Process Design Appraisal Research from the Artige Company

Statistical Process Control (SPC) is the main tool of the quality control engineer that is being driven by a TQM philosophy. Although the term "process control" is associated at the present time with automation and computer driven devices, process control is a general term that relates to how a process is operated, with the thought being the process must run within known or given bounds. SPC is the application of statistical analysis to monitoring and controlling a process. It provides the unassailable information that states whether a process is in control or not, and when out of control, by how much. In order to do this, the process must be analyzed to determine what parameters can be monitored in numerical terms, such that statically analysis can then be applied to the data collected. Since numbers are being worked with, SPC readily lends itself to automaton, which is why process control is frequently associated with computer devices. As a side note, SPC relies upon the variable named sigma, with is a unit of deviation for the process. Back in the good old days, plus/minus three sigma was satisfactory, but currently manufacturers have a fascination with six units of sigma.

The original implementation of this visually-oriented system was to create an analog alarm in real time that recognized process difficulties in manual production systems. That implementation was focused on work-in-progress (WIP), where sequential production was limited to single units, and work on successive steps could not progress until the previous step was completed. Areas are set aside and marked off in highly visual means (florescent tape or painted areas on worktables). When the stock or WIP piece was not [obviously] present in its dedicated area when required, the worker would notify the team leader and ask the downstream worker for the cause in a non-confrontational manner. The team leader would mark the cause on a visible display, adding it to the list of process items that need improvement. Since this activity is expected to occur many times in a work shift, there would be no barrier to identifying process difficulties that need to be corrected. For larger WIP items, colorful cards would be used, where their appearance in a slot or designated area would indicate whether the WIP unit was ready for its next step.

The concept of Kanban has always been available in automated systems such as those driven by process control systems. As long as sensors can collect the presence of WIP and stock, the alarming subsystems of process control systems would handle the functions that Kanban seeks to provide. For manual processes, a shop floor control system can track the flow of WIP, and automate some of the Kanban functions, such as simplifying the identification of process failures, plus the data would be collected immediately and statistics run on it and reported to management without delay.

Business Process Design Appraisal




		If you wish more information on our products or services, then please send a message to sales@artige.com or call us at (1) 717-354-5541, and one of our representatives will be happy to discuss your needs.



		A summary of this work is available here. This lengthy document is contains the research to back up the synopsis.

		Check out our new Quality Matrix to see comparisons between many of the different business process methodologies.

	Available on this page:		Overview		Background		Process Design
			Methodologies Defined		Methodologies Compared		Conclusion




Overview		One of the questions that is frequently asked of the Artige staff goes along the lines of "what is the difference between (take your choice): TQM; Quality Management; Kaizen / continuous improvement; ISO 9000; reengineering; change management / BPR; BPM; six sigma; lean manufacturing". So we decided to write up a response and post it as an article in a living document for all to read. This allows us to update the article as new OR buzzwords make their way into the business world. For those folks whose glass is half-full, we also discuss the similarities that these terms have with each other.

		This article will use the reductionist tools of scientific analysis that were derived during the scientific revolution (maker of the "paradigm shift"), to describe what a process is, and determine what properties are available for us to measure. This analysis will rely upon mathematics for modeling and natural laws of physics for constraints. To finish, we also apply Kantian principles to the selection and continuance of the constraints. In this way we can come up with an apples-to-apples comparison of the different OR methodologies. We rely upon these doctrines in an attempt to explain process design principles in a transparent manner. We feel that using cause-and-effect based methods is more practical and comprehensible than relying solely on model-based methods.



Background		Process design has its roots in operations research (OR) and statistics. It is an exercise in sequencing tasks so that resource consumption is minimized. To help us understand process design, we need to separate this concept into its smallest components. We can resort to physics to give us some definitions that we need:


				Work is change in energy, and is the same as force expended over distance.

				A task is a well-defined set of work performed over a course of time, and can be expressed as the summation of all work involved, divided by the time it took to perform all of the well-defined work.

				If the tasks in question are related to manufacturing, it is possible that a stock item is being acted upon, that the work units are being executed to modify a resource. This in turn requires a resource that can be modified as desired.

				If the tasks in question are related to the service industry, then in most cases the task of value will be focused only on force expended over distance or on energy being transferred, at the right time, and no external resources will be involved.

				Regardless of industry that the tasks are executed in, the tasks will generate output, which includes the desired end result, and possibly additional output that is generated by the designed tasks, but cannot be avoided.


		Note that the following keywords above were highlighted: work, energy, distance, task, time, stock, value, time, output and output. These are the only terms one will need to do the terminology comparison, which is important when we want an apples-to-apples comparison. Actually, for our purposes, we can consolidate the term work into task, as long as we realize that a task is a summation of work. And yes, time was repeated twice, which points out that we will be dealing with time on different levels of granularity: time to perform a task (rate) and time between tasks (lag). Also, output was repeated twice, reflecting a societal norm that designates desired output as product, separate from undesired output, typically named waste.

		Finally, we need to align our processes within a systemic view. A system is defined to be a black box that accepts an input and transforms it to create an output, with the ability to accept feedback from the process. With this definition of a system, the alignment of a process is straightforward; the inputs are energy and optionally stock, the transformation is the work, and the output is modified stock or satisfied customers that has had work applied to them.

		There is no reason to stop applying the system model with work units and tasks. One can also consider the task of designing a process as a task in itself, which means the system model applies to process design. Later on you will see that this is concept is important to the continued successful operation of a process, plus critical if one wishes to have a rational method of improving the process. This system even has a name, called change management.



Process Design		Now we can define process design using the terms set above.

		Process design is the systemic arrangement of tasks such that one can provide stock, and energy as input and expect that a product will appear in the system output, with the expectation that the arranged tasks performed transformations along the way, with the additional expectation that the output will appear at a certain time, and that the process operator can modify the output over time based on the system feedback.

