Dealing with Labor
A considerable contributing factor in the engine that can drive toward effective asset performance management is a fundamental change in mindset and culture that is a holdover from the industrial revolution. Changing such a mindset requires that we first understand what it is and where it originated. As industrialization started to ramp up in North America and Western Europe, one resource that was abundant was people to work the plants and factories. Unfortunately, the vast majority of the available human resources were uneducated and unskilled. From the perspective of today's culture it may be hard to relate to how uneducated these people really were.
Most could not read, write or do even basic arithmetic. This led to a huge industrial challenge - how to take advantage of such a resource. This challenge was met by Frederick Taylor, who developed an approach called Scientific Management, which focused on gaining maximum value from an uneducated workforce. In today's vernacular, Scientific Management essentially turned people into minimally functional robots, each performing a well contained and well defined function within the context of the operation of the entire plant or factory. For example, a person may have been trained to watch a gauge and keep it in a certain range. When the needle moved out of the range, the worker would turn a hand valve in one direction. When the needle moved out of range in the other direction, he turned the valve in the other direction. This person might join the workforce of the factory at 16 years old and retire 50 years later having performed that contained task his entire career. This led to the concept of a labor force in industrial companies which was so unskilled that management believed it could not be trusted to perform duties beyond menial tasks. In essence, the laborers were almost treated as a kind of industrial slave.
This view of the labor force was exacerbated with the introduction of automation technologies. In many cases, the automation technologies were developed to perform the same functions laborers had performed. For example, automatic controllers providing direct manipulation of control valves essentially were replacing the laborer who had previously been stationed at that valve. Early automation advancements may have allowed a single laborer to perform the scope of functionality that six or eight laborers had previously been doing. As computer-based automation systems were introduced, single operators may have been able to oversee functionality that single operators may have been able to oversee functionality that would have required fifty people in the past. The basic value proposition for the introduction of automation technology was typically based on headcount reductions that could be achieved. Many manufactures seem to have viewed these reductions as a double benefit to the company. First was the cost reduction for not having to pay the displaced laborers. But second was the thought that there would be less of the low-level laborers to have to manage and worry about.
The culmination of the technology replacing people trend took place in the 1980s when a number of management scientists and engineers supported a notion referred to as "lights-out manufacturing." The thought process behind this trend was that technology may have advanced to the point at which no front line workers would be required at all, and without people in the plants there would be no need to turn on the lights. This was a short-lived movement due to the fact that the technologists found they could not anticipate every possible issue or problem that may arise in a plant and that at least some number of people must be in the plant, if for nothing else, at least contingency responses.
All of this has left a residual mindset in both industrial management and engineering that frontline personnel are a necessary evil that would be eliminated if possible. This has further led to an attitude prevalent across industry that the actions and activities of these frontline laborers have to be contained to only those essential to keep the plant operating. A good example of this mindset can be found in the design approach taken to the software in industrial workstations. This software is designed around the concept of "operation by exception," which basically means that the process operator is not supposed to do anything if the process is operating in a reasonable manner (except, perhaps read the sports page). When something unexpected happens, an alarm will cause the operator to follow a predefined procedure that should bring the alarm condition under control. Once the alarm condition has been addressed, the operator goes back to the newspaper. Additionally, engineers have developed and deployed advance control and other advanced techniques designed to operate the plant better than the operators could by themselves. The attitude of protecting the plant from the frontline laborers has continued, even while the average education and skill level of the labor force has been steadily rising. I have been in control rooms in which the frontline process operators all had college educations, and were still viewed as the unskilled, uneducated laborers of the early industrial revolution.
Having worked with industrial organizations for over three decades, I have frequently heard the rejoinder that "islands of automation" are to blame for the difficulties in developing higher performing operations. Although there is certainly much truth to this, I have found that "islands of organization" within industrial companies present a much more formidable barrier to performance improvement. As industrialization took hold and grew, the complexities introduced to manufacturing businesses became very challenging. In early industrial plants the same person might operate and maintain the equipment, design and commission new production areas and even account for the business. As more complex manufacturing systems have evolved, this level of generalization is just not feasible, which has led to the era of specialization.
Professionals specialized in engineering, accounting, management, purchasing of materials and shipping of finished products while frontline labor specialized in operations and maintenance of the equipment. This naturally resulted in separation of departments by function which, in turn, led to organizational silos. The development of specialists was necessary to the operation of the increasingly complex plants, but the development of organizational silos resulted in huge inefficiencies across organizations. Today it is not unusual to find maintenance departments that never directly communicate with operations or production teams. In some organizations they don't even like or trust each other. Adding to this, many IT organizations don't like or trust engineering, and the feelings are mutual. And nobody seems to get along well with accounting.
