Context and Correction in Systems of Systems

Context

SIPOC

Since the late 1980’s quality management systems (Six Sigma, Lean Manufacturing and Business Process Management) have been using the SIPOC methodology to provide a high level overview of a specific business process, define a new processes, identify suppliers and customers internal or external to the organization and emphasizes the inputs of the business process and the business process outputs while establishing clear boundaries. (Courtnell, 2020)

SIPOC is an acronym, standing for Suppliers, Inputs, Processes, Outputs and Customers. It is thought that following this acronym from S to C (or sometimes C to S) a business process is followed from the beginning of its supply chain right through to customer delivery. A standard representation of a SIPOC is shown in Figure 1. Following this model, the supplier supplies products, or services to the process. This provides the starting trigger that kicks off the process. The process includes all the steps needed to transform the supplier inputs into the customer desired outputs. Outputs are the result; they include everything that comes out of the process but primarily are focused on whet the customer is expecting. (Antony, Singh Bhuller, Kumar, Mendibil, & Montgomery, 2012) (Shankar, 2009) The quality of the outputs and the timing of delivery are typically specified contractually.

Extending the SIPOC - HLPM

From the SIPOC a high-level process map is developed to show what actions undertaken in the process. the “high-level” process map restricted to roughly four to eight key steps regardless of whether the output is a product or service. This describes the series of steps or activities that takes inputs and transforms them into an output for the customer. To describe it in a more business manner, the process is where value is added to the input to provide a suitable output that, at a minimum, meets the customer’s need(s) and related quality requirements. The relationship between the SIPOC and the high-level process map is shown in figure 2.

From these high-level maps, the process can be further decomposed to show greater and greater detail until a process map is developed that accounts for every process step. The development of these detailed maps allows for consistent, and potentially automated processes to reduce costs by decreasing waste and rework, increasing efficiencies, and improving consistency. To remain competitive, these processes must be constantly reviewed and updated to avoid the risk of increased costs, decreased revenue, and obsolescence.

With their impact on the organization, well-defined and well monitored processes are integral to sustained success. The procedural views lack the representation of how the processes are controlled and monitored for issues. To overcome this, a switch to a system view is required.

Productive Systems

A system is a set of elements working toward a common goal by acting on inputs to produce outputs, this is not unlike the SIPOC. To create what is called a productive system, feedback loops are added to provide control mechanisms so that corrective action can be performed. (Haksever & Render, 2018) An illustrative productive service is shown in Figure 3.

The feedback loops provide the inputs and mechanisms for the control, of the system. While feedback and control may seem like a tack on to the original SIPOC model, they are however essential for the continued success of the system.

No system run forever at some point a breakdown will occur, whether it is caused by the humans or the machines associated with the system, neither of which is infallible. Human error is inevitable, and as machines wear make mistakes, machines go out of adjustment or break down. By adding in the feedback loops management can take corrective action when the system fails to work as expected (shown by the addition of the control system in figure 3). While shown as a single block the control system is controlled by varying input and produces varying outputs to provide the desired response. (Tutorials Point, 2020)

While providing control and correction to the higher-level system, it should be noted that each system can be decomposed into smaller more granular systems, building ultimately into a system of inter and intra dependent systems bound together by one or more common objective and collectively acceptable methods to achieve them. (Gharajedaghi, 2011) These interactions, interconnections, how the sequence and flow of activities, and the surrounding business rules define how the systems operate. They define how systems operate, and how the elements are the dynamic process of using elements of the structure to produce the desired functions. In a nutshell, to put it in the simplest of terms it is about understanding the box that lies between system input and system output. The better the understanding of the boxes, the easier it is to find measures of how the system is performing.

Measures

SLA

Following the approach described for processes/systems above, it is possible to use these models to determine where to inert data points. Starting with the SIPOC, it is only possible to build simple measures. We can measure inputs and outputs but armed with only this cursory knowledge of the process not much can be inferred. Looking at the graphs in figure 4 we can only state that 33 of the 602 inputs received have yet to be converted to outputs. Additional static information must be added to provide context. The added detail provides that the goal is 99%, but the measure is at 95%, 27 more inputs need to be processed to reach a threshold set at 93%, 27 more are required to reach 99% (the SLA measure) and all 33 would make the measure reach 100%. Combined with a date and time, and knowledge of the periodicity of the measure, it is possible to infer a general sense of how the overall system is operating.