		Note that this definition does not make any claim as to effectiveness or efficiency of the process. It just states the definition using terms where we can now compare different methods of designing a process up and running it, and see how one compares to another. So we should be able to modify the inputs and the transformations and expect different outputs to appear, possible at different points in time. It also points out that the system itself may change the output over time, as feedback plays into the picture.

		We should also provide a definition for Operations Research (OR) and to provide the context that we like to use it in. OR is the application of mathematical models to assist in making decisions. OR is typically associated with Decision Science. OR is most frequently [perhaps always] used for making decisions in a business environment. These decisions can be tactical or strategic, with the idea that a benefit needs to be maximized for a given set of constraints. As will be shown later on, the decision is being applied to the operating parameters of a system, and the system can be a primary (tactical) or secondary (strategic) type.

		When the term Operations Research is used in this article, we are referring to the fact that a system is being designed and operated in a deterministic manner, with the desire to maximize benefit. This is the intrinsic meaning behind OR. This can be shown by taking the opposite tack, that operations research would not be used to make decisions, neither operational nor structural, in an ad hoc manner. For such a scenario, maximum benefit would only be a result by chance. So for us, OR means not only a set of mathematical models, but also the purposeful application of those models to design and operate systems for maximum benefit, that OR always means we are dealing with systems.

		This analysis relies upon Kantian Principles to assist in the design of the system. Natural laws in themselves do not provide enough information to construct or operate a system. A set of guidelines is then required to determine if the selected products and constraints are viable. Kant provided some simple and consistent dogma that works well with system design. It also covers the lifecycle of the products and constraints. Especially pertinent to the short sighted and corrupt business practitioner, Kantian principles do not allow one to modify the constraints at whim.

		Now why the focus on physics and mathematics? In order to generate the greatest amount of benefit (second order, i.e.: maximum maximized benefits), we need to secure the most favorable constraints. This can only be accomplished by removing constraints to the point that only natural laws hinder process operation and design. This will also assist us in providing the apples-to-apples explanation that will overcome resistance to important concepts that have been tarnished through overuse in business press. This abuse is colloquially known as "management du jour", which has poisoned the minds of managers to the point that they would avoid deploying productive process design practices.

		So what would be an example of an unnatural constraint? Take for instance the Monday morning production planning meeting. Why is it held every Monday? The test we like to apply when analyzing a process is this: what natural law is this process being constrained by? Weekly meetings are an example of periodicity, where an event occurs every 168 hours. As far as we have been able to determine, there is not a single physical law that runs at a period of 168 hours. So weekly meetings are an example of a process being constrained by an unnatural law, and open for improvement. It is a rare occasion that we find a natural law dictating the constraint, instead driven by behaviors, which means there is typically an opportunity to increase benefits for every process we have come across. This also points out that bureaucratic organizations are operating under inherently unnatural constraints, but you knew that already. If you are wondering what a perfect process looks like, see the toothbrush factory example below.

A Perfect Process		The Toothbrush Exercise

		We have a favorite example to explain how the parameters interact with each other in process design. Toothbrushes are frequently made from Nylon, made in multiple steps, and more than likely, from multiple manufacturing sites. If one were to disassociate a Nylon toothbrush into its smallest possible components, one would need to supply water, organic compounds and energy. We can represent those organic compounds with the time-honored "lump of coal". So, if we were to design the process to manufacture a toothbrush, we would go to the factory store and purchased a toothbrush black box. We would then plug it into an electrical outlet, place a lump of coal inside the box, pour water through the hole in the top of the box and press the on button. After a period of time (listed on the black box nameplate) we should expect a brand new toothbrush to be spit out. Depending in the model of the black box, we might have been able to provide additional input as to size and color of the toothbrush.

		This toothbrush process took into account all aspects of the process, including waste streams. The only waste would be a result of natural constraints, as discussed previously. A toothbrush would be manufactured as soon as possible, with no time lost or wasted. Only the exact amount of materials required would be consumed, and no external waste would be generated. The process was purposefully designed as such, and any waste possible would have been reincorporated into the toothbrush. The process will not incur any economic cost beyond that required to provide the resources, operate the machine and deliver the product. This process should appease our green friends and financiers simultaneously. Shows that there is no natural law to prevent a process from being a designed environmentally friendly.

		If this exercise sounds silly, it is because you have not yet been to the factory store. This store is hard to find, but its existence is essentially what Deming, Juran, Crosby, Schonberger and Hammer have been touting all these years. The easy way to design a process is to use what is familiar and readily at hand. However, process design has tradeoffs, and the tradeoffs are directly associated with the interaction of the physical parameters listed in the previous section. So it is most likely the case that a simple ad hoc process design will not be the minimum essential design, rather will incur lost efficiencies, consume extra inputs and may even result in undesirable outputs. On top of that, if the discipline to operate the process properly and monitor the feedback is not provided, then the transformations will change over time, resulting in outputs that change over time, perhaps not to your liking.

		Today a toothbrush is made from Nylon fiber, which was purchased from a fiber supplier who probably spun the fiber from Nylon chips, which in turn were purchased from a polymer company that fabricated the chips from organic acids purchased from a chemical company who in turn manufactured the acid from some petroleum products purchased from a refinery, who had to purchase crude oil from an oil field. Then there is the handle, which was molded ...

		You get the drift by now, lots and lots of steps, even lots of parties that need to communicate with each other. As a matter of fact, most manufacturing processes involve many human interactions, just from the point of trading the stock from one company to another, let alone the operational aspects of any automated systems. This is where the supply chain comes into play, and boy do they bring lots of processes right there.



Methodologies Defined		Now we need to know what these OR related process design methodologies are, so we can see what parameters are most important to them. Below are the definitions that we use at the Artige Company. They are listed in the chronological order that they have appeared in the public domain. Note that your mileage may vary, and no one has any official definitions for all of these methodologies. Also note that the definitions provided below are not entirely descriptive of every facet of each methodology. Rather, the succinct points were extracted, focusing on the physical parameters listed above, to enable the apples to apples comparison later on.


