queasy? The phrase suggests that you’re in trouble. This isn’t actually accountability—it’s scapegoating. This is a connotative meaning, and the connotations of accountability are negative. The fear of negative consequences can lead to all sorts of dysfunctional behavior.

If the meaning of the measures isn’t managed, then accountability is more likely to instill a culture of fear than it is to motivate employees and support the organization’s strategic goals. 

To motivate with measurement-based accountability, the meaning of the measure must be managed. Use measures to describe expectations before the employee works to achieve results.  Articulate both the formal, denotative meaning (how the measure works) and the connotative meaning (the implications for the employee).

• If you’re building a measurement system, remember that the connotations are probably more important than the measures. Consider how the measures will be seen by employees. Develop a list of actions employees could take to influence the data.  Be sure to include actions that the system intends to encourage as well as unintended actions. Adapt your system accordingly to especially encourage the intended and discourage the unintended.
• If you’re managing employee accountability with measures, be sure to talk with employees about both the denotative and connotative meanings. It’s important to develop a shared vision: This is a true leadership communication task.  If there is clear agreement on the meaning of the measure as well as the level of performance expected, accountability can be positive. 

Accountability is simply responsibility. Measurement can help top build responsibility for results and the rewards or consequences of the results.

Incentives and Measurement

If goal setting and accountability work, why not add incentives to make them work even better? Why not juice the motivation system? Many of us have worked, or currently work, in an incentive system. Sales people work on commission. Managers get bonuses and stock options.

There is a whole industry of compensation consultants trying to create incentives that work. Since the industrial revolution, we’ve been trying to get incentives right—and some of us are starting to wonder if incentives are just wrong.  

Research summarized by Daniel Pink suggests that incentives lead to lower performance in completing tasks that are complex or involve creative thinking. I’m sure we’ll find that this relationship is true in many circumstances. 

I am also sure there are many circumstances in which incentives lead to improved performance, even in complex and creative tasks. As with many aspects of human performance, there are complexities.

Life isn’t one-size-fits-all; there are individual differences and nuances of context that influence how incentives affect performance. Long-term goals, which are difficult to study experimentally, may work better with incentives.  Mr Pink presents the world in black and white; i am confident there are many shades of gray.

There is a bigger problem with linking incentives to measures, however.  Incentives, or consequences, have the tendency to put the focus exclusively on moving the needle—on affecting the data and the measure rather than addressing the underlying goal. Too much focus on the connotations of the measure, as opposed to the meaning of the measure, leads to gaming.

The Dilbert comic strip may seem ridiculous, but as is always the case in Scott Adams’ cartoons, absurdity reflects reality to an uncomfortable degree (many of his cartoons are based on real-life examples submitted by readers). Incentives can have unintended consequences, often encouraging employees to behave unethically. For example, if you were earning a subsistence wage as a packer for Green Giant, and the company announced that a bonus would be paid to every employee who could find and remove insect parts from packages of frozen peas, what would you do? Possibly what many of the employees did—bring insect parts from home to earn the incentive.

There are, of course, more troubling examples of the dark side of measurement-based motivation. In the sad story of system-wide cheating in Atlanta Public Schools, 178 employees, including both teachers and principals, are now suspected of inflating scores on standardized tests to earn the significant rewards that come with rapid improvements in school performance. Outright swindles, such as Bernie Madoff, are all too common.

In sales departments there are more subtle examples of gaming incentive systems. Sales departments have been known to count all sales in the current quarter toward commissions—even though many of the sales are not actually closed. 

Conclusion

It’s dangerous to rely too much on measures for motivation: The more you emphasize measures, the more apt the measures are to cause dysfunctional, even unethical, behavior.  If you need to use measures for accountability and incentives, be careful.  Measures can’t replace management; they are a management tool.  It is necessary to make sure that the measures are reasonable – not gamed – and that accountability is understood and positive. 

Employee Performance Measurement is Increasing But Will Not Increase Efficiency

Employee performance has been measured since the start of the industrial revolution in the late 18^th century. By 1910, scientific management—the attempt to improve efficiency by applying engineering principles and measurement, in manufacturing —was reaching its peak. 

