Quality Control is the control of the operational techniques and the activities used to fulfill requirements of quality. One way to control quality of a product or service is to provide in-process and/or final inspection. The problem with inspection is that experts maintain that even with 100% inspection, inspection is only 80% effective – at best.
A process owner is fairly self-explanatory, but many times in lean six sigma, the process owner is not identified until the ‘ideal process state’ has been put into place.
The importance of the process owner reveals itself in the last two subtasks of the control phase of DMAIC. The process which is usually somewhat of a new process needs to be managed by one person who will ensure that the improvements are sustained. Without a process owner, it is only human nature that a process will tend to go back to the old way.
Project management is the discipline of Initiating, Planning, Executing, Monitoring and Controlling, and Closing projects to achieve specific goals. The Work Breakdown Structure is at the heart of .
The Work Breakdown Structure breaks down the project into smaller components. It defines and groups a project’s work elements in a way that helps organize and define the total work scope of the project.
A Pull System is a control-oriented system that operates by receiving signals that more production is needed. A pull system is contrasted with a typical push system that is common with mass production.
In a pull system, the requirement to produce more occurs as a ‘signal’ from one process to the previous process. This is counter to the way a lot of processes operate.
There are many definitions of quality. Some define quality as ‘conformance to requirements’. Others define quality as merely ‘meeting-the-customers’ specifications’. When the mentality is to focus on the most marginally-accepted product or service (focusing on either the upper or the lower specification, or both), it isn’t as competitive as focusing on the most ideal product target. This author defines quality as: Product uniformity around a target value(s). The target might be a larger-is-better quality characteristic or a smaller-is-better quality characteristic. It might also be a nominal-is-best where some named value is the target. With any of these quality characteristics, it is important that there is relentless attention paid to having as little variation around the target values as possible, or competitive.
Quality Function Deployment (QFD) is a process for planning products and services. It starts with the Voice of the Customer (VOC) which becomes the basis for setting requirements. From that, the organization identifies the what’s – the most important needs of the VOC. From there, a team will identify the how’s– those areas of process focus that address each of these identified requirements.
A radar chart is a graph in which each data series has it’s own axis and ‘radiates’ out from a central point.
A radar chart is particularly useful when there are multiple characteristics that need to be compromised to a point (a ‘sweet spot’) that is most pleasing to most of the customers in a particular niche. For example, there really is no such thing as a perfect car that will please everyone.
Random sampling means that every sample drawn from a population has an equal probability of being chosen.
Random sampling is important because when drawing from the population, one wants to have no bias in the sample drawn. For example, if samples are taken from only the top of the pile in a batch, those pieces that are on top were probably all made at a point in time that was much later than those parts that are on the bottom of the barrel. If something changed in the process that is time related, and if only the parts chosen as samples are taken from the top of the pile, the variation due to time would not be detected. In many cases, random number generators are used to eliminate bias when drawing samples.
To get an understanding of the ratio scale, there needs to be an understanding of the interval scale. The interval scale has the intervals between each measurement increment equally split. Think of the temperature on a thermometer. The difference between 71 degrees and 72 degrees is the same as the difference between 31 and 32 degrees. The increments are in even intervals.
Repeats in a design of experiments methodology is a trial that is conducted; then, each trial is immediately repeated before moving onto the next trial condition. This methodology is useful in learning about experimental error. Contrast this with replicating. With replicating, all of the trials are conducted; then only after all have been completed, they all are conducted again. This also is useful in learning about experimental error, but replication is better because it spreads out the element of time across the whole set of experimental trials.