				TQM (Total Quality Management) / TQC (Total Quality Control)

				This is the foundation of the quality methodologies, as conceived by Deming. However, Deming never wanted to codify quality-oriented practices under a theory. He was more interested in the practice end of things. So that left a vacuum that others have rushed in to fill. This means that there is no true body of TQM theory that will increase the quality of products delivered. Instead, Deming left us with his "14 Points of Management", a toolkit that lists all of the concepts that have to do with quality one must be aware of when designing processes. A description of what a quality organization is, not how to accomplish quality.

				Strange though that at face value those fourteen points left by the quality practitioner seem to be more philosophy than a framework or methodology in which one can design a quality business process. Each one of those points can give rise to a framework and practice in themselves, and there is no fixed manner in which those frameworks can be expressed. So it seems that TQM is readily left open to the reader's interpretation.

				TQM includes both an empirical component associated with statistics, and an explanatory component that is associated with management, of both people and processes. The terms "hard" and "soft" are commonly used to represent these two components. TQM brought recognition to the fact that tasks can be categorized as value adding or not. The obvious corollary is that non-value adding tasks would be eliminated and the value adding ones improved. Many process design and operation tools have been highlighted in TQM, such as statistical process control, Kanban, and flexible organization, just to name the tip of the iceberg. Some of these TQM components are explained in detail below:


				All of the methodologies listed below can claim a heritage to Deming and TQM. One can pick and choose any of the below, and even others not listed, and claim to be a TQM organization. The fact that these separate methodologies exist might indicate that establishing and maintaining a TQM organization is hard work, and possibly financially unattainable.

				Quality Management (if you saw it, that must have been it)

				We like to use this category as a catch-all for the quality methods that are not founded by obvious cause and effect methods. That is, those methods that are driven more by personalities than particulars. We also like to lump in the traditional quality control methods, as it can be shown that the quality management practices are a result of the original "measure and reject" philosophy of quality control.

				Quality management is based upon heuristic and ad hoc methods, based on previous experience. Note that there is nothing wrong with that, and it can be deployed very successfully. It is just that certain individuals like a predictable and deterministic methodology, and an ad hoc method does not fit such a constraint. The quality management methodology has its roots in traditional quality control in production, where conformance to requirements is the major point of interest.

				We use a waterfall method to explain where quality management comes from, which is based upon quality control and quality assurance. This is described by the three definitions below:


				This category of process design is extremely interesting, in that it is the reason why all of the other methodologies listed in this analysis on this page have come about. We cannot emphasize more clearly, "no one set of rules exists that one can deploy to improve a process". All of the methodologies listed in this report are an attempt to overcome this shortcoming. Since there is no cause and effect relationship that can point one in the one and only direction to improve their processes, all of these other processes are acceptable, and can never be proven invalid or unsound in every case. All that one can do is determine which of these processes fits one's way of operating and still be able to meet the criteria set by the customer.

				Since this is the methodology that brings up the point that there is no cause-and-effect method to improve processes, we like to use quality management as the moniker to point to the ad hoc and people-oriented methods. That is because as the quality improvement concept advanced, this method would be the first that recognized human intervention was needed to improve process outcome. Most of these process improvement ideas came about in an ad hoc manner, just from trial-and-error and previous experience. Since previous experience is such a great factor in promoting prevention and improvement, the practitioners of quality management will always be focused upon people-oriented methods that pull out best practices from the experienced workers.

				Quality management is also where that concept of announcing the company financials to production workers came about in a big way. The idea here is that education and knowledge of the workforce is important, in order for them to understand where their paycheck comes from, and that they have an input in the processes that allow them to maintain their cash flow. This focus on financial measures then proves to the organization membership that ignoring quality holds a price for non-conformance.

				Note that the workforce education component is a tricky area. There are two facets to this factor. On one hand one could just announce the data and expect that the workforce will consume the data to everyone's advantage. This also allows certain executives to claim they are open and honest with their workforce. There is another step that is probably necessary. One will need to explain, and even tutor, what the meaning is behind the raw data. In other words, how will the workforce gain knowledge from the financial data? The first option to just announce the data opens the door for misunderstanding and mistrust, which is why the second option is preferred.

				All of the above tasks would be categorized as incremental in terms of implementing change.

				Kaizen / Continuous Improvement

				One translation of the term Kaizen from Japanese is "to take apart and put back together in a better way". One will immediate recognize that we have a term that relates to quality, as Kaizen relates directly to improvement. This definition even comes with a recommendation on how to accomplish the improvement. Note that the definition does not have a subject, which means that it can be applied to any matter, such as processes, resources or people.

				For all intents and purposes, Kaizen is most often associated with continuous improvement. This is the improvement methodology where small steps are taken in an attempt to improve an existing process. Feedback is collected, and then these results are analyzed. If the review of the outcome proves successful (based on predetermined metrics), the small step change is accepted. Otherwise the process is rolled back to its previous state. This is not a radical methodology. Rather, many small steps are taken to produce desired process improvements instead of one big overhaul. Kaizen is seen by many business people to be a safe way to implement quality.

				Thanks to the folks at Toyota, the term Kaizen has been applied in a quality context to refer to a human based approach that requires feedback from the people that are familiar with the process and interact with it. The idea here is that the workers experience and knowledge includes both positive and negative aspects of the operation and would be best to know what changes could potentially work or not. There are both advantages and disadvantages to this line of thought, which is why Kaizen is applicable in certain instances. Obviously, the presence of workers with experience is a requirement, which means high turnover facilities will not be a good candidate. It is also well positioned for manual tasks, especially in the service sector.

				Another offshoot of Kaizen is the thorough use of the employee suggestion box. The Kaizen effort relies upon employee suggestions and cross-functional teams, where empowered employees are encouraged to speak up. Without human intervention, there can be no process improvement using Kaizen. As such, employees are expected to make suggestions in order to raise the quality of the processes they are responsible for.