In many cases, the performance measures used to evaluate the performance of one group are in direct conflict with those of a second group. For example, maintenance teams are often measured on the availability of critical equipment assets while operators are measured on the utilization of the assets. Asset availability and asset utilization are inverse functions. That is, to increase utilization often requires the sacrifice of some availability and vice versa. Under this scenario, it is no wonder operations and maintenance teams seldom get along well.
As industry has invested huge amounts of capital into efficiency-increasing automation and information technologies, organizational silos have worked to destroy any potential value that may have been created by the technology. I was recently attending an industrial conference in which an engineer estimated that over 80% of all advanced control that has been implemented in industrial plants has been turned off by the process operators because the operators don't trust it. If engineering and operations had a better working relationship, based on common goals and objectives, this might not be the case. Organizational silos have tended to sub-optimize plant performance by sub-optimizing the human performance within the plants. Perhaps it is time for industry to start moving away from long over-worn prejudices and consider using the valuable human resources more effectively to drive better plant performance.
You are probably familiar with the common adage is: "people perform to their measures." I believe that this is very true. Most people want to be evaluated positively, and if they know that measures of performance exist for which they will be held accountable, they will strive to make those measures move in the correct direction. This is true whether the measures are driving desired behaviors or not. For example, measuring maintenance on asset availability and operations on asset utilization does not encourage the cooperative behaviors most industrial leaders would like to see.
In the early periods of industrialization, prior to the many inventions that drove the industrial revolution, most shops measured performance as each product was produced. Production was so slow that accounting for the business on the basis of piecemeal production was easily achieved. Management and operators of these firms knew exactly how they were performing compared to their plan at all times. But with the introduction of tools, such as the power loom in the textile industry, the pace of production increased to the point that piecemeal accounting was no longer feasible. The result was that industrial operations compromised and began measuring the business performance through monthly accounting methods. The primary output of these systems for measuring manufacturing performance was, and in most cases today still is, the variance report. Variance reports basically report the cost per unit product made for each product produced over the past month and displays this against a previously predicted expected value, referred to as the standard cost for the product class. This information may be acceptable for reporting manufacturing performance, but it has little value in enabling the plant personnel to change their behaviors to improve the performance of the operation. The information in the variance reports is both too little (providing a broad plant-wide perspective) and too late (after the month is over) to be of any value to the people actually working to keep the plants operating.
Monthly accounting systems for reporting of manufacturing and business performance represented a compromise introduced to industry out of necessity. The tools just did not exist to measure plant performance as the plant was running. Over many years, industry got lulled into believing that monthly financial reporting was a best practice that should never be challenged. Accounting professionals earned Masters Degrees on how to do monthly accounting. Once degrees are conferred on how to do any practice, it is very challenging to ever question the validity of the practice again. Therefore, when digital computers were generally introduced into industrial operations during the 1960s and 1970s, nobody seemed to raise the question as to whether accounting and performance measurement systems might be able to be developed to account for operations as originally intended - as the products are made - in real time.
Since monthly accounting measures from in cost accounting systems proved to be fairly useless in directing the actions of the operations and maintenance teams, a number of leading industrial companies started to develop a different set of operations performance measurements to supplement the accounting systems by providing more actionable feedback to plant personnel. The measures produced by these systems are commonly referred to as key performance indicators (KPIs). These KPIs were not developed to replace the accounting measures, rather they were developed because engineers and managers did not view the measures produced in the accounting systems as adequate for directing performance and improving actions in the plant. KPIs were typically developed to measure different operational silos within plants, such as maintenance, operations and engineering. By focusing on specific functions, they tend to offer better resolution, as well as better timeliness, than accounting measures. However, by being functionally focused, they also tend to discourage cooperation between organizational groups. Even though daily measures provided a great leap forward from traditional monthly measures, frontline personnel often find daily measures too long a timeframe to offer actionable feedback. A single operator may make hundreds of specific actions each day, and an overall daily measure does not provide the timeliness for them to understand the performance impact of any specific action.
