Six Sigma Answer to Material Shortages

One of Six Sigma¡¯s strengths is its facility for revealing causes and solutions that run contrary to our initial assumptions. When a persistent condition resists all attempts at improvement, or when an obvious fix to a newly discovered problem turns out to be lacking, a methodical approach like Six Sigma¡¯s can uncover even the most unlikely of causes and deliver results.

In the following case study the continuous improvement team was in for just such a surprise. Conventional wisdom was wrong, and the path the team started down hid unexpected complexities.

Definition

The XYZ Pump Garage program overall performance was poor. Future customer orders would not have been forthcoming without substantial improvements in quality and delivery.

• On-time delivery was 80% vs. >99% goal.
• Direct labor overtime was running 15% vs. a goal of zero.
• Field reported defects were found in 50% of system shipments vs. 0.5% goal.
• Project margin was approximately 22% vs. a 33% goal.

A process improvement team was formed with members from Customer Service, Manufacturing, Production Control, Engineering, Operations, and Purchasing.

Measurement

' Initial ' majority team consensus was that the program¡¯s poor on-time delivery was the result of material shortages due to understaffing in Purchasing. More buyers seemed to be the probable solution. The team suspected that the field defects were principally a result of poorly trained assembly staff.

The team began daily monitoring of data for number of daily kit shortages, overdue suppliers, and daily purchasing workload based on Material Requirements Planning (MRP) demands. Field personnel were interviewed for detailed descriptions of field defect rework.

Briefly summarized, the data showed:

• Typical labor overtime occurred near the end of the manufacturing process.
• 100% of all kits were issued with shortages.
• The key suppliers were >3 days late 50% of the time.
• The MRP system was posting material demands inside the material lead times!
• The requested delivery dates for material in the MRP system did not match well with project ship dates!
• A majority of customer-reported defects appeared to be the result of incomplete or incorrect manufacturing documentation.

1. Overtime was being worked to make up lost time due to late material deliveries.
   
2. Understaffing in Purchasing was not the problem! An army of buyers would not result in on-time material when the MRP ¡®buy¡¯ signal came too late or not at all. The team¡¯s true analysis problem was to understand why the MRP System was giving wrong signals. The team decided to focus on one specific sales order line item that exemplified the problem set for a typical system.
   
   What they found:

1. The sales order was coded incorrectly in a fashion that would generate several MRP problems.
2. Item master attributes were not properly populated for many of the material items that had MRP problems.
3. Customer engineering change orders (ECO) had been accepted without renegotiating product delivery dates with the customer to allow time for ECO implementation, including new material delivery.
4. A check of other customer order line items showed similar problems.
   

1. Customer ECO information was not being properly transmitted and propagated throughout the organization, resulting in out-of-date manufacturing instructions and field defects.
   
2. Problems would not have occurred if program participants had properly followed the procedures and work instructions documented in the Quality Management System.

Improvement

The improvements we implemented can be summed up in one word: training. The company had grown significantly during the past year and while all employees had received training, it had sometimes been rushed or had not been completely absorbed by the new personnel. Mandatory training was scheduled immediately for all Customer Service, Engineering, and Operations personnel on the documented procedures for sales order entry, customer engineering change orders, creating item masters, and creating engineering masters. Retraining took 7 working days with approximately 30 personnel participating.

ERP data for all active purchase orders was audited for the most common errors the team had recently discovered. This process required 5 working days.

New delivery dates were negotiated with the customer¡¯s buyer based on the new solid data foundation. This was difficult, but fortunately the customer¡¯s buyer is a mature personality with a long-term partnership attitude.

Results:

• Within 4 weeks material shortages had improved considerably.
• On-time delivery reached 100%.
• Overtime labor became negligible.
• After 8 weeks there had been no field defects found in the 6 systems shipped in the prior 5 weeks.
• Margin has improved to 28%, but this needs further investigation.
• Teamwork between organizations improved as a result of greater appreciation for the needs and complexities of their respective jobs.

Control

On-time delivery, customer field defects, and margin remain the bottom-line metrics for process control on the XYZ program. However, most importantly, as a result of the XYZ team findings, a new continuous improvement team was formed: the Enterprise Resource Planning Data Integrity Team (EDIT). EDIT is tasked with developing a set of strategies and process control tools to insure there are no repeats of the XYZ difficulties on other programs.

