In 1950, the Japanese Union of Scientists and Engineers (JUSE) invited legendary quality guru W. Edwards Deming to go to Japan and train hundreds of Japanese engineers, managers and scholars in statistical process control. Deming also delivered a series of lectures to Japanese business managers on the subject, and during his lectures, he would emphasise the importance of what he called the “basic tool” that were available to use in quality control.
One of the members of the JUSE was Kaoru Ishikawa, at the time an associate professor at the University of Tokyo. Ishikawa had a desire to ‘democratise quality’: that is to say, he wanted to make quality control comprehensible to all workers, and inspired by Deming’s lectures, he formalised the Seven Basic Tools of Quality Control.
Ishikawa believed that 90% of a company’s problems could be improved using these seven tools, and that –- with the exception of Control Charts — they could easily be taught to any member of the organisation. This ease-of-use combined with their graphical nature makes statistical analysis easier for all.
The seven tools are:
- Cause and Effect Diagrams
- Pareto Charts
- Flow Charts
- Check sheet
- Scatter Plots
- Control (Run) Charts
What follows is a brief overview of each tool. If you would like to know more or be trained in their use, please get in touch using the form at the top of the page.
Click on the images to enlarge; you will need to disable your pop-up blocker.
Cause and Effect Diagrams
Also known as Ishikawa and Fishbone Diagrams
First used by Ishikawa in the 194os, they are employed to identify the underlying symptoms of a problem or “effect” as a means of finding the root cause. The structured nature of the method forces the user to consider all the likely causes of a problem, not just the obvious ones, by combining brainstorming techniques with graphical analysis. It is also useful in unraveling the convoluted relationships that may, in combination, drive the problem.
The basic Cause and Effect Diagram places the effect at one end. The causes feeding into it are then identified, via brainstorming, by working backwards along the “spines” (sometimes referred to as “vertebrae”), as in the diagram below:
For more complex process problems, the spines can be allocated a category and then the causes/inputs of each identified. There are several standard sets of categorisations that can be used, but the most common is Material, Machine/Plant, Measurement/Policies, Methods/Procedures, Men/People and Environment –- easily remembered as the “5M’s and an E” –- as shown in the below:
Each spine can then be further sub-divided, as necessary, until all the inputs are identified. The diagram is then used to highlight the causes that are most likely a contributory factor to the problem/effect, and these can be investigated for inefficiencies/optimization.
Control (Run) Charts
Dating back to the work of Shewhart and Deming, there are several types of Control Chart. They are reasonably complex statistical tools that measure how a process changes over time. By plotting this data against pre-defined upper and lower control limits, it can be determined whether the process is consistent and under control, or if it is unpredictable and therefore out of control.
The type of chart to use depends upon the type of data to be measured; i.e. whether it is attributable or variable data. The most frequently used Control Chart is a Run Chart, which is suitable for both types of data. They are useful in identifying trends in data over long periods of time, thus identifying variation.
Data is collected and plotted over time with the upper and lower limits set (from past performance or statistical analysis), and the average identified, as in the diagram below.
Based upon the Pareto Principle that states that 80% of a problem is attributable to 20% of its causes, or inputs, a Pareto Chart organises and displays information in order to show the relative importance of various problems or causes of problems. It is a vertical bar chart with items organised in order from the highest to the lowest, relative to a measurable effect: i.e. frequency, cost, time.
A Pareto Chart makes it easier to identify where the greatest possible improvement gains can be achieved. By showing the highest incidences or frequencies first and relating them to the overall percentage for the samples, it highlights what is known as the “vital few”. Factors are then prioritized, and effort focused upon them.
A Scatter Diagram, or Chart, is used to identify whether there is a relationship between two variables. It does not prove that one variable directly affects the other, but is highly effective in confirming that a relationship exists between the two.
It is a graphical more than statistical tool. Points are plotted on a graph with the two variables as the axes. If the points form a narrow “cloud”, then there is a direct correlation. If there is no discernible pattern or a wide spread, then there is no or little correlation.
If both variables increase as the other increases – i.e. the cloud extends at roughly 45 degrees from the point where the x and y axes cross – then they are said to be positively correlated. If the one variable decreases as the other increases, then they are said to be negatively correlated. These are linear correlations; they may also be non-linearly correlated.
Below is an example of a Scatter Diagram where the two variables have a positive linear correlation.
Like Pareto Charts, Histograms are a form of bar chart. They are used to measure the frequency distribution of data that is commonly grouped together in ranges or “bins”. Most commonly they are used to discern frequency of occurrence in long lists of data. For instance, in the list 2, 2, 3, 3, 3, 3, 4, 4, 5, 6, the number 3 occurs the most frequently. However, if that list comprises several hundred data points, or more, it would be difficult to ascertain the frequency. Histograms provide an effective visual means of doing so.
“Bins” are used when the data is spread over a wide range. For example, in the list 3, 5, 9, 12, 14, 17, 20, 24, 29, 31, 45, 49, instead of looking for the occurrence of each number from 1 to 49, which would be meaningless, it is more useful to group them such that the frequency of occurrence of the ranges 1-10, 11-20, 21-30, 31-40 and 41-50 are measured. These are called bins.
Histograms are very useful in discerning the distribution of data and therefore patterns of variation. They monitor the performance of a system and present it in a graphical way which is far easier to understand and read than a table of data. Once a problem has been identified, they can then also be used to check that the solution has worked.
A flow chart is a visual representation of a process. It is not statistical, but is used to piece together the actual process as it is carried out, which quite often varies from how the process owner imagines it is. Seeing it visually makes identifying both inefficiencies and potential improvements easier.
A series of shapes are used to depict every step of the process; mental decisions are captured as well as physical actions and activities. Arrows depict the movement through the process. Flow charts vary in complexity, but when used properly can prove useful for identifying non-value-adding or redundant steps, the key parts of a process, as well as the interfaces between other processes.
Problems with flow charts occur when the desired process is depicted instead of the actual one. For this reason, it is better to brainstorm the process with a group to make sure everything is captured.
Also known as Data Collection sheets and Tally charts
Like flow charts, check sheets are non-statistical and relatively simple. They are used to capture data in a manual, reliable, formalised way so that decisions can be made based on facts. As the data is collected, it becomes a graphical representation of itself. Areas for improvement can then be identified, either directly from the check sheet, or by feeding the data into one of the other seven basic tools.
Simply, a table is designed to capture the incidences of the variable(s) to be measured. Tick marks are then manually put in the relevant boxes. As the ticks build up, they give a graphical representation of the frequency of incidences. Below is a typical example.
The seven basic tools of quality can be used singularly or in tandem to investigate a process and identify areas for improvement, although they do not all necessarily need to be used. If a process is simple enough – or the solution obvious enough – any one may be all that is needed for improvement. They provide a means for doing so based on facts, not just personal knowledge, which of course can be tainted or inaccurate. Ishikawa advocated teaching these seven basic tools to every member of a company as a means to making quality endemic throughout the organisation.
Ishikawa, Kaoru. Guide to Quality Control. Kraus International Publications, White Plains, New York, 1982.
Tague, Nancy R. The Quality Toolbox, Second Edition, ASQ Quality Press, 2004.