To make matters worse, KPIs tend to have little credibility with accountants, whose job it is to measure the business performance. Although many KPIs may report in monetary terms, accountants often have great difficulty reconciling the values reported though the KPIs with the values in the accounting reports. When this happens, the accounting information clearly takes precedent. I actually heard one CFO say, "If one more engineer comes to me with one more KPI telling me how much value he has created, I'll fire his In most plants there are simply too many measures for any one frontline person to deal with in real time. When working in real time environments, such as driving a car or operating a plant, ergonomic research has determined that most people can only consider up to four competing measures at a time. The question is which four measures are most appropriate for each person in the operation. This can be determined by taking the current manufacturing strategy into consideration. Dr. Thomas Vollmann developed a strategy analysis approach that can be very helpful in determining the DPMs for each person in the operation. The Vollmann Triangle diagram (See Figure 2) is helpful in understanding his approach. He points out that every plant should be working to a strategy designed to maximize the economic value of the plant output within the external and internal environment in which the plant is operating. Each manufacturing strategy should be defined by a set of actionable strategic objectives for the plant. An action plan, in which each action step is measurable, should be developed for each objective. The measures that fall out of the action steps are the strategic performance measures of the plant. These measures can be decomposed through the physical areas, units and major assets of the plant to determine the most important measures for each process unit according to the current strategy. This can then be used to prioritize the real time KPI and accounting measures for each person that impacts the performance of the operation. This information can then be presented on a performance dashboard contextualized to each person's responsibility. These are the DPMs of the frontline operators.amp;*!"
Dynamic Performance Measures
The value of an effective and comprehensive performance measurement system cannot be overstated when it's working to drive increased levels of performance from plant assets. Industry has reached the point where the performance measures that encourage the organizational silo mentality have to be abandoned in favor of measures that drive collaboration between traditionally competing functions. A new approach to performance measurement is required that combines the goodness of accounting and operational measures, provides performance measures for every person in the operation, within the time frame in which they do their job and for the same domain for which they are responsible. Such performance measures are referred to as dynamic performance measures (DPMs, See Figure 1).
The first issue that has to be addressed in developing a DPM approach is the availability of a database that provides real time input data. Fortunately, in most industrial plants, such a database is readily available in the form of plant sensors. Plant sensors continually measure physical and chemical properties, such as flow, level, temperature, pressure, speed and composition of process variables in real time. They are typically accessible by the installed automation systems and are used to monitor and control the process. Since both accounting and operational measures can be defined via equations, an experienced engineer can develop models of the equations in the automation system and determine which sensors can be used to populate the models needed to calculate the DPMs. The net result is a set of performance measures for each process unit or work cell in the plant.
In most plants there are simply too many measures for any one frontline person to deal with in real time. When working in real time environments, such as driving a car or operating a plant, ergonomic research has determined that most people can only consider up to four competing measures at a time. The question is which four measures are most appropriate for each person in the operation. This can be determined by taking the current manufacturing strategy into consideration. Dr. Thomas Vollmann developed a strategy analysis approach that can be very helpful in determining the DPMs for each person in the operation. The Vollmann Triangle diagram (See Figure 2) is helpful in understanding his approach. He points out that every plant should be working to a strategy designed to maximize the economic value of the plant output within the external and internal environment in which the plant is operating. Each manufacturing strategy should be defined by a set of actionable strategic objectives for the plant. An action plan, in which each action step is measurable, should be developed for each objective. The measures that fall out of the action steps are the strategic performance measures of the plant. These measures can be decomposed through the physical areas, units and major assets of the plant to determine the most important measures for each process unit according to the current strategy. This can then be used to prioritize the real time KPI and accounting measures for each person that impacts the performance of the operation. This information can then be presented on a performance dashboard contextualized to each person's responsibility. These are the DPMs of the frontline operators.
Developing these DPMs requires a real time computer engine that has builtin modeling capability. This is exactly what a standard automation system is. These DPMs must then be aggrandized to provide performance measures in real time for every other function within the plant. This can easily be accomplished by using a standard process historian which can also develop hourly, shift, daily, weekly and monthly accumulations of the DPMs. The availability of a comprehensive, real time, bottom to top performance measurement system provides the potential to drive improved performance in a number of ways previously unavailable to industrial operations. The basic value improvement that can be realized through better individual performance of frontline personnel, who can immediately see how their actions impact plant performance, has been proven to provide huge performance gains. However, this is only a starting point.
A New Perspective on Asset Performance Management
The availability of DPMs enables asset performance management in ways previously unavailable. As previously mentioned, traditional asset management involves operators driving the assets to maximize asset utilization and maintenance maintaining the assets to drive maximum asset availability. It is important to understand that neither asset availability, nor asset utilization, is a measure of the business objectives of any plant. Since they are inverse functions, operators and maintenance teams are frequently at odds with each other. So, in essence, traditional performance measurement systems tend to discourage cooperation and collaboration.