Measures at this level provide an incomplete picture of the health of the system.  Without the inclusion of the added feedback loops (see figure 3), it is impossible to tell that in 63% of the measurement cycles over 20% of the inputs entered the process late, as shown in figure 5. And 21% of the time, the number of actions entering the system late exceeded 30%,

Measures at this level also do not provide enough detail to inform the system owners that 30 to 40% of the systems outputs fail to reach the standard for quality.

Feedback

By taking the productive system view, it is now possible to address the timeliness of the inputs, as well as the quality of the outputs. Controlling for both feedback measures would directly impact the efficiency and effectiveness of the system. As the inputs are reviewed for how they impact the transformation process, additional controls should be created for each (see figure 7). The ability to remove or correct for variations in input, ensures the system is adequately supplied.

While the above example focuses on the input side of the productive system, the same process should be taken for the outputs. If the output is rejected due to a quality concern, or if it is late the feedback loops should activate the controls to prevent the reoccurrence. The feedback loops, and the analytics they provide (quantitative and qualitative) serve to eliminate the gaps between observation and action. (Davenport, Harris, & Morison, 2010)

Iterative Success

It would be simplistic to assume that once the primary impacts to a system are discovered and documented, feedback loops are established and are operating to keep the system performing at optimal levels the task is complete. In truth, the act of achieving success by reducing late, incorrect, and incomplete inputs to acceptable levels dissolves that problem and transforms it to an entirely new set of concerns. (Gharajedaghi, 2011)

A system experiencing an increase in input (see figure 8, PP25) may not have the resources to handle 20 to 30% more inputs once the timeliness issue is addressed. This effectively transforms the (presumedly) external issue of input timeliness to an internal resource issue. Corrections to one area, will expose issues in another (or create them). (Goldratt, 2004)

Business process provides the context of what the system should be doing, while the addition of feedback loops provides the mechanisms to ensure it is. A solution utilized to overcome an issue can shift the problems from one area of the system to another, and even from one system to another. These self-inflicted issues “…often go undetected because …those who solved the first problem are different from those who inherit the new problem” (Senge, 2006) To ensure the continues success of the system, it must be consistently under review, it must be consistently changing – it must be measured against the expectations and in the context of its design.

References

Ackoff, R. L. (1981). Creating the Corporate Future. New York: John Wiley & Sons.

Antony, J., Singh Bhuller, A., Kumar, M., Mendibil, K., & Montgomery, D. C. (2012). Application of Six Sigma DMAIC methodology in a transactional environment. International Journal of Quality & Reliability Management, 29(1), 31-53.

Courtnell, J. (2020, June 29). What is SIPOC? How to Create a SIPOC Diagram. Retrieved from Process.st: https://www.process.st/sipoc/

Davenport, T., Harris, J., & Morison, R. (2010). Analytics at Work: Smarter Decisions Better Results. Boston: Harvard Business School Publishing.

Gharajedaghi, J. (2011). Systems Thinking: Managing Chaos and Complexity. Burlington, MA: Elsevier.

Goldratt, E. (2004). The Goal. Barrington, MA: North River Press.

Haksever, C., & Render, B. (2018). Service and Operations Management. Hackensack, NJ: World Scientific.

Senge, P. (2006). The Fifth Discipline. New York: Crown Business.

Shankar, R. (2009). Process improvement using six sigma: a DMAIC guide. Milwaukee, WI : ASQ Quality Press.

SIPOC.Info. (n.d.). SIPOC Diagrams. Retrieved from SIPOC Diagrams: https://sipoc.info/sipoc-diagram/

Tutorials Point. (2020). Control Systems - Introduction. Retrieved from Tutorials Point: https://www.tutorialspoint.com/control_systems/control_systems_introduction.htm