Although scientific management as a school of thought had faded by the 1930s, it continues to influence the way we measure and manage, and it provides fascinating historical insights into industry. Scientific management was largely focused on per-worker output, how to increase output by finding the right employees, incentive schemes and best practices to ensure that systems function optimally. 

Does this sound familiar? It should. Management, and our society in general, continually focus on these topics. Our use of measurement has only expanded. Today, 97% of organizations have an employee appraisal process, and many organizations are working to increase the number of appraisals to several times each year.  

With current social and technological trends, I expect human performance measurement will continue to increase. I also expect that people will continue to be surprised at the real outcomes and consequences of measurement. I’ll make another prediction: We’ll see little value from all this measurement, unless we start cultivating wisdom in our use of measures.

Frederick Taylor’s Underestimated Influence

The influence of Frederick Taylor, generally considered the founder of scientific management, has been compared to Darwin and Freud.  As Taylor wrote in his 1911 best seller Scientific Management:

… the end of our coal and iron is in sight. But our larger wastes of human effort, which go on every day through such of our acts as are blundering, ill-directed, or inefficient … are less visible, less tangible, and are but vaguely appreciated. … And for this reason, even though our daily loss from this source is greater than from our waste of material things, the one has stirred us deeply, while the other has moved us but little.

Taylor is the father of modern management.  At the turn of the century, management primarily were concerned with budgeting and hiring. Teams of workers defined the tasks. Workers in a sense were still in somewhat of a guild system and had some level of autonomy. Ideas presented by Taylor, and then Henry Ford’s production line, changed things considerably.

   Frederick W. Taylor

While famous for time and motion studies, Taylor had two bigger ideas associated with measuring employee performance:

• System improvement (also known as organizational learning)
• Employee accountability (to achieve targeted levels of performance).

These ideas continue to drive improvement efforts to this day. Taylor believed that per-individual output should be measured, and that the measure should be studied to understand both the system of work and how it can be improved. He also advocated for developing challenging targets for employee output.  

Taylor’s most famous example was of moving pig-iron, which was made completely by men’s labor.  To illustrate scientific management, he described an intervention at Bethlehem steel. He found that the average man moved about 12 ½ tons of pig iron each day, but the best handlers loaded 47 ½ tons. (I find both values astounding.)  By selecting the right men for the job, specifying how the loading task should be completed, and setting appropriate goals, he was able to improve per-employee pig-iron loading output. The loading tasks were streamlined and proper rests were enforced to ensure that a high level of output could be maintained for the entire day. Pig-iron handlers were accountable to achieve the goal of 47 ½ tons per day. Failure to consistently achieve the goal would result in reassignment.

Similarly, through careful study Taylor found the optimal shovel load was 21 pounds. This finding suggested many system-wide changes—in tools, team organization, and supervision. The organization provided different shovel sizes and shapes for different material density and characteristics. Workers that had been managed as teams were now managed as individuals, each accountable for performance targets.  Management structures changed to oversee these major changes.  

While Taylor’s time-and-motion studies are no longer relevant, given robotics and increasing service work, the measurement themes continue, and the themes of system improvement (organizational learning) and employee accountability continue to this day. In fact, these are two ideas with growing currency.

Organizational learning, which goes by many names, is simply the idea that organizations are entities that can learn and adapt. By looking deeply at the organization’s environment, processes, structures, routines, and culture, we can understand and improve them. There are many methods for encouraging organizational learning.  Whole conferences and societies have been developed around organizational learning, feedback, and measurement. In many situations, measurement of employee performance is the basis of organizational learning. 

Given the increased need to adapt organizations in our changing political, economic, social, and technical environment, I expect that measurement will increase. I worry, however, that organizations do not think clearly about the performance measures they track.

As organizations grow more complicated and employees more specialized, accountability for results has replaced the idea of managing tasks. Taylor was concerned with a moving pig-iron and micro-tasks such as lifting and carrying between stations. Now an employee is likely to have more complicated assignments comprised of dozens of tasks (for example, creating a new software feature or maintaining a database). Now we measure value-added and organizational results.  