It's quite useful to use an analogy from the world of sports since nearly all professional sports are performance- driven. In automobile racing, the driver is analogous to the operators in industrial plants and the pit crews are analogous to the maintenance teams. In interviewing a NASCAR driver and a pit crew chief, I noticed how well they tended to cooperate. I asked them if, as is common in industrial plants, the pit crew was measured on the availability of the car and the operator measured on the utilization. They told me that although utilization and availability (or maintained state, which may be a much better measure than classic availability) are important, the primary measure of both is winning the race. I asked the pit crew chief if, upon detecting a problem with the car that might negatively impact the maintained state, he would call the car into the pit. He said, "Only if the problem means we won't win the race." Then I asked the driver if he would refuse to come into the pit if called in by the crew chief. He said, "no way, I know he is calling me in because I'll lose the race if something is not done." You see, for both parties, the primary focus is winning. And since they have a shared focus, they not only trust each other, but they cooperate extremely well. So how can we define "winning" for frontline maintenance teams and operators in industrial plants to engender the same level of cooperation and even collaboration?
Most plant management teams are measured on driving the maximum production value from the plant assets over an extended period of time. Certainly the utilization and availability of each plant asset impacts business value, but neither should be treated as the primary measure of performance of any industrial operation. The real victory in industrial plants is driving the maximum business value from each plant asset over time. If every operator and maintenance person has a primary measure based on this win, the behaviors of each will change drastically and the behavior of the plant will follow suit. Industrial companies must empower frontline teams with the information, in the form of DPMs, which will drive both collaboration and continuously improving business value from all plant assets.
Asset Performance Management (APM) driven by DPMs results in operations and maintenance working together to balance plant operations for optimal business value in all circumstances. The primary measure of both frontline teams is business value. Secondary measures for maintenance include the maintained state of the equipment and the probability of a failure over time. Secondary measures for operations include operation to maintained state and the probability of a failure over time. With all DPMs prioritized to the manufacturing strategy in place, every person in the organization will be pulling in the same direction. They will all be focused on "winning." They will all be focused on doing their part, but, even more productively, doing it within the context of the overall performance of the operation.
An interesting symmetry develops between operations teams and maintenance teams when a true asset performance management approach is taken. Both operations and maintenance have advanced in three steps with the evolution of technology in each area over time. Technology impacted operations by first providing regulatory control, followed by advanced control, then followed by process optimization. Maintenance had a similar progression from reactive, to preventive and then to predictive maintenance. As each progression was underway, the KPIs for each function were used to measure progress. The next step, asset performance management, occurs when the two frontline functions converge around new measures of performance that combine accounting and operational measures into a comprehensive, prioritized performance measurement system called DPM. This is the point at which cross silo collaboration takes hold and breakthrough levels of performance are attained.
Industry is on the verge of a major new wave in performance improvement driven by collaboration across organizational silos guided by Dynamic Performance Measures. For this new wave to really take hold, industrial management and engineering have to escape from the residue of the industrial revolution and stop thinking of operations and maintenance teams as an unskilled, uneducated labor force.
Frontline personnel are responsible for making, or losing, most industrial operations more money minute by minute than any other group in industry. It is time we start treating them as the performance managers they are by empowering them with DPMs.
On top of this, industrial management must start to break down the organizational silos that have existed in plants for decades while simultaneously preserving the specialized knowledge and capability of each team in the plant. Again, this can be achieved by empowering the teams with the correct performance measures that define the "win" for the business. When this is accomplished, the result is a new performance-generating collaborative approach to plant operation called asset performance management. Asset performance management is the industrial performance wave that is just starting to crest. Those industrial concerns that catch this wave will be the performance leaders of this new millennium.
Peter G. Martin, PhD, D. Eng., joined The Foxboro Company in the 1970's and has worked in a variety of positions in training, engineering, product planning, marketing and strategic planning. He left Foxboro to become Vice President at Intech Controls and also at Automation Research Corporation before returning to Invensys in 1996. Since his return, he has been VP of Marketing for Foxboro and Chief Marketing Officer for Invensys Manufacturing and Process Systems prior to moving into his current position, VP Strategic Ventures. He has written two books: Bottom Line Automation and Dynamic Performance Management: The Pathway to World Class Manufacturing. Dr. Martin holds multiple patents, including the patent for Dynamic Performance Measures, Real-Time Activity-Based Costing, Closed-loop business control, and Asset and Resource Modeling, which are the basis for Fortune recently naming him a Hero of U.S. Manufacturing. He was also recently named as one of the 50 Most Influential Innovators of All Time by the Instrument, Systems and Automation Society (ISA). Dr. Martin has BA and MS degrees in Mathematics, an MA degree in Administration and Management, a Master of Biblical Studies degree, and a D. Eng in Industrial Engineering and a PhD in Biblical Studies.