A SIMPLE WAY TO DIGEST SPC. Belmont, NC Cypress, TX 77429

Transcription

1 A SIMPLE WAY TO DIGEST SPC Christy Morrow Loa Olafsson McNeese Industrial Distribution Group Business Process Improvement Belmont, NC Cypress, TX KEY WORDS SUMMARY Statistical Process Control Training The purpose of this paper is to demonstrate a simple way of teaching an introduction to statistical process control (SPC) to associates using peanut M&M's. This paper contains the teaching guide. This teaching guide includes an overview of how to introduce each SPC tool, the general steps in constructing the SPC tool, and how to use M&M's to practice the SPC tool. This training is given as part of a two-day introduction to quality to all associates in an industrial distributor. INTRODUCTION Industrial Distribution Group Charlotte (IDG) is an industrial distributor based in Belmont, North Carolina. IDG has been involved with quality since Each associate attends a two-day introduction to quality within the first six months of his or her employment. This training introduces the associates to IDG s quality process. This process has many components. On the soft side of quality, there is teamwork, motivation, recognition, reinforcement, social styles and feedback. On the hard side of quality, there is ISO 9000, problem solving methodologies, and, of course, the basic SPC tools. The initial SPC training developed for inclusion in this introductory training involved showing examples of each SPC tool. These tools included Pareto diagrams, process flow diagrams (PFD), histograms, cause and effect (fishbone) diagrams, and control charts. Participants, in teams, would construct a process flow diagram and a fishbone on some self-chosen topic. Feedback on this approach indicated that most participants found that this approach was very boring. In addition, what they learned did not stay with them for very long on the job. A new approach was definitely needed. This approach had to be fun and had to be something that associates could easily remember - something like the red bead exercise. Everyone remembers this over time, especially if they were one of the willing workers. It was decided to build this approach around peanut M&M's. Associates leaving this training should know how to construct a process flow diagram, cause and effect (fishbone) diagram and a Pareto diagram. They should also be able to interpret a histogram and a control chart. More advanced SPC training is available from other sources. The M&M's approach to teach the initial SPC training is described below. OUTLINE OF TRAINING Associates are taught the concept of variation before the SPC tools are introduced. The variation module includes the following points: The definition of quality. Why we should focus on our processes. What a process is. The detection and prevention modes of operation.

2 The definition of variation. Why we must trace variation back to its source to eliminate it. How we have handled variation in the past (specifications and standards). What common and special causes of variation are. The Shewhart (control chart) approach to handling variation. The lessons of the red beads. The red bead exercise is performed during this module on variation. This background information is needed before moving into the SPC tools. The following is covered for each SPC tool: 1. What the SPC tool is and looks like. 2. When to use it. 3. What you can learn with this tool. 4. How to use it. Actual company examples of each SPC tool are included. The introduction to the SPC tool should only take 10 to 15 minutes. The rest of the time is spent applying the tools using peanut M&M's. The training process is described below as if you were teaching the workshop. TRAINING PROCESS Before starting, you need to have peanut M&M's available. Use the packages that contain about 22 M&M's per bag. You will need a bag for each person, one for each team and a couple for yourself (in case you get hungry). The first SPC tool introduced is the process flow diagram. The following points are made about the process flow diagram. What is a PFD? When do you use a PFD? What can you learn from using a PFD? How do you use a PFD? Where have we used a PFD? A PFD is a detailed picture of a process. It represents the logical flow of activities from the beginning of the process to the end of the process. You use a PFD when you want to improve the process. This is the usually the first step in process improvement. PFDs are also useful for training. You gain agreement from everyone on what the steps in the process are. You also can identify potential areas for improvement and measurement. You make a PFD using post-it notes following the steps in Figure 1. Entering an order; picking a line item in the warehouse, changing a work procedure, handling a customer complaint You are now ready to introduce the story of peanut M&M's. Tell the class that we are going to explore the use of the SPC tools by learning about peanut M&M's. There is a lot of information on the web site for M&M's that you can use to introduce the subject. 1 For example: M&M's Plain Chocolate Candies were first manufactured in 1940 when Forrest E. Mars, Sr. formed a company in Newark, New Jersey to make chocolate candies. The original idea was to offer chocolates that could be sold throughout the year, including summer, when many people stopped buying chocolate