Implications

This single Six Sigma project thus had far-reaching implications for the XYZ Pump Garage program. First, in fulfilling the immediate purpose of improving our performance, we achieved customer retention for the near future. On a broader level, we also seized an opportunity to enhance our overall long-term approach to improvement. The value of reaching beyond obvious solutions having been so dramatically reinforced, we created a new continuous improvement team charged with making the pursuit of quality a more proactive endeavor.

Six Sigma Case Study: Defect Reduction in the Service Sector

by Chris Bott

This case study discusses the effective use of Six Sigma tools to improve our plastic issuance processes. It will take you through a project American Express completed, “Eliminate Non-received Renewal Credit Cards.?This analysis demonstrates how we applied Six Sigma techniques to reduce the defect rate with ongoing dollar savings.

Define and Measure the Problem
(Data has been masked to protect confidentiality.)

• On average (in 1999), American Express received 1,000 returned renewal cards each month.
• 65% (650) were due to the fact that the card members changed their addresses and did not tell us.
• The U.S. Post Office calls these forwardable addresses. Please note: Amex does not currently notify a card member when we receive a returned plastic card.

Analyze the Data

We applied various Six Sigma tools to identify Vital Xs, or the root causes of the defect. The use of Chi Square indicated the following:

• By type of card/plastic: We isolated significant differences in the causes of returned plastics among product types. Optima, our revolving card product, had the highest incident of defects but was not significantly different in the percentage of defects from the other card types.
• Issuance reason: Renewals had far and away the highest defect rate in the three areas in which we issue plastic—replacement, renewal, and new accounts.
• Validated reason for returned: Because we suffered scope creep early in the project, it was important to confirm what our initial data was telling us. After testing the five reasons for returns, returns with “forwardable?addresses were overwhelmingly the largest percentage and quantity of returns.

Improve the Process

An experimental pilot was run on all renewal files issued. This “bumping?against the “National Change of Address?service was implemented on all renewal cards in mid August. Due to the strict file matching criteria, this solution will impact 33% of the remaining population (or 333 cards monthly).

As a result of a successful pilot, we were able to reduce the defect rate by 44.5%, from 13,500 to 6,036 defects per million, reflecting annual savings of $1,228. Figure 1 outlines the combined test results.

Fig. 1 Combined Test Results

Non-Received Renewal Credit Cards	Baseline	Test Results
Defect rate	1.35%	.6%
DPMO	13552	6036
COPQ		$3,360
Total annual savings		$1,228
Sigma level	3.71	4.01

Control the Process

To ensure that we perform within the acceptable limits on an ongoing basis, it is important to monitor the new process. To achieve “control?status, we will be using the p chart, a tool that tracks proportions of returns over time.

In addition, our vendor has constructed reporting, which gives us the ability to monitor the defect rate on a monthly basis. The report will tell us if any credit cards that were “bumped?against the "National Change of Address" database were returned back to our warehouse.

Impact on Customer Satisfaction

Using the "National Change of Address" will enable over 1,200 card members to get their credit cards. Prior to this implementation, these card members would have never received their cards automatically. Revenue and customer satisfaction will undoubtedly increase.

Capabilty Analysis: Can a process produce output within spec?

A process that is in control is not necessarily producing an output that meets customer or engineering requirements. To find out if your process is capable of producing outputs that are in spec, you should perform capability analysis.

Capability analysis is a set of calculations used to assess whether a system is statistically able to meet a set of specifications or requirements. To complete the calculations, a set of data is required, usually generated by a control chart; however, data can be collected specifically for this purpose.

Specifications or requirements are the numerical values within which the system is expected to operate, that is, the minimum and maximum acceptable values. Occasionally there is only one limit, a maximum or minimum. Customers, engineers, or managers usually set specifications. Specifications are numerical requirements, goals, aims, or standards. It is important to remember that specifications are not the same as control limits. Control limits come from control charts and are based on the data. Specifications are the numerical requirements of the system.

All methods of capability analysis require that the data is statistically stable, with no special causes of variation present. To assess whether the data is statistically stable, a control chart should be completed. If special causes exist, data from the system will be changing. If capability analysis is performed, it will show approximately what happened in the past, but cannot be used to predict capability in the future. It will provide only a snapshot of the process at best. If, however, a system is stable, capability analysis shows not only the ability of the system in the past, but also, if the system remains stable, predicts the future performance of the system.

