Metrics, Trust and Job Performance

This is a cautionary tale. A common thread ties together the work of Frederick Taylor and the pioneers of scientific management at the beginning of the 20th century, the industrial engineers who followed with methods engineering, and now the metrics and analytics specialists. The thread is the idea that if human work efforts are analyzed, tracked and systematized, then productivity, quality and efficiency can be improved. 

Now it’s happening to Uber drivers. As reported in Harvard Business Review, the company developed an algorithm and an app that instructs drivers which passengers to pick up and which routes to take. For Uber, as for the scientific management experts, the goal was standardization of the way workers did their jobs. Management was exerting control. The message was clear—workers cannot be trusted.

In direct contrast, W. Edwards Deming, the Total Quality Management guru, also focused on tracking and acting on metrics, but he trusted front line workers to address quality issues. His “14 points for management” are the keys to high performance.

Deming was followed in the early 1990s by Michael Hammer and his process reengineering movement.  He believed the power to make decisions should be given to the people performing the process. That opened the door to employee empowerment and the goal of raising performance levels. That decade saw a revolution in the work management paradigm.

Uber drivers, according to the HBR article, are angry and feel disempowered. They are reacting “by identifying clever ways to work around [the ap].” As the Yankees Yogi Berra once said, that’s “deja vu all over again.”

Unfortunately for government, the still common, bureaucratic approach to work management was solidly entrenched long before Deming and Hammer. Now it’s reflected in the use of metrics, analytics and AI. This increasingly important function is organizationally and mentally miles from front line workers. 

The Cost of Distrust

Multiple studies in different cultures confirm the correlation between worker empowerment—read “regain control”—job satisfaction and performance. Compared with low trust companies, high trust organizations report 50% higher productivity, 13% fewer sick days, 76% higher engagement, 40% less burnout and 74% less stress. When employees are not involved in decisions that affect them, they take steps to thwart, impede or, as with Uber drivers, work around management’s actions. Simply stated, high trust organizations are a better, more productive place to work.

Looking back, Taylorism, as it’s called, triggered a level of distrust that permeated management-worker relations for a century. Initially it impacted the work environment in manual operations but carried over to white collar working relationships when those jobs mushroomed after World War II. This was the era pictured in the 1960 movie “The Apartment” with rows of white-collar workers sitting at desks. Management made all important decisions. 

That phenomenon naturally carried over to government. Workers did the work and management did the thinking. The focus on an "acceptable level of competence" is still reflected in law and on the Office of Personnel Management website where the discussion of performance management starts with, “Developing Performance Standards.” That’s Taylorism. The cost in underutilized talent is incalculable. 

It’s not government, but I lived with that philosophy years ago with a summer job building truck tires in a BFGoodrich factory. Several of the men working with me had been doing the job for two decades or more. We worked under an industrial engineered system that specified productivity quotas. It was obvious my co-workers knew how to beat the system. One gentleman in his early sixties was able to finish the work expected to take eight hours in under four. After six weeks or so I met my quota in under six hours. Then we were free to sit around and read a book, play cards with co-workers, etc. If they had allowed us to clock out and leave, I am sure I could have satisfied the quota more quickly. 

The lesson that stayed with me and has been reinforced repeatedly is that failing to tap worker capabilities to solve job-related problems is always costly. My co-workers knew everything about assembling truck tires but the supervisors never asked for their help. Instead, it was a worker-versus-management environment. The industrial engineers are comparable to the analytics specialists today. 

Imposing top down performance standards will always be resisted. Specialists who have never performed a job cannot fully understand the factors influencing performance, especially when a job is performed at multiple work sites. It sends the message that employee ideas have no value. 

Metrics and Analytics are Tools

Manufacturing operations are straightforward—raw materials are made into products. Each day is the same. Machines typically control output and productivity. Metrics are basic to monitoring performance.

There are agencies with multiple locations organized around similar, ongoing operations—prisons, hospitals, national parks, air traffic centers. But in contrast to factories, the operations differ in size, ‘“customers” served, and other characteristics. At each work site, local labor market dynamics, worker 

demographics, and local management affect talent management. Metrics and analytics too often ignore the “soft” issues that influence employee performance.

Another significant difference is the constraint that is the civil service system. It imposes an inflexible work management system that precludes timely responses to new operating problems. The system focus is on compliance, not performance. There are no incentives for improved results.

It’s always possible to develop metrics and generate performance data. Software to track performance and support statistical analyses is readily available. A core issue is that all analyses should be explainable and intuitively valid. Few people, however, have the background to understand the statistical methods or what the analyses generate. That makes it difficult to begin questioning or refuting analytics and algorithms, and adds to the distrust. 

Agencies that rely on metrics to define performance standards miss an opportunity. In large organizations, the mental distance of the headquarters analytics staff from front line workers is a barrier to collaboration and cooperation. It would be far more productive for the analysts to work with the workers and their supervisors to identify opportunities to improve performance.

Building a High Trust Work Environment

A study by the American Psychological Association showed that supervisor support is key to increasing a work group’s shared sense of trust. Employees who feel supported “were more than twice as likely to report being satisfied with their job, valued by their employer, and willing to recommend their organization as a good place to work.”

Proven practices to show support and build a sense of trust include:

• Recognize goal/project achievement in gatherings of peers 
• Assign difficult but fully achievable goals or tasks
• Empower employees to manage jobs or projects in their own way
• Give employees discretion to set priorities
• Share organizational goals, operating plans, and progress frequently
• Plan events where employees can socialize and build relationships
• Show support for professional growth and career progress.

For government, nothing on the list requires legislative change, nothing involves significant costs. My exposure to government work environments is broad but somewhat superficial, however, low levels of trust appear to be common. Work management practices appear to be antithetical to the APA conclusions.

At its core, this is a leadership issue. Where distrust exists, it’s incumbent on leaders to advocate for the changes to assess and initiate corrective actions. That was the point of another HBR article, “If Employees Don’t Trust You, It’s Up to You to Fix It.”