				ISO 9000 series

				It is one thing to claim that one's firm produces quality products, but it is another for your trading partners to know that this is true. So a bureaucracy was established where third parties could verify a company's claims of quality products. Question is, how do you measure quality? The answer is to document a firm's practices and audit the firm for compliance to its own procedures.

				This methodology involves administrating a culture of rules and documentation. A natural fit for enterprises that already operate under a bureaucratic culture. The documentation includes not only current practices, but also the methods for implementing process changes. The audit is meant to verify whether the firm follows the documented rules it wrote up. So in other words, ISO 9000 does not ensure that a product or service is has quality about it. Rather, ISO 9000 certifies that certain process were used, and provides for the manner in which the fact that these processes were used will be confirmed. The assumption is that always following the same method will deliver the same product (of quality or otherwise).

				Reengineering

				This methodology promotes major change over incremental change. Starting over is better than modifying existing processes. This notion is resorted to when one realizes that a process cannot be fixed, typically measured in financial terms, such as ROI. Reengineering would strive to replace an existing toothbrush manufacturing line with the previously exemplified toothbrush black box. As such, process design and innovation are stressed, along with simplification. This methodology stresses revolution over appeasement, so radical change is a necessity. One interesting corollary of this methodology is that it suggests a means to its end; the use of Information Technology in the reengineered processes.

				Change Management / BPR (Business Process Reengineering)

				First thing to note about change management methodology is that (unless otherwise noted) it is organizational. On occasion one will actually encounter the modifier "organizational" in front of the phrase "change management", but most often not. So this methodology deals with the soft components of TQM, and how one modify an organization to accept changes in processes, and then to have the organization accept that the processes will change continuously over time. Change management is the over-encompassing program that implements a culture of quality in an organization, setting a strategy of change in place. BPR provides one set of structured methodologies to implement the desired changes, the tactical tools to make the changes happen.

				BPM (Business Process Management)

				It is very interesting that BPM is compared to TQM, et al, and shows that there is confusion in this topic area. A more pointed question would be, "how does BPM compare with ERP (Enterprise Resource Planning) systems?" One could ask this since BPM is concerned with automating business processes or tasks. The tasks in this case are entirely service oriented (not manufacturing) and are automated through the use of IT (Information Technology).

				However, the level of automation in BPM is complex, as that compared to traditional IT systems, where a monolithic application was written to cover a single business process, no matter how big or small the task at hand being automated was. BPM requires that the automation be provided at the same level of granularity as the analysis of the business process was made. This provides two advantages over monolithic applications. First, libraries of small code modules are delivered, which provides the promise of program reuse. Second, BPM brings along the requirement of continuous change, in that it is NOT good for business processes to be static, rather the new processes must be flexible enough to handle any business situation that can be encountered. Libraries of programs offer a hope of meeting this level of automation, as long as a virtual control panel is provided, such that one can modify the automation of a business processes to match the desired physical changes that are going to be made.

				The big hint that BPM is more like ERP is that web services are being constantly touted as the best method to deploy a BPM. Web services are being designed (or hyped, your choice) to provide interfaces at a task level that matches the analysis of a business process. One would build an application through orchestration, which is where the output of one web service is pointed to the input of another web service. Had the ERP vendors had written their wares using web services, then the discussion of BPM would have moot, as an ERP would deliver the goals of a BPM. Of course the ERP vendors are currently scurrying to modify their wares to behave like web services.

				Lean manufacturing / Lean management / Lean thinking

				The main concern of lean manufacturing design is to eliminate waste. The main desire is to reduce the production cycle, which eliminating waste should accomplish. Lean also has a focus on retaining tasks that add value, and eliminating non-value adding tasks. Other concepts having to do with time and waste are important to lean manufacturing. Lean manufacturing is normally driven by customer demand. This brings up the point about what the driver of a business process should be. The two concepts are push and pull. Most concepts of lean involve a pull scenario. This is in comparison to the "traditional", "out-of-date", or "old-fashioned" push scenario. In the good old days companies manufactured to stock, filling warehouses with product that marketing was responsible for emptying out. The push method involves carrying costs and results in various types of waste, especially as the product lifetime came to an end.

				In a pure pull scenario, the customer demands the product, and the manufacturer creates or delivers the desired product at the moment the demand signal is received. Based on today's technology, the marketing department closely monitors the customer' needs, or the customers themselves can directly make their own demands, so the firm is able to react very quickly to market conditions. Note the word "react". It is a term that lean and TQM aficionados would like to eliminate from business vocabulary, as it foreshadows the waste to come. To accommodate pulling, minimal amounts of work-in-progress and inventory will be desired in the process design, otherwise there must be additional steps that are adding delay, which will result in waste being generated.

				When considering Lean manufacturing, one also has to take into account the concept of flow, which is driven by a production beat, that being called "takt time". Flow reinforces the notion that lean manufacturing requires constancy and cannot tolerate interruptions, otherwise additional amounts of waste will be generated. The term "takt time" describes the average amount of time it takes to manufacture a product or deliver a service, expressed in terms of a cycle. In other words, one might be able to manufacture one unit of a certain part in 120 seconds. However, if one needed to manufacture 2000 units of the part, will it still take 120 seconds per part? Takt time takes into account the flow of production, and requires that a process to run at a consistent rate, to the constant beat of a production clock. The concept of takt time recognizes that many business processes need to run at a consistent rate in order to maintain the highest quality and still deliver product at a particular volume. With takt time, one can visibly see when a problem might be brewing. If the production rate becomes erratic and inconsistent, or changes from a given norm, then some aspect of the process has failed. However, a period of erratic production may occur when the takt time period was purposefully altered.

				This last concept of purposefully altering takt time helps explain an inconsistency that one might think exists at first glance between pull-driven manufacturing and takt time. If a lean manufacturing process was based upon a consistent, never-varying takt time, how can it deal with changes in demand? The answer lies in the fact that a takt time system does not react to every demand whim on a first order basis. Rather, a second order function is used, that watches the rate of change in demand. This rate change will manifest itself in a change to the takt time period of the production line. The takt time change may result in some waste (time, resources, cost, depending on the situation). In the end the total cost of operating the production line with a takt time is supposed to compensate for the waste that may occur with individual takt time period changes.