Capability analysis is summarized in indices; these indices show a system¡¯s ability to meet its numerical requirements. They can be monitored and reported over time to show how a system is changing. The main indices used are Cp and Cpk. The indices are easy to interpret; for example, a Cpk of more than one indicates that the system is producing within the specifications or requirements. If the Cpk is less than one, the system is producing data outside the specifications or requirements. This section contains detailed explanations of various capability indices and their interpretation.

Capability analysis is an excellent tool to demonstrate the extent of an improvement made to a process. It can summarize a great deal of information simply, showing the capability of a process, the extent of improvement needed, and later the extent of the improvement achieved.

Capability indices help to change the focus from only meeting requirements to continuous improvement of the process. Traditionally, the focus has been to reduce the proportion of product or service that does not meet specifications, using measures such as percentage of nonconforming product. Capability indices help to reduce the variation relative to the specifications or requirements, achieving increasingly higher Cp and Cpk values.

(Illustration) Measure Step-4

Measure step - 4

Instructor : The last tool we are going to look at is a Failure Modes and Effects Analysis, or FMEA, for short. FMEA is
Instructor : a structured approach to identifying the ways in which a process can fail to meet critical customer requirements, estimating the risk of specific causes with regard to these failures,
Instructor : Evaluating the current control plan for preventing these failures from occurring, and
Instructor : Prioritizing the actions that should be taken to improve the process. FMEA is an ideal approach to documenting and tracking risk reduction actions, particularly when the consequence of a failure is severe.

Instructor : There are a number of terms that have specific definitions when used in the context of an FMEA. These are three of the most critical.
Instructor : The Failure Mode describes how a part or process can fail to meet the specification. This is usually associated with a defect or nonconformance. Going back to the car example mentioned earlier, the failure mode would be brake malfunctioning.
Instructor : A cause describes a deficiency that results in a Failure Mode. Causes are sources of Variability associated with Key Process Input Variables. For example no brake fluid would cause the malfunction.
Instructor : The Effect describes the impact on a customer if the Failure Mode is not prevented or corrected. The customer can be downstream or the ultimate customer. In the car example the effect of brake malfunction could be quite severe, even fatal.

Instructor : So how does the FMEA work? In short, it identifies ways in which a product or process can fail, and then plans how to prevent those failures. The key steps are
Instructor : Identify potential failure modes
Instructor : Rate the severity of their effect
Instructor : Evaluate objectively the probability of occurrence of causes and the ability to detect the cause when it occurs.
Instructor : Rank the risk of failure modes according to a numerical index called the Risk Priority Number, or RPN for short.
Instructor : And, focus on eliminating product and process concerns and help prevent problems from occurring.

Instructor : The overall FMEA goal, much like the goal of a fishbone diagram, is to link the causes, via the Failure Modes, to the potential effects over a time span. Next, we'll look at exactly how you do this using the FMEA document.

Instructor : Now that you understand the FMEA purpose and process at a high level, in the next few screens we'll take a look at the entire F M E A process, from preparation through improvement. Throughout this section, remember that the FMEA is a living document that will be used throughout the life of the product or process, not just for this six sigma project.
Instructor : While we will start putting the FMEA together in this phase, some of the information will be added in the Analyze, Improve, and Control phases. So even though we are going to look at each area of the FMEA form in this section, you will fill in various parts of it throughout your Six Sigma project.

Instructor : FMEA completion begins with Preparation.
Instructor : First, you must select a process team.
Instructor : The team then develops a process map and identifies the process steps; this helps to define the scope of the FMEA.
Instructor : Next list the key process outputs necessary to satisfy internal and external customer requirements; this helps to define what is and is not a Failure Mode.
Instructor : Then list the key process inputs for each process step,
Instructor : and define the relationships of the product outputs to process variables; this helps to determine the effects and causes for the corresponding Failure Modes.
Instructor : And finally, rank the inputs according to importance in order to determine the severity of the risk.

Instructor : With all the preparation complete, you can address the heart of the FMEA process. Let's recap those steps in detail.
Instructor : List the ways process inputs, or causes, can vary, along with any other sources of variability, and identify associated Failure Modes and Effects.
Instructor : Assign severity, occurrence and detection rating to each cause, and calculate the Risk Priority Number for each potential failure mode scenario.
Instructor : After performing the FMEA, it's time to address the necessary improvements.
Instructor : Begin by determining the actions your team recommends to reduce the RPNs.
Instructor : Then establish timeframes for corrective action.
Instructor : Take the appropriate actions and put controls into place.
Instructor : Finally, recalculate the RPNs.

