PDCA cycle and the MASP methodologyRemember when we talked about Deming? He is known for the PDCA cycle (Plan, Do, Check and Act), which he spread widely, especially in Japan from 1950 onwards. The original idea came from Walter Shewhart, Deming's mentor, which is why the cycle is also called the Shewhart cycle or the Deming cycle:PDCA CycleImagine yourself in a doctor's office: you know you are ill, but you do not know the name of the illness. When the doctor analyzes the symptoms, orders tests, finds the cause of the problem and decides on the treatment, we are in the Plan phase (P).When you follow the prescribed treatment, we are in the Do phase (D). When the doctor asks you to come back for another appointment and compares the results with what was expected, he is Checking (C). Finally, in light of the results, he Acts (A): if the treatment worked, he keeps it and ends it; if it did not, he corrects the course and the cycle starts again.The four phases of PDCA can be used to solve both problems in our lives and problems that are part of any company.MASP (from the Portuguese Método de Análise e Solução de Problemas, a problem analysis and solving method widely used in Brazil, based on the Japanese QC Story) breaks the four PDCA phases down into eight steps, which the figure below numbers from 1 to 9 because it counts the question "was the block effective?" as a step of its own. The sequence begins with identifying the problem; continues with observing it and with the analysis, in which testing and prioritization tools are used to reach the root cause; then comes the action plan to block that cause, the action itself, checking whether the block was effective and, finally, standardizing what worked and the conclusion.Problem-solving methodology and analysisAlthough MASP and PDCA provide a concrete roadmap for solving a problem, on their own they do not explain how to identify the problem, much less which are the best solutions to solve it. PDCA by itself is just a flow of tasks to make sure the problem is dealt with. MASP goes further: besides the flow, it indicates that we should use tools that help with decision-making.For example, imagine you are part of the marketing team for a children's movie that is not attracting the expected audience. How can you find out what the problem really is? The great thinkers of Quality and MASP are clear: there are some basic quality tools that all employees of a company should know.Ishikawa (1993) classified statistical control techniques into three groups of increasing complexity. The first group is made up of the seven basic tools, which should be known by everyone in the company and which, according to him, can solve up to 95% of quality problems. They are: Pareto chart, cause-and-effect diagram, histogram, control chart, check sheet, scatter diagram and stratification. Many authors, as in this article, put the flowchart in place of stratification.In MASP, the team carried out the action plan, but at the checking step defects remained at the same level. What is the next step?Answer: Go back to observation and analysis, because the root cause was not blocked — in the MASP figure, a 'no' answer to the test 'was the block effective?' sends the flow back to observation. Standardizing only makes sense after checking confirms that the action worked.HistogramA histogram is a bar chart that shows the frequency distribution of a variable: how many measurements fall into each range (class) of values. For example: how is the height of the students in a school distributed? Are there more students in the 1.56 m to 1.60 m range or in the 1.68 m to 1.72 m range? Or think of the age distribution of two different populations: a city with a young population and a city in Japan, a country with an ageing population. In the Japanese city, the histogram will probably show more people in the older age ranges.HistogramIn the histogram above (titled "Height of the students of School A.B.C."), each bar is a 4 cm range, centred on the value on the horizontal axis. Of the 40 students, 11 are in the central range (around 162 cm), which is the most frequent; frequencies decrease on both sides, and the highest range (174 cm) has only 3 students. In a single chart we can see where the data are concentrated, how much they vary and whether the distribution is symmetrical, information that would take much longer to extract from a list of 40 numbers.Scatter diagramThe scatter diagram shows whether two variables are related. The table below contains a sample with the age, weight and height of students at a college (in the image: Idade = age, Peso = weight, Altura = height). Let's analyze two diagrams built from this same sample:Table with dataFirst, let's check the correlation between weight and height. With the data in the table, we build the following chart (height in centimetres on the horizontal axis and weight in kilograms on the vertical axis):Correlation between weight and heightThe points cluster around an upward-sloping line: as height increases, weight also tends to increase. There is, therefore, a strong positive correlation between the two variables (the correlation coefficient calculated from the table data is about 0.90).In the age and weight diagram below, however, the points are much more scattered and do not line up around a straight line. There is a slight tendency for older students to weigh more (coefficient of about 0.55), but the correlation is weak: knowing a student's age helps little in predicting their weight.Correlation between weight and ageThe figure below summarizes the possible types of scatter diagrams, indicating whether there is a correlation between the variables and of what kind (from left to right and top to bottom: strong positive, weak positive, strong negative, weak negative and no correlation).Types of correlationAn important caution: correlation does not prove causation. Two variables may move together because a third one influences both; the scatter diagram points to relationships worth investigating, not to ready-made conclusions.A bakery builds a scatter diagram of oven temperature against the number of burnt loaves per batch, and the points form a narrow upward-sloping