				One will see the terms of lean manufacturing, lean management and lean thinking used interchangeably. From what we have found, there is no difference between these terms. They are all driven by the same methodology of cutting waste. The vast majority of lean manufacturing implementations have been applied to manufacturing, but there is no reason why it could not also be applied to service processes. Lean manufacturing would be considered incremental in the rate of change being applied.

				Six sigma

				Six Sigma has come to mean two things. First, it is a focal point or slogan used as a means to coach a company into improving its performance. For example, one firm's Six Sigma program is "a highly disciplined process that helps us focus on developing and delivering near-perfect products and services". Second, Six Sigma is a designation for a regulated program that a firm might elect to use to establish a quality management system in an effort to improve the quality of products produced or services delivered, and then desires to maintain that improved level of performance. The latter definition is the one referenced most often in the popular business press.

				For both definitions, Six Sigma draws upon the general system theory and relies heavily upon statistics, especially statistical process control (SPC), and requires quantitative parameters that can be measured on an on-going basis if it is deployed as a quality management system. This methodology utilizes traditional process control at its best, making Lord Kelvin proud. Process control is the practice of operating a system, measuring externally available parameters, and modifying the process based upon the measurements. The calculations are not random, but based upon statistics, especially standard deviations.

				Actually, Six Sigma gets its name from the table of probabilities for the normal (Gaussian) distribution that is used in many statistical calculations. The standard deviation variable is typically symbolized by the small Greek letter sigma. A standard deviation in this context is the amount of the population of samples that are expected to be perfect. The higher the standard deviation, the fewer rejects are expected. The amount becomes exponentially smaller as the number of standard deviations increases, which indicates that it will be more difficult to maintain a process within higher levels of standard deviation.

				Back in the good old days when SPC was common practice, a process was considered to be in control when it ran with +/- three standard deviations, or three sigma. Note that number sums up to six total standard deviations. Today, one is not satisfied unless the process runs within six-sigma deviation, leaving little room for error or defects. To give some numbers to these sigma values, one could consider the number of defects one could expect in the different scenarios. For three sigma, one could expect 2.6 defects per thousand units. For six sigma, the rate would be one half defect per billion units. However, there is an additional factor that needs to be taken into account, that of the drift in the process being measured. SPC takes that into consideration, so the typical defect rate realized with six sigma processes increases to 3.4 defects per million units. Note that these values are typical, and a properly run SPC regime will measure the true defect rates.

				As you can see from the previous paragraph, it is possible to deploy a Six Sigma program with steadiness and purpose. However, SPC is only one portion of a Six Sigma program. Essentially, Six-Sigma extends the process control concepts to process design and improvement. It requires that one take a system view of the business or manufacturing processes and treat them in a systemic manner. An acronym associated with Six Sigma is DMAIC, which stands for the continuous improvement process of Define, Measure, Analyze, Improve and Control. This is a circular process, where the results of the first pass are used to run the second iteration.

				The hardest part of Six Sigma is defining the system that describes the business process. It is completely up to the business or process owner to select the best places for splitting an enterprise into monitorable systems. The first two parameters of Define and Measure are the numerically manageable parameters. Metrics and goals need to be defined, seeking out those that can be measured and consistently reported upon, that reflect upon some sort of process output. As the process is operated, process measurements are collected and recorded on a periodic basis.

				The final three parameters of Analyze, Improve and Control act upon the metrics that were recorded, and are of a more qualitative nature. Here one compares the results against the self-determined boundary conditions and goals. The process is investigated when the boundary conditions are exceeded, and problem solving is engaged in an attempt to determine what went wrong and what could be done to improve the process. The metrics also allow for one to be proactive, and start problem solving based on trends that are observed before the boundary conditions are crossed. Main point is that the DMAIC process is never halted, otherwise complacency will set in.

				The Six-Sigma methodology requires human intervention, as this process occurs around and about the business processes. The data collection aspect can be automated, but does not have to be. The analysis and improvement aspects cannot be automated at this time. The requirement for human intervention, along with its inefficiencies, brings along indirect issues, such as group dynamics and process ownership. To address these inescapable issues, many Six-Sigma methodologies incorporate personnel practices, summarized through the use of mentoring and granting of titles to the practitioners, based upon colored belts.

				Note that there is difference of opinion on the effectiveness of hierarchical organizations, and quality organizations typically run flat. On top of that the notion that the practitioners of Six-Sigma are limited by the level of expertise that they possess and are only able to draw upon is curious for a quality organization to pursue. Nonetheless, discipline is strongly promoted, as the tighter the control limits (higher the number in front of sigma), the more tedious the effort to maintain control will be. For the most part, Six-Sigma is an incremental methodology.

		After reviewing the popular process design methodologies, we see that not all of them are true methodologies that can be compared. The true methodologies that we can compare are listed in the summary below:

TQM	NO	List of tools available to the quality company
Quality Management	YES	Based upon Quality Control
Kaizen	YES	Small improvements to processes
ISO 9000	YES	Process that are audited
Reengineering	YES	Process replacement and elimination
Change Management	YES	Method to institutionalize change of processes in an organization
BPM	NO	Just tools to create enterprise applications
Lean Manufacturing	YES	Process design to eliminate waste
Six Sigma	YES	Process control applied to process operation, which drives changes to process design




Methodologies Compared		So how do these process methodologies compare to each other? We categorize each methodology by the physical parameters we listed at the beginning of this article during the background discussion. We will first list our definitions of these parameters, so you can understand how the comparisons are being analyzed.

		Primary Parameter Definitions:

		Energy		A resource that provides the wherewithal to act or to perform an action. A common synonym would be power. Nothing can happen or move without energy. Our concern is the amount of energy consumed in the process being designed.

		Distance		Physically this refers to the difference in space between two points, where an item is located on one position and there is a desire to have it appear in another position. Temporally it represents the difference between actions that have, or will occur, such as different levels of a state in a state machine.