Instructor : We are going to use a simple example of an FMEA and show you how the Risk Priority Number can be calculated. We are going to look at our drive to work. Let's start by filling in the first row of our FMEA. The first step is to define the first process step and its potential failure modes.
Instructor : The first entry we make is in the Process Step or Part Number column.
Instructor : This column defines the specific process steps or individual parts we are going to examine.
Instructor : In this case, we will look at starting the vehicle as our first process step. The second column is labeled Potential Failure Mode.
Instructor : This is where we define what can go wrong, or otherwise result in a defect.
Instructor : Our first failure mode is simply when the vehicle won't start.
Instructor : The third column is labeled Potential Failure Effects
Instructor : This is where we will define the impact of the failure mode upon the customer.
Instructor : In this case, the effect is that the driver cannot get to work.
Instructor : The fourth column is labeled S E V, for Degree of Severity.
Instructor : In other words, how severe is the effect on the customer?
Instructor : In this case, we will use the scale previously displayed in the Tell Me More example, and say that the degree of severity is seven, high degree of customer dissatisfaction due to the loss of function without a negative impact on safety. This degree of severity will cover all of the various potential causes because it's linked to the failure mode, not one of the specific causes.

Instructor : The Fifth column is the Potential Causes column.
Instructor : This column is where you list the possible causes of the failure mode.
Instructor : In this case, we have listed two possible causes, being out of fuel and a dead battery, for the failure mode Vehicle won't start. You may often list more than one cause for a given failure mode.
Instructor : The column labeled O C C is where we put another score.
Instructor : This score is the Likelihood of the specific failure mode occurring due to that specific cause. It is based on the ratings standards you saw in the previous Tell Me More. This file is also available in the Tools section.
Instructor : In this case we have assigned a score of three, or a low failure rate, to the vehicle failing to start because it's out of fuel and a six, or a moderate failure rate, for the vehicle failing to start because of a dead battery.
Instructor : The next column is labeled current controls.
Instructor : It asks you to describe the current controls which would act to prevent the specific failure mode from the cause listed.
Instructor : In this example, we have Observe Fuel Gauge as the control to prevent failure to start due to being out of fuel and Check Battery Life as the control to prevent failure due to a dead battery.
Instructor : Next we have the D E T, or Ability to Detect field.
Instructor : This is where we enter a score indicating how well one can detect the Cause or Failure Mode.
Instructor : We have given watching the fuel gauge a score of two, indicating almost certainty that the cause of the failure will be detected. The control of checking the battery life receives a score of eight, which indicates that there is a very low likelihood that the potential failure will be detected before reaching the next step.
Instructor : The next column documents the RPN or the Risk Priority Number. RPN is an index used to evaluate the risk of the failure mode associated with the potential causes and corresponding controls. Risk reduction plan should be initiated to address those items with high R P Ns. This is an index that can be used to prioritize our project focus.
Instructor : This number is calculated from the three previous scores entered in the Severity, Likelihood of Occurrence, and the Ability to Detect columns. By multiplying those three numbers together we get a final result.
Instructor : Let me show you how to calculate the RPN associated with running out of fuel. Take severity rating from the SEV column multiply by the occurrence rating from the OCC column the delectability rating from the DET column
Instructor : This results in a RPN of 42
Instructor : In this case, the RPN score for failing to start by running out of fuel is much lower than the score for failing to start due to a dead battery. This would indicate that the second cause is at a higher risk and a corresponding risk reduction plan should be produced.
Instructor : The remaining sections of the FMEA should be used to document the identified solutions to reduce the risks. The RPN score should also be revised after the implementation of the risk reduction plan.

Instructor : The Risk Priority Number is calculated by taking the products of the Severity score, the Likelihood of Occurrence score, and the Ability to Detect score.
Instructor : In the case of one failure mode, all of the severity scores will be the same. In this case, they are seven.
Instructor : We regarded to likelihood of occurrence much higher for the dead battery than the fuel outage, so its score is higher.
Instructor : Likewise, we regarded the chance of the controls failing to detect the problem in advance to be far worse for the battery than the fuel level, so that cause also got a higher Detect score.
Instructor : If we multiply the three scores in each column, we come up with a Risk Priority Number, or RPN, of 42 for the vehicle failing to start due to an undetected fuel outage and 336 for the same failure due to an undetected dead battery.