band. What is the correct conclusion?Answer: There is a strong positive correlation, and temperature is worth investigating as a possible cause — points aligned in a narrow rising band indicate a strong positive correlation. But correlation does not prove causation, nor does it rule out other causes, so the 'it has been proven' option goes too far.Pareto chartThis chart is linked to the so-called 80/20 rule. The Italian economist Vilfredo Pareto observed, at the end of the 19th century, that about 80% of the land in Italy belonged to about 20% of the population. In the 1940s, Joseph Juran brought the idea to quality and named it the Pareto principle: a vital few causes account for most of the problems, and the many remaining causes are trivial. The 80/20 ratio is a rough reference, not an exact law.Let's consider the following case: you work in the quality department of a hotel and are responsible for drawing up an action plan to improve guest satisfaction.The first step, of course, is to analyze the complaints from recent months. Looking at the list, you find more than 100 different types of complaint. Where do you start? How do you prioritize them? Which ones weigh more? The Pareto chart helps with this, because it orders the causes by frequency. It is a bar chart with occurrences in descending order, accompanied by a line showing the cumulative percentage.Let's use the hotel example. The table below shows the categories of problems (in the image: Frequência = frequency, Percentagem = percentage, % Acumulada = cumulative %); to keep it simple, we have limited the example to 10 categories: internet speed, menu variety, lack of daily room cleaning, parking cost, dirty bed linen, noise, number of TV channels, staff friendliness, slow check-in and mattress. The Frequency column shows how many times each problem was recorded (384 complaints in total), and the Percentage column is the frequency divided by that total; for example, 120 ÷ 384 = 31.25%.Problems presented at the hotelWith this table, we build the following Pareto chart. Don't worry about how to build it yet; let's just analyze it together:Pareto diagram regarding the hotel problemsThe problems are in descending order of frequency. Where would you start? With staff friendliness? With the number of TV channels? Probably not. The first three problems add up to 250 of the 384 complaints, that is, about 65% of the total (the cumulative percentage in the third row of the table). Solving internet speed, menu variety and the lack of daily room cleaning would be the starting point: 3 of the 10 categories (30%) account for almost two thirds of the complaints. Note that, in this example, the ratio is not exactly 80/20: 80% of the complaints is reached with 6 of the 10 categories.After the main problems are solved, the chart should be redone with new data. Problems that used to be secondary move to the top of the list, and the prioritization cycle starts again; this is how Pareto supports continuous improvement.In the hotel example, how many complaints would disappear if the first two problems (internet speed and menu variety) were fully solved, and what percentage of the total does this represent?Answer: 200 complaints, about 52% — 120 + 80 = 200, and 200 ÷ 384 = 52.08%, which is the cumulative percentage in the second row of the table. The 65% corresponds to the first three problems, including daily room cleaning.FlowchartA flowchart is a representation of a process that uses graphic symbols to describe, step by step, the nature and flow of that process. The aim is to show simply the path of information and materials, highlighting the sequence of operations in the work being performed.The steps are represented by geometric shapes, such as rectangles, diamonds, circles and arrows, and each symbol has a meaning. Flowcharts should be used when we want to:understand how a process works;study a process in order to implement improvements;explain to other people how a process works;document a process.Below are the basic shapes of a flowchart and their meanings (in the image: start or end, decision, process, predefined process, manual operation, connector, information, preparation, document and manual input). The rectangle represents an activity (process), the diamond a decision with outputs such as yes and no, and the oval shape the start or end of the flow:Basic flowchart figuresIn the example below, a loan application, the diamonds show the two decision points (was the loan approved? was the exception approved?) and each answer leads along a different path to the end of the process:Flowchart exampleCause-and-effect diagram (fishbone or Ishikawa diagram)Proposed by the Japanese engineer Kaoru Ishikawa in 1943 and refined in the following years, this diagram is also called the fishbone or cause-and-effect diagram. Its purpose is to help the team organize the possible causes of a problem and guide the search for a solution.The diagram is usually built as a group, with the participation of the people involved in the process being analyzed. First the problem (the effect) to be studied is defined; then the possible causes are listed, for example through brainstorming, and only then is the diagram built.The diagram classifies the causes into six categories, the famous 6Ms: method, material, manpower, machine, measurement and mother nature (environment).MethodThis is where causes related to the way of working go. The problem does not lie in how an activity is performed, but in the method used. In other words, the execution may be correct and the method wrong.MaterialThis groups the causes related to raw materials. For example, we need to check whether a nonconformity in the final product may come from using raw material that is out of specification, or even from a poorly written specification.ManpowerThis covers causes related to people's actions: lack of technical preparation, failure to follow procedures and even carelessness.MachineThis groups causes related to machine failures, which may occur, for example, due to lack of maintenance. Many of these failures, however, come from incorrect operation of the equipment; in that case, besides a cause in the machine category, a cause should