		Work		The combination of energy that was used to move an item over a distance, or change states in a state machine. Stated another way, work is the effort to undertake an action. The focus will be on measure the amount of work a certain action requires, in order to compare similar actions and see which one expends less energy or uses less distance.

		Task		This is where one or more acts of work is executed in a fixed sequence that can be repeated as desired, in the exact identical order, with each sequence yielding the exact same desired end result. That end result may be a physically modified item or an act that satisfies an entity.

		Stock		This would cover any resources that a task requires as input to accomplish its actions successfully.

		Product		This would cover the desired end results of the process, and is the reason the process was designed in the first place.

		Waste		Undesired artifacts created by the process, while the succeeding work actions are executed to create the product. It has been accepted that there is some sort of correlation between the generation of waste and the consumption of stock and energy, which makes waste doubly undesirable.

		Rate		The inverse of the amount of time a task will consume as it creates the desired end product. The smaller the rate, the faster the task is being executed, as fewer units of time is being consumed, which is desired.

		Lag		The amount of time between work acts within a given process. The larger the amount of lag, the more time units that are being consumed.

		Value		The measurement in a product of a combination of desirability and the amount of success in accomplishing the desirability. As desirability is in the eye of the beholder, there is no subjective measure available. In other words, did executing the work acts in your process that consumed energy, stock and time create the desired product with minimal waste?

		You might have noticed that the process design methodology definitions listed in the previous section focused on some secondary, or indirect, considerations that are not related to physics. For the most part these secondary considerations are related to the need to change the processes. Without taking these secondary considerations into account, many of the above process design methodologies could not be differentiated. These secondary terms are defined below.

		Secondary Parameter Definitions:

		Focus		The definition of TQM pointed out that there are hard and soft components that need to be addressed when a process is designed. For whatever reason, certain entities like to focus on one type of component or another, to the point that different methodologies have evolved that will concentrate on one component over another. There is no advantage to focus on one component over another, and concentrating on one component over another is not going to get one the most effective process possible.

		Intensity of change		An important consideration for the process methodologies is not only the design of the process, but its ongoing management. The process methodology definitions presented above frequently cited whether the process changes were incremental, implemented over time, or a complete overhaul, replacing one process with another, as many times as required. This point addresses the level of effort and energy the process owners will need to provide to operate their enterprise using the selected process methodology. There is no metric to determine what intensity of change is best. The absolute indicates that any waste is not desired.

		Strictness		A process can be operated or changed within a set of criteria laid down ahead of time as a set of rules. The closer the rules are followed, the stricter the operating or change process is. On the other end of the spectrum, an ad hoc approach can be used. By definition, if process rules are not documented or communicated by rote, then the process is being run on an ad hoc basis. There is no metric to determine what level of strictness is best. Since it is secondary, waste may or may not be affected by strictness in rules.

		Automation		Does the process require human intervention in order to operate, or need human monitoring to determine that a change to the process is required? A high level of automation will not require much human intervention, not all processes can be automated, nor do all methodologies promote automation. Automation would normally address the hard components of TQM.

		Now we can perform the comparisons for each of the process definition methodologies, using the physical parameters listed above. The analysis for the primary parameters is listed first below.

		Energy, Distance, Work, Task, Stock, Waste, Lag:

		Quality Management		Seeks to minimize, through prevention and predictions if possible.

		Kaizen		Seeks to minimize, through iterative observations.

		ISO 9000		Parameters to measure

		Reengineering		Seeks to eliminate

		Change Management		Inconsequential

		Lean Manufacturing		Seeks to minimize

		Six Sigma		Parameters to measure, seeks to maintain status quo.


		Product:

		Quality Management		The requirements that need to be checked for conformance.

		Kaizen		A parameter to measure

		ISO 9000		A parameter to measure

		Reengineering		A goal to aim for

		Change Management		Inconsequential

		Lean Manufacturing		Demand drives design and operation of processes

		Six Sigma		A parameter to measure


		Rate:

		Quality Management		Seeks to minimize

		Kaizen		Seeks to minimize

		ISO 9000		Inconsequential

		Reengineering		Seeks to minimize

		Change Management		Inconsequential

		Lean Manufacturing		Seeks to minimize

		Six Sigma		A parameter to measure


		Value:

		Quality Management		Requirements reflects desirability, so they will be checked for conformance.

		Kaizen		Undesirable results are noted, in hopes of remediation.

		ISO 9000		Inconsequential

		Reengineering		Process is not tolerated if results are not desirable.

		Change Management		Inconsequential

		Lean Manufacturing		Demand reflects desirability, so it too drives design and operation of processes.

		Six Sigma		Inconsequential


		The analysis for the secondary parameters is listed below.

		Focus:

		Quality Management		Soft

		Kaizen		Soft

		ISO 9000		Hard

		Reengineering		Everything

		Change Management		Soft

		Lean Manufacturing		Soft

		Six Sigma		Hard


		Intensity of Change:

		Quality Management		Incremental

		Kaizen		Incremental

		ISO 9000		Incremental

		Reengineering		Radical overhaul, as many times as required.

		Change Management		Tweak at first, replace when all else fails.

		Lean Manufacturing		Replace as needed, or tweak if possible.

		Six Sigma		Incremental


		Strictness:

		Quality Management		Ad hoc

		Kaizen		Ad hoc

		ISO 9000		Strict

		Reengineering		Ad hoc

		Change Management		Ad hoc

		Lean Manufacturing		Strict

		Six Sigma		Strict


		Automation:

		Quality Management		Manual

		Kaizen		Manual

		ISO 9000		Auto

		Reengineering		Manual

		Change Management		Manual

		Lean Manufacturing		Auto

		Six Sigma		Auto


Similarities and Differences		Having gone through this exercise, one will notice that there seem to be many similarities between the process design methodologies. This would make sense as they are all related to the same topic of business process design methods. Actually, we demonstrated earlier that the process design methodologies all stem from TQM, which is derived from operations research and statistics. So by definition all of the process design methodologies will share time and resource attributes. It would have been a surprise if one had found that any of these methods to be completely different from each other and unrelated to process design. To better understand the similarities and see where the process design methodologies are alike or differ, we need to map out the process design attributes by grouping the methodologies. That will be accomplished in this section of the article.