3 because of the heat. Many stores either discontinued stocking chocolate or greatly reduced buying because this was before air conditioning. M&M s Plain Chocolate Candies are made of milk chocolate inside a thin and colorful sugar shell. The chocolate used in M&M s melts just like fine chocolate. This gives the product its great taste -- but the thin and colorful sugar shell surrounds the chocolate and gives it the crunchy texture. In other words, the milk chocolate melts in your mouth, not in your hand. M&M s Peanut Chocolate Candies are made of whole peanuts covered with milk chocolate inside a thin and colorful sugar shell. The Peanut variety of M&M s Chocolate Candies debuted in Tell the class we are going to learn about peanut M&M's through the use of our SPC tools. Ask the class the following questions: Have you ever had trouble opening a bag of peanut M&M's? Have you ever tried to open one end, could not and turned the bag around to try the other end? Have you ever ripped the bag open only to see the peanut M&M's go everywhere, ending up on the floor? What is your horror story about opening a bag of M&M's? It is evident that the people at Mars, Inc., the makers of peanut M&M's, need our help in developing the process for opening the bag. What SPC tool can you use to describe the process of opening a bag of peanut M&M's? The answer to the last question, of course, is the PFD. Then: Divide the class into teams of 4 5 people to make a process flow diagram on how to open a bag of peanut M&M's using the PFD in Figure 1. Use the same teams throughout the SPC exercises. Give each team one bag to look at, but tell them not to open the bag. As a class, determine the starting and ending steps for the PFD. Each team constructs a process flow diagram using flip chart paper on the wall with post-it notes. Remind the teams that it is not a very good PFD without a decision box in it. When all teams are finished, have the teams move to the PFD to their left. Using that PFD, see if the team can open the bag of M&M's. Then have the participants return to their seats and debrief the exercise. Figure 1: Steps in Constructing a PFD Determine the Decide on Tape flip chart starting and a process paper on the wall ending points of to study the process Put the ending point on As a team, sketch out Put the starting point on a Post-it note and place the process step-by- a Post-it note and place it in the lower rightstep, writing each step in the upper left-hand hand corner of the flip on a Post-it note. corner of the flip chart chart Does everyone Draw in the lines to Yes agree on indicate the the steps? direction of activity flow No Share the PFD Continue to discuss with all those until consensus is closest to the reached process for more input When debriefing the PFD, ask the class the following questions: How many teams could open the bag of M&M's based on the PFD? Did the PFD you used agree with the one your team developed? Why are there differences in the PFDs?

4 What insights into PFDs did this exercise give you? An example of a PFD from one of the classes is given in Figure 2. Figure 2: Process Flow Diagram for Opening a Bag of Peanut M&M's Hold bag of M&M's firmly in left hand Is bag vertical? No Turn bag to vertical position Yes Firmly grasp left top corner of bag between index finger and thumb Move right hand above left top corner of bag Slowly tear corner of bag until you reach the middle of the bag Fold torn part of the bag to the right The next SPC tool introduced is the Pareto diagram. The following points are highlighted about the Pareto diagram. What is a Pareto diagram? When do you use a Pareto diagram? What can you learn from using a Pareto diagram? How do you use a Pareto diagram? Where have we used a Pareto diagram? The Pareto diagram is a special type of bar chart used to determine which problem to work on first to improve a process. It is based on the 80/20 rule (80% of our problems are due to only 20% of the possible causes). You use a Pareto diagram to determine what problem to work on first. You can learn what the vital few problems or causes of a problem are. This allows you to focus your time and attention where you will get the most return. You make a Pareto diagram using the steps in constructing a Pareto diagram shown in Figure 3. Reasons for credits, reasons for customer complaints, supplier errors by supplier, reasons for computer downtime. Now return to the M&M's. Ask the class the following questions: What are the colors in M&M's peanuts? What color occurs most frequently? How many peanut M&M's are there in one bag? How much variation is there in the number of peanuts in one bag? What do we need to do to answer these questions? The answer to the last question is data. Then:

5 Tell the class we are going to find out the answers to these questions by Figure 3: Steps in Constructing a Pareto Diagram collecting data. Hand out one bag of M&Ms to each Determine the Total the frequency of participant. Tell them not to eat the Select the time problems or causes to occurrence (or cost) for each period to be covered be used in the problem during the time M&M's until you have said it is on the diagram. diagram. period. OK to do so. Tell the class to open the bags and Draw in the bars. record the number of different color Under the x-axis, write in The height of the bar Draw the x-axis and y- the most important will correspond to M&Ms they have. axis putting the proper problem (largest the frequency of units on the y-axis. frequency) first, then the occurrence for each Then have the class return to their next most important, etc. problem on the x- axis. teams and construct one Pareto diagram for the combined results of Calculate the Plot the Title the graph and their team. This is done by totaling cumulative cumulative include any other the number of each color for a team. percentages. percent line. important information. The teams should use the steps in constructing a Pareto diagram shown in Figure 3. You should provide each team a blank Pareto diagram with the y-axis filled in. When each team is done, have the participants return to their seats and debrief the exercise. Tell them they can now eat their M&Ms. When debriefing the Pareto diagram, ask the class the following questions: What did you learn from your Pareto diagram? Which color appeared most frequently? Least frequently? Why don t all the Pareto diagrams look the same? An example of a Pareto diagram done by one team is shown Figure 4. Figure 4: Pareto Diagram for Colors in Peanut M&Ms Bags (September 2000) Frequency % 92% 82% 72% 50% 36 29% Brown Red Yellow Blue Green Orange 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Percent As the teams are making the Pareto diagram, take a data collection sheet and go around to each team and record the number of red, green, yellow, brown, blue and orange M&M's in each bag. This is the data you need to complete the rest of the SPC training. There should be one line completed on the data collection sheet for each person in the class. An example of the data collection form is shown in Figure 5. This type of sheet is easily setup in a spreadsheet program such as Microsoft Excel. Make copies of the completed data collection sheet for everyone in the class.

6 Figure 5: Data Collection Sheet Design for M&M's Bag Number of Number of Number of Number of Number of Number of Number Red Green Yellow Brown Blue Orange Total Number The next SPC tool introduced is the histogram. The following points are made about the histogram. What is a histogram? When do you use a histogram? What can you learn from using a histogram? How do you use a histogram? Where have we used a histogram? Histograms present a picture of how the process "stacks up" over time. Histograms illustrate how many times a certain data value or range of data values occurred in a given time frame. You use a histogram to examine the variation in a process. You learn four things from a histogram: how much variation there is in the process, what the mode is, what the shape of the distribution is and the relationship of the specifications to the data. You make a histogram using the steps in constructing a histogram shown in Figure 6. Freight charges, time to close a customer complaint, lines picked per hour, number of credits per week. Now return to the M&M's. Ask the class the following questions: How many M&M's did you have in your bag? Did everybody on your team have the same number of M&M's in his or her bag? Why do you think there is not the same number each time? What would happen to Mars, Inc. if there was too much variation in the number of M&M's in a bag (e.g., from 2 to 40)? There appears to be valid reasons for Mars, Inc. to control this variation. What SPC tool can you use to determine the amount of variation present in the number of M&M's in a bag as well as determining which value occurs most frequently and the shape of the variation? The answer to the last question is, of course, the histogram. Then:

7 Pass out a copy of the data collection sheet. Divide the class back up into their teams to do a histogram on the total number of M&M's in a bag (the total is on the data collection sheet). Have a blank histogram chart ready to hand out with the x-axis and y-axis already filled in. Ask the class simply to count how many times each number occurred and plot that as a bar on the chart. When the histograms are complete, bring the class back together and debrief. When debriefing the histogram, ask the class the following questions: Figure 6: Steps in Constructing a Histogram Select the number of classes (K) to be used on the histogram using the Select the time frame (history) you are interested in and gather following guidelines: No. of data points K the data (at least 50 points over Calculate the class width Calculate the overall (width of bar on the range (R), where R equals histogram). The class the maximum value minus width, h, is determined the minimum value. by h = R/K. Select the class Round h to the nearest convenient number. boundaries so that data points do not fall on the boundary between two classes. What is the maximum number of M&M's in a bag? The minimum? What is the mode (the most common value)? What does the shape of the distribution look like? What are the advantages to Mars, Inc. in keeping the variation in the histogram at a minimum? The combined histogram from two classes is shown in Figure 7. Record the number of data points in each class. Label the histogram and include any other important information such as the time period covered. Plot the histogram. Figure 7: Histogram for Number of Peanut M&M's per Bag (September/October 2000) Frequency Number per Bag The next SPC is the cause and effect (fishbone) diagram. The following points are made about the cause and effect diagram. What is a fishbone diagram? A fishbone diagram is a tool that shows the relationship between a quality characteristic (effect) and possible sources of variation for this quality characteristic (causes).

8 When do you use a fishbone diagram? What can you learn from using a fishbone diagram? How do you use a fishbone diagram? Where have we used a fishbone? You use a fishbone to determine the causes of a problem. The fishbone diagram helps organize the results of a brainstorming session on causes of the problem. It helps you identify the possible causes, discuss the causes and determine the most likely cause. You make a fishbone diagram using the steps in constructing a fishbone diagram shown in Figure 8. To find causes for late deliveries to customers, for uncollected freight, for picking errors, and for credits. Now return to the M&M's. The class is going to develop a fishbone diagram on possible causes of variation in the number of M&M's in a bag. To do this, it is helpful to give them a little background information on how M&M's are produced. Again, this information is available from their web site. For example: M&M's" begin their lives as liquid chocolate -- made with all-natural ingredients like whole milk, cocoa butter, sugar and chocolate liquor (which comes from grinding cocoa beans). The chocolate is poured into tiny molds to create the M&M s chocolate centers. After they harden, they're moved on a conveyor belt to have their candy shells applied. The chocolates are rotated in large containers as the liquid candy is sprayed onto them. They get several coatings to ensure a complete and even coating on every piece. As the liquid dries, it turns into M&M s famous crunchy shell. The single-colored batches of candy are combined into mixtures of red, yellow, blue, green, brown and orange. They're sifted to eliminate misshapen pieces, and then transported to the machine that will put the "m's" on their shells. Next comes the big moment -- when the "m's" on the candies are put on. A special conveyor belt moves the pieces along, each in its own little indentation, past rubber etch rollers that gently touch each candy, printing that distinctive "m" on its shell. Figure 8: Steps in Constructing a Fishbone Diagram Pinpoint the problem you want to solve or the goal you want to reach. Select the categories for the causes of variation in the problem or goal. Brainstorm detailed causes for each main factor. These detailed causes are written on branches off the categories Materials Machines Measurement Write the problem or goal (effect) on the right hand side and draw a horizontal line to the left. Draw in the categories. Always try to branch further by continuously asking, "What causes variation in this branch?" Problem or Goal Methods People Environment A special packaging machine then weighs the M&M s, pours the proper amount into each bag, and heat-seals each package to ensure product safety and freshness. The finished packages are moved along a conveyor belt to a machine that assembles the shipping cartons, fills them with the right number of packages, and seals the cartons closed. The cartons are then loaded into trucks and transported to distribution centers, and then to stores across the country! Now ask the class the following questions: Did you think that there are ever any problems with the making of peanut M&M's? What SPC tool can be used to determine the causes of a problem The answer to the last question is the fishbone diagram. Then:

9 Divide the class back up into their teams to do a fishbone on the reasons for varying numbers of M&Ms in a bag. Debrief by going around the room for each team to highlight a few of their answers. An example of a fishbone diagram from one class is given in Figure 9. Figure 9: Fishbone Diagram on the Causes of Variation in the Number of Peanut M&M's in a Bag People Environment Machine Someone eats them while packaging No inspector Too hot Operator distractions Counter not calibrated Clogged hole Operator error No data collection Scales aren t calibrated Heavier ingredients Different number of M&M's per bag No documented procedure Bags measured by weight instead of pieces Different sizes of peanuts Methods Measurement Material The next SPC tool is the control chart. The following points are made about control charts. What is a control chart? When do you use a control chart? What can you learn from using a control chart? How do you use a control chart? Where have we used a control chart? A control chart is a picture of the variation in your process over time. You use a control chart to monitor the key variables in any process. You can learn if the process is in statistical control whether there are just common causes of variation present or if you have special causes present also. There are five basic steps in developing a control chart: 1. Gather data. 2. Plot the data. 3. Calculate the average. 4. Calculate the control limits. 5. Interpret the chart. Picking accuracy, service level, customer complaints, and average time for delivery. The details of how to construct a control chart are not covered in this introduction. The focus is on interpreting the control charts. The following tests are taught to the class: Points beyond the control limits. Seven points in a row above or the below the average.

10 Seven points in a row trending upward or downward. Time is spent examining control charts to determine if the chart is in or out of statistical control. The p, c and individuals control chart are covered. These three charts are all most distributors need. Now, return to the M&M's. For the entire class, determine the percentage of different colors in an average peanuts M&M's bag (add up all of one color and divide by the total number of M&M's )? The percentages from the September 2000 class are shown below. Ask the class: % Red 17% % Yellow 25% % Brown 20% % Orange 6% % Blue 19% % Green 14% How do you think this compares to what Mars, Inc. says is in an average peanuts M&M's bag? Figure 10 is the % of M&M's per bag on average based on the data from the web site. Figure 10: Color Percentages in M&M's 25% 20% 20% 20% 20% 20% Percent 15% 10% 10% 10% 5% 0% Blue Brown Red Yellow Green Orange Color Ask the class the following questions: Can we predict, based on our sample, what a bag of peanut M&M's may contain? If yes, what kind of SPC tool will tell us if a process is consistent and predictable? The answer to the last question, of course, is a control chart. Then: Divide the class back up into their teams.

11 Ask each team to pick a color and use this SPC tool to determine the amount of variation present in a bag of M&M's for that color. A blank control chart form is used for this. The average and control limits (based on the average subgroup size) have been calculated previously using the data collection sheet contained in Microsoft Excel. When the exercise is done, tell the class to return to their seats and debrief the exercise. When debriefing the control charts, ask the class the following questions: How did the averages on the control charts compare to what Mars, Inc says should be in a bag of M&M s on average? How much variation can you expect within a bag? Are there any special causes of variation? An example of a p control chart for one class is given in Figure 11. Figure 11: p Control Chart: % of Yellow M&M's per Bag Percent of Yellow 50% 45% UCL 40% 35% 30% 25% 20% Avg 15% 10% 5% 0% Avg.=19.6, UCL=45.1, LCL=None Bag Number Scatter diagrams are not covered in this overview to SPC. However, it would be easy to include scatter diagrams and plot the number of one color of M&M's against the number of another color of M&M's in the bags. CONCLUSIONS This session has introduced how to use peanut M&M's to introduce the concept of SPC. After using this approach at IDG, the feedback from participants has been excellent. The enhancements to the SPC portion of the training have helped evaluations improve. What s more important, however, is that the participants have been able to take the concepts back to the workplace and use them. We found a simple way to digest SPC. REFERENCES 1