Process Improvement in Six Sigma

We all have annoying relatives who ask us about our secondary school percentage, don’t we? Ever got bothered by their, “Oh how much percentage did you score in your boards?” or “Oh you scored only 99.6%? Where did the remaining 0.4% go?” and thought, well what difference will that 0.4% make anyway? What if I lost 0.4% or 0.1 % or 0.01 or even 0.001? Am I still not the topper of my school! You are and might laugh these little variations in numbers off, but a multinational giant with tasks on the scale of millions will not. 

Let’s look at the sales figure of Amazon on Cyber Monday in 2013 when the company processed 36.8 million orders1. Let’s assume that it cost Amazon $40 for each order error. Below is a chart that explains the cost difference between getting 99.97% orders right and getting 99.9996% orders right.

Order error	Total Orders	Errors	Average cost per error	Total cost of error
Cost to Amazon at 99.97% accuracy	36.8 million	8,674	$40	$346,960
Cost to Amazon at 99.9996% accuracy	36.8 million	125	$40	$5,000

That’s right! Amazon could’ve lost a whopping $346,960 operating at 99.97% accuracy! Why exactly did we choose these figures of 99.97 and 99.9996? In mega-industries and among statistics connoisseurs these figures are termed in Sigma (denoted by σ), respectively five sigma & Six Sigma, and this case study of Amazon is an example of how improving from five sigma to Six Sigma in order error can reduce losses and optimize customer satisfaction. Companies like Amazon, Toyota, Ford, GE, Walmart, etc. are trying to optimize their operations by reducing such errors (variations) in their existing processes and are striving to meet the standards of the so-called ‘Six Sigma’.

But what is the six sigma methodology anyway? And how can it be used to improve any process? In simple terms, Six Sigma is a statistical representation of a “perfect” process- a tool to identify variations in a process, validate assumptions and brainstorm appropriate solutions. Using this tool, a firm can visualize and predict the outcomes of a particular process with a high level of accuracy and take safe financial decisions. Not to mention that a process is any activity or task performed usually in a well-defined sequence on some tangible inputs such as raw material or intangible inputs such as information or data to add some value to it and transform them into some tangible outputs such as finished goods or services required by the customer. 

Any such process will always have some scope of variation in it and could be because of either common causes, which can’t be statistically controlled such as the mental state of the operator/physical/geographical conditions, or special causes such as tool wear, calibration error, etc., which can be studied, analyzed and statistically controlled.

Using the Six Sigma method for a particular project, a project leader creates a dedicated team that focuses on studying various aspects of variations in an existing process and generates a road map known as DMAIC. DMAIC consists of five phases namely-Define, Measure, Analyse, Improve, and Control. This structural approach to process improvement includes identifying the critical causes that are bringing variations in the process, verifying those causes, brainstorming the appropriate solutions or course of action to reduce or eliminate them, implementing these solutions, and creating a robust control plan to ensure that the improvement continues.

Process Improvement in Six Sigma

Define: In this stage, the Six Sigma team defines the scope and severity of the problem, its financial and quality impact, its deliverables on completion, its milestones, and its timeline. It is also in this stage that the team estimates the required capital investment in implementing this project and if carrying out the Six Sigma activity will improve the process in terms of profits. Measure: Once the team is well aware of the project, its deliverables, and its economic viability, it moves from the Define stage to the Measure stage. Usually, this transition has to go through a ‘Gate review’ in which the team lead evaluates if the team members are clear about the project’s milestones and if they are ready to move into the measure phase. During the measure phase, the team refines the problem statement, indicates how a process should be measured, and measures the current performance of a process so that it can be compared with the performance after the improvements have been implemented. A major challenge in this stage is to correctly decide on ‘what to measure’. Sometimes the team members end up measuring parameters that are not relevant to the process and do not provide any answers. A successful measure phase requires good observation skills, the right knowledge of statistics, and an understanding of data types. Analyze: After the measure phase is done, the project moves into the analyzing phase in which the measured data is studied and processed. In most projects, there is no well-defined line between the Measure phase and Analyse phase as it is in Define and Measure. More often than not, the team measures, analyses, and measures some more. The team considers all such relevant data and attempts to identify the root causes and plot them into a statistical model such as Pareto Charts, Cause and effect Diagrams, Run charts, Process maps, and Value analysis. Improve: During the improvement phase of the six sigma activity, the team selects a solution and starts its implementation on the assembly line or in the service portfolio. Since these solutions are expensive and can cause downtime or losses, care must be taken while implementing them or a backup plan must be put in place. When not one final solution has been derived, rather a few small solutions, it can be hard to determine which solution will improve the process. However, it is usually the best practice to implement one change at a time, measure its improvement as compared to the original state of the process, and choose the best one from them. Control: This is the fifth and final phase of the six sigma process improvement. In this phase, the team documents its observations, establishes a methodology to evaluate the process, and develops a statistical model to continuously control process variations in the new improved process so that the improvement is maintained. Tools used by the team in this phase are process dashboards, control charts, Cp, Cpk, etc. The six sigma philosophy is used widely across all sectors and industries and is equally effective in them. Here are some statistics that show the efficiency, 55 Process Improvement Case Studies & Project Results [2022] (aimultiple.com)
  One major challenge while implementing it is to measure the right parameters in industries where the output values are intangible, such as hospitality business or entertainment sectors. However, with the correct brainstorming and apt usage of statistical tools, one can use this method and the concept of DMAIC to bring about remarkable improvements in industrial processes.