		The Six Sigma, ISO 9000, and process control types of design methodologies are geared more towards trading in a single-minded, inflexible bureaucracy for a bureaucracy that is driven by statistics, in order to give it legitimacy, and a set of meta-rules that allows for the bureaucracy to handle outliers. In the end, maintaining status quo is the most important characteristic of these design methodologies.

		Quality Management and Change Management seek to modify an organization's behavior. Both methodologies take direct aim at the culture of an organization.

		If one were to select reengineering as the process design methodology, one will also need to accept and embrace innovation. Reengineering will need lots of ideas to accomplish the big bang type of radical changes that are anticipated. Experimentation and failure will be part and parcel of the typical reengineering effort.

		To better understand what this analysis has deduced, a figure was generated that maps the various process design methodologies against two axes. The vertical axis shows how much support the design methodology provides for the primary parameters vs. secondary parameters. The horizontal axis shows how aggressively the methodology is advocating change, varying from incremental to radical. The figure is presented below:


		Figure 1

		Process Design Considered Against Parameter Category vs. Level of Change

		Note how the above figure shows that the process design methodologies all emanate from TQM, which in itself emanates from operations research and statistics.


More on Primary and Secondary Parameters		By this point the analysis has pointed out many times over that a process can be measured by primary and secondary parameters. This differentiation is not haphazard, and is a suggestion of a more basic phenomenon that must be recognized and understood if one is to comprehend the nuances of systems design, even moreso if one wishes to become adept at performing system design. It is probably evident that a process can be evaluated against more than one set parameters. It just so happens that this simple fact is the basis for comprehending business process design. One must simultaneously design a process for multiple sets of criteria. Simply stated, but difficult to execute effectively. To illustrate this impression we like to use the analogy of a slippery fish. Once you think you have your hands around it, it pops out of your hands. Simultaneous design of processes at multiple levels is similar. Once you think you have met all of the criteria, a process cycle will transpire that shows otherwise. This is a great lead-in to our fish story...

		A Fish Story

		We usually bring out our fish story when we are asked to explain how it is possible to interpret models in more than one way, and also how it is that systems need to meet multiple sets of criteria. To begin with, we present two figures below, which may be familiar to persons that have been exposed to articles concerning topics on quality.



		Figure 2

		Fish State Diagram, in the "Current State"

		The diagram presented in Figure 2 above shows a number of symbols that resemble fish, inside a box, such that the fish are pointed in various directions, persuading the reader to think these fish are randomly aligned. These fish are meant to symbolize processes, based upon the transformation symbol turned on its side. The fish symbol is convenient, in that there seems to be two ends, with an assumption that one end is an input and the other end is an output. This figure of random alignment is typically used as a metaphor for the current state of processes for a given enterprise.



		Figure 3

		Fish State Diagram, in the "Ideal State"

		The diagram presented in Figure 3 above shows the same set of fish symbols, this time presented such that they are lined up in parallel rows, just about in an end-to-end fashion. Reflecting on the allegory that a fish is the symbol of a process, one might even be persuaded to consider that these aligned fish resemble properly placed processes, with one output feeding the input of another. Keep in mind that there is more than one way to interpret a model. Our contention is that these fish models are simplistic. How does one handle processes with multiple inputs and outputs with these fish diagrams? In short order we will show what makes these diagrams simplistic. But first we must place the use of the misaligned fish diagram in proper perspective, which we will do with another analogy.

		This metaphor provides an illustration that is very similar to the misaligned fish diagram, that being the messy desk syndrome. Certain individuals are uncomfortable when presented with a desk that seems to be in disarray. One can raise two questions when confronted with a messy desk. Is the user of the messy desk successful in accomplishing their goals? More than likely the answer is in the affirmative. In these instances, the user of the messy desk is "in control" and is capable of operating successfully in such an environment. The second question then would be whether the messy desk is efficient or not? That depends upon the criteria being used to evaluate the system. Why does it depend? Because there are TWO sets of criteria. One set of criteria is the one being applied by the messy desk user, with the other criteria fixed by the visitor. This means there will be more than one "correct" answer.

		If one is presented with a situation that results in an outcome that can be evaluated to different conclusions, then one is dealing with one or both of two circumstances. In one circumstance there could be multiple sets of criteria being applied. More than likely, it is the other circumstance where multiple sets of criteria could belong to different processes that are using the same transformation! If it is possible that a transformation can be used by multiple processes, then models would need to show this somehow, as it will take more than fish lined up end-to-end to show this concept. The process design practitioner will need to recognize what sets of criteria are being applied, and to what processes they belong to.

		When one is dealing with multiple sets of criteria, then one will frequently run into the "orthogonality" question. It is possible that one set of criteria have absolutely nothing to do with any other. So a process that satisfies one set of criteria is seen as successful by one process owner, but seen as a non-compliant process by other owners. One might now begin to recognize a pattern here of why some business design processes are successful to some and fail utterly for others, even within the same enterprise. There is an explanation for this, which will be presented next, and should bring hope to those that have been worn out by the many iterations of "program du jour", and have to work in dotted line environments.

		Multiple Process Mapping

		As started many times in this article, systems have inputs, outputs, transformations and feedback. That is in a sense what the fish is showing in the two above figures, although that shape is the symbol for just a transformation, turned on its side. So, for any given set of inputs and outputs of a system, any number of transformations can be applied successfully. It should be empirically apparent to people having lived some time in our civilization that there is more than one way to perform a task. No claim can be made that one and only one transformation exists for any given set of inputs and outputs.