also be recorded in the manpower category.MeasurementHere we are basically talking about measuring instruments and their calibration. For measurements to be reliable, the following must be considered: use of the appropriate instrument, the measurement method, calibration and instrument wear (when it should be replaced).EnvironmentThese are causes coming from the production environment, such as temperature, humidity, noise, vibration, lighting and indoor pollution, which can influence products or services.Example of a diagramBelow is a completed example for the problem "defective products" (the categories in the image are método, máquina, material, mão de obra, medida and meio ambiente, the Portuguese 6Ms):Ishikawa DiagramWhile analyzing parts with wrong dimensions, the team finds that the caliper used in inspection is past its calibration date. Under which of the 6Ms should this cause be recorded?Answer: Measurement — measuring instruments and their calibration belong to this category. The temptation is to put it under machine, but that category deals with the equipment that makes the part, not the equipment that measures it.Control chartHere we explain what a control chart is and what it is for. In the next lesson, on statistical process control, we will go deeper into the subject and see how to build one.A control chart is a graph used to monitor a process over time. It has a centre line, which is the process average, and two statistically calculated limits: the upper control limit (UCL) and the lower control limit (LCL). In the chart below, the limits appear in red and the centre line in green. The aim is to check whether the process is in statistical control, that is, free of special causes of variation.Control chartTo understand the chart, imagine a company that manufactures a part whose approval depends on its diameter, which must be between 73.98 mm and 74.02 mm. Outside these values, the part is rejected. These are the specification limits, set by the customer or the design.The chart's control limits are something else: they are calculated from the process's own data and show how far its natural variation goes. In the chart, the UCL is 74.0143 mm and the LCL is 73.9881 mm, narrower than the specification. Samples 14 and 18 fell outside the control limits, although they are still within specification: the parts were approved, but the process gave a signal of a special cause, which should be investigated before rejected parts appear. Samples 3 to 9 are also marked, because they form a run of seven consecutive points above the centre line, another pattern that indicates a special cause.By separating common causes from special causes of variation and showing whether a problem is local (calling for action at that point in the process) or requires management attention (calling for a change in the system), control charts avoid frustration and the cost of tackling the problem in the wrong place. See the difference between the two types of cause:Common causesSpecial causesDefinitionThe accumulated effect of many small causes, always present in the process, each with little individual influence.Occasional failures that arise during the process, each with great individual influence.ExamplesSmall vibrations, normal variations in temperature and humidity, limitations of the work method itself, among others.A batch of substandard raw material, operator errors, incorrect machine setting, tool wear, among others.By helping to improve the process, control charts provide:an increase in the percentage of products that meet customer requirements;less rework and scrap and, consequently, lower manufacturing costs;a higher probability of producing acceptable items;information for process improvement.In the example control chart, sample 18 was above the UCL (74.0143 mm) but below the 74.02 mm specification limit. What does this mean?Answer: The parts are accepted, but the process showed a signal of a special cause that should be investigated — the specification decides whether the part is approved; the control limit tells whether the process has changed. Ignoring the signal just because the part passed means missing the chance to act before rejections appear.Check sheetA check sheet is a form (chart, table or spreadsheet) for organizing data collection at the moment things happen, which makes both collection and analysis easier.Verification SheetIn the example (a final inspection of moulded plastic parts), the inspector marks each defect found with a stroke; each crossed-out square counts as 5. The defects are surface marks (17), cracks (11), incomplete part (26), deformation (3) and others (5). Of the 1,525 parts inspected, 42 were rejected, but 62 defects were recorded, because one part can have more than one. The sheet already delivers the data ready for a Pareto chart: incomplete part (26) is the most frequent defect.Conclusion on MASPMASP is the sum of the PDCA cycle and the quality tools. In other words, throughout the MASP steps the tools are used to support the search for a solution.PDCA + toolsThe figure above illustrates this sum (PDCA + tools), in a condensed seven-step version. In identification, for example, we see the check sheet, the Pareto chart and the control chart; in observation, the flowchart and the schedule; in analysis, the brainstorming meeting, the cause-and-effect diagram and hypothesis testing; in the action plan and in standardization, 5W1H (what, why, where, who, when and how).A team needs to decide which type of defect to tackle first and then find the possible causes of that defect. Which pair of tools meets these two needs, in this order?Answer: Pareto chart and cause-and-effect diagram — the Pareto chart prioritizes defects by frequency and the Ishikawa diagram organizes the possible causes of the chosen defect. The check sheet collects data and 5W1H organizes the action plan, but neither of them prioritizes or identifies causes.Exercises(1) You would like to improve the quality of your school or of the course you are taking (for example, at CEFET-RJ, a Brazilian federal technical education institution). How would you start? Which tools would you use?
ردود الفعل
0
0 تعليقات
كن أول من يعلق
لا شيء هنا بعد.