Today, the essence of progressive management thinking could be stated as, “I do not want to micro-manage, so I am going to assign accountability for results and let my employees use their creativity and capability to get the job done.  In order to hold them accountable, I am going to measure their performance or productivity.” Many management theorists, such as Daniel Pink, argue for a result-oriented/only work environment (ROWE).  (Note that Taylor was indeed a micro-manager, so the parallel ends quickly.)  

I am a fan of managing results, not tasks. If you have ever been micro-managed, you’re probably a fan of managing results as well. I’m not completely sure, however, that we have the expertise to make it work. I see many organizations struggle with measures of performance—with pushback from employees and a blind push from management. ROWE works best when there are clear measures of performance that are linked to valued organizational outcomes.  These measures do not exist in many situations-- for example in support functions.

The final driver of increased measurement of employee performance is the computer-based work environment. Due to increasingly inexpensive databases, this environment allows more measurement and data-tracking. As computers facilitate more work, keystrokes, conversations, tasks, and transactions can be—and often are—automatically recorded. These digital trails can be compiled very cheaply. Footsteps, driving routes, and bathroom breaks can all be easily recorded with easily available technologies. It is inevitable that new measures of performance will be dreamed up, developed, and implemented. 

What Have We Learned About Managing and Leading With Measures?

My concern is our ability to manage, learn from, and improve from all this measurement. If we’re not careful, the additional performance measurement will only lead to additional misunderstanding and organizational chaos, misalignment, gaming of the system, and, in general, dysfunctional behavior. I’m not sure that our wisdom has kept pace with the amount of data and performance measures.  

In fact, problems associated with employee performance measurement appeared soon after the advent of scientific management.  Measuring employee performance often leads to strange outcomes. 

This is demonstrated by what’s known as the Hawthorn Effect, which showed, as far back as the mid-1920s, that when you measure employees’ performance, they react. Sometimes they react positively, sometimes negatively. Not surprisingly, it can be difficult to predict which outcome you will find.

The Hawthorn plant manufactured telephone equipment, and was attempting to identify the best practices to increase employee output. In one study, they focused on lighting. They increased the brightness of the lights in the plant, and employee output increased. They dimmed the lights, and, surprise—the same thing happened. Employee output increased whether the lights were brightened or dimmed. It wasn’t the quality of the lighting, but the change that employees were reacting to.

Additional studies from the Hawthorne plant, referred to as the "bank wiring room studies," revealed that social forces were affecting the output. Employees were aware of the measure and how they were studied by management. As a result the employees used the measure to communicate to management that they were, in fact, doing a reasonable job. Concerns, such as layoffs based on the measures, affected employee output. In these studies, employees attempted to communicate to management that they were working in a steady and reasonable manner.  

These studies have made it clear that measurement does not cause efficiency —there is always an underlying and very human mechanism that is a reaction to measurement. The reaction may or may not be improvement. These very human processes hinge on reactions to the measures that include communication and messaging, image management, and social interaction.

In day-to-day work environments, we continue to ignore what studies published 90 years ago clearly demonstrated. Many attempt to improve performance with measures, without giving adequate thought to managing the underlying human processes that will lead to improvement. 

Consider the testing mania present in American public schools.  These schools are expected to improve, largely because they are being measured and expected to improve. There is a blind belief in accountability. 

Similarly, consider the 97% of organizations that have performance appraisal systems that attempt to provide feedback to individuals in the hope that it will automatically improve performance, learning, and motivation.

Unknown to most, the ghost of Fredrick Taylor is alive in many of our modern organizations. Blind beliefs in simplistic ideas about measurement continue. We need to use measurement to increase efficiency, but a naive belief that productivity will increase without addressing the very human element of work is defeating the purpose, and the effectiveness, of the measures.

If you have examples of simplistic thinking behind employee performance measurement, I’d love to hear your story.

Charley Morrow