		When a process is designed, one will be considering what systems are involved and how they will interact with each other. Often times these processes are the primary ones, those that are being driven by the primary parameters. This is especially the case when the design assignment is made without thought, a "just do it" request to satisfy a business need. Of interest to our analysis is that the individuals carrying out this assignment may not be consciously aware that there is a secondary, higher level, process in progress, and that the process being designed will be incorporated into a secondary process, if not multiple secondary processes. Consider it a part of the "just do it" mentality, the focus of the task will be on the primary process design, to the detriment of the entire enterprise, because the other, possibly several, sets of criteria are not being taken into account.

		It should become evident then that a prosperous process design regime will need to consider all of the processes that are relying upon a single transformation, no matter what level that process is running at. All of the sets of criteria must be met simultaneously in order for the enterprise that is hosting the processes to operate in a valuable manner. Once this factor is recognized, then is a straightforward task to take simultaneous sets of criteria into account. It is through the mapping of inputs and outputs between the various levels that will give one a handle on designing processes that operate within multiple sets of criteria. It is necessary to identify which criterion at a secondary level is affected by particular primary transformations, in order to determine what transformations will affect the unconscious secondary criteria. This is what the mapping between levels is meant to accomplish.

		Note that while it is straightforward to take the simultaneous sets of criteria into account through mapping, it is going to be a tedious task, as there are many transformations being assembled by process design. While it is human nature to summarize complex patterns, summarization in business process design is the first step towards purposefully introducing defects into a process. This is the price one must consider when the reductionist methods are invoked. All we can say at this point is that automation is our friend, and is why automation has seen widespread use in process design, as well as operation and evaluation of processes.

		Note too, that the actual mechanics of inter-level process a mapping is outside of the scope of this business process design appraisal. We expect that we will write a separate article on this sometime in the future.

		As an aside, we do have another fish state diagram that we use to explain the inter-level process mapping. It also takes a valiant stab at revealing how processes interact with each other within an enterprise, useful for those that enjoy fish diagrams. So if one insists on a visual model to represent the overlapping primary and secondary processes, then our Jonah state diagram would be more representative of the mapping, presented in Figure 4 below.



		Figure 4

		Fish State Diagram, in the "Jonah State"

		The diagram presented in Figure 4 above presents the system transformations in a misaligned manner once again. But more importantly, it shows that some transformations are contained within other transformations. Subsequently, our Jonah state diagram allows us to maintain the simplistic allegory of fish representing a system that certain individuals are comfortable with. Through this diagram type we can show how secondary processes will include a variety of primary processes, by showing the primary processes inside of the secondary process. A few things to keep in mind:

		-	The orientation and alignment of fish is not relevant to the interaction between processes, neither at primary nor secondary levels. Alignment of the fish is orthogonal to explaining their interaction.

		-	The fish will move constantly, which is representative of constant change in the business environment.


		This small fish inside big fish analogy is an example where primary and secondary system design will need to be performed, and that they are dealing with different systems. It also shows an example of managers at different levels dealing with different tasks. The caretaker manager will want to insure that each little fish is interacting with the other as designed, while the strategic manager will be moving little fish from one big fish to another.

		The other item of interest is how the little fish interact with the big fish. This is where we explain why the messy desk user is successful, and why it does not matter (orthogonality) that the fish are lined up end to end. The point of importance when systems are being designed, especially when designing a secondary level system, is the mapping of inputs and outputs from the lower level systems to the upper level system. If this is done properly then one will be successful in accomplishing their goals and all systems will be operating as best as can be expected, given the multiple sets of criteria they will be evaluated against.

		The above diagram is still not completely representative of process interaction. It is not possible to show in a two-dimensional figure how transactions can simultaneously belong to multiple processes. It is also difficult to claim which fish should be the containing, or outermost fish. Interaction between processes is not an encounter that can be delineated within a single structure. At some point there will be inputs and outputs that must "leave the page". At this point in time we are only aware of mapping tools as our only resort to comprehending the interaction of processes and our ability to work with them.



Conclusion		Looking at the comparisons above, one will note that some of the process design methodologies did not address the physical parameters (noted as being inconsequential). For the methodologies that did address the physical parameters, they treated the parameters similarly, when compared against each other. From this lack of differentiation one can distinguish that the methodologies can be grouped in two categories of primary and secondary process design. If a process design is meant to handle the primary parameters, then it will function as well as any other process design methodology that handles the primary parameters.

		So that means the differentiation between process design methodologies is either that they handle both primary and secondary process design considerations, or only secondary design considerations. There is definitely a difference between the handling of the secondary parameters between the methodologies. Problem is that the methodologies did not arrange themselves neatly within each secondary parameter. So one cannot rely solely on the secondary parameters acting as a grouping function to differentiate the process design methodologies.

		The only fact then that comes out of the parameter comparisons is that the process design methodologies can be categorized by its ability to handle primary parameters (a Boolean yes or no), and that the secondary parameters are handled in a variety of ways, which is where the major differentiation can be made.

		One additional finding from this appraisal was that process design must take into account primary and secondary parameters. A successful process design will allow for multiple sets of criteria to be satisfied simultaneously. This will take into account the primary needs of the task, while satisfying the higher level processes, such as those driving the direction of the business, which are continuously causing the change in the business environment.



Summary		The process design methodologies can be grouped by their being able to handle primary parameters or not. For those methodologies that do handle primary parameters of work, energy, distance, task, rate, stock, value, lag, product and waste, there is no discernable difference between their ability to create processes that can manage them effectively. All of the process methodologies handle the secondary parameters differently. The reason to select one process design methodology over another will then have to do with organizational culture and fit, as the methodologies can be matched to existing cultures. Of course, one can select a process design methodology that does not match the current organizational culture, perhaps with the hope that a different culture can be instilled. As we like to say in operations research, that's one way to do it. Plus, there is no correct answer, although we recommend going with the methodology that best fits your culture.

		Another point is that one can deploy multiple process design methodologies. This would be required if one were to select a methodology that only addressed the secondary parameters.




		Return to Services Division page	Navigate to home page

Legal

Last updated:
4-July-2005 18:25z