(Illustration) Measure Step-3

Measure step - 3

Professor : In step 2, we established a definition for the process we are measuring and a standard against which we'll compare the performance. Now in step 3, we are ready to record the current performance of the process by creating a data collection plan, evaluating the measurement system we use, and properly recording the results. We come away from this step with valid data in a format that is ready for analysis in step 4.

Professor : When you've completed this step, you'll know the purpose of a data collection plan and will have established one for the Rockledge case.
Professor : You'll know the meaning of several terms related to measurement system analysis.
Professor : And, you'll recognize possible sources of variation in a measurement system.
Professor : You will be able to use the Measurement System Analysis checklist to guide your validation of a data source.
Professor : And, in preparation for the analyze phase, you'll become familiar with the software used to record and analyze data.

Professor : A data collection plan is the first step towards gathering accurate data.
Professor : It's intent is to provide a clear, documented strategy for collecting reliable data.
Professor : It gives all team members involved in the measurement process a common reference.
Professor : And, it also helps to ensure that resources are used effectively to collect only data that is critical to the project.
Professor : In some cases, new data collection might not be the only option. Look for any historical data that is available and consider the benefits of new data versus the costs of the collection process.

Professor : Well, talk about timing. Master just told me a very important message.

Master: Hey there, professor Do you remember when Mr. Alberti, the Rockledge General Manager, referred to some research that GE did on nut removal?
Master: Well, it seems that they collected several data samples on turbine casing nut removal and installation over a period of one year.Thought this information might be useful to you and the greenbelt. I'll send the data over to you.
Professor : If the data turns out to be valid, it would save us considerable time and money to use it instead of conducting another study. In order to determine the validity of the data, you need to first evaluate how the measurements were taken. We call this process measurement system analysis. We'll come back to the case in a few minutes.

Professor : Before we look at the Rockledge data, let's revisit the concept of variation.
Professor : Remember that in a perfect world, a process is done exactly the same way every time and every product that comes off an assembly line is identical.
Professor : But in the real world we have Variation or differences from the ideal. In a Six Sigma project, the variation we find in the process is called Actual Process Variation. Our ultimate project goal is to reduce that variation, thus satisfying our customers' needs.
Professor : So at this step in the Measure phase of D M A I C, we measure the output of the actual process. It is the output of the Measurement process that becomes our data.
Professor : But, collecting the measurement is a process itself, with the same potential for variation. Just like a scale on a production line might be slightly off, the timer that is being used to measure nut removal time might be malfunctioning.
Professor : Actually, this variation in your measurement process can come from a few sources.
Professor : The gage, or device, used to measure
Professor : The operator of the device
Professor : And other less common sources, such as the environment in which the measurement takes place. There's master again.

Professor : Instead of telling you how it relates, I'd like you to relate that diagram you just saw to our case. Here are the processes and some examples of potential variation in our Rockledge case. Drag these items to the appropriate place on the variation sources diagram.

Professor : There are three types of error that can result from the gage: accuracy, precision, and resolution. In this context, those words may have a somewhat different meaning than expected, so I'll walk through an explanation.
Professor : Let's address accuracy and precision first, by looking at the analogy of a target.
Professor : The circle here represents the ideal performance of a measurement instrument.
Professor : These dots represent the outcomes from a precise instrument, but not an accurate one. Because they are very close together in the same area, you can assume that the behavior of the instrument is predictable, but unfortunately it's predictably not inside the circle where you want it!
Professor : These dots represent the outcome from an accurate system, but not a precise one. They are all arranged around the center of your circle, so the results are all close to the target, but they are quite distant from one another so no precision here!
Professor : Finally, these dots represent outcomes from a precise and accurate instrument. They are close together, so the instrument is precise, and inside the circle, so they are accurate. With a precise and accurate instrument you get consistent results within the appropriate range. Problems with accuracy can be addressed by calibration of the instrument. Problems with Precision can be addressed using a method called Gage R and R which we will talk about shortly.

Professor : Now, let's take the third variation type for a gage -- Resolution. Here you see a ruler. Imagine you are supposed to measure the length of paperclips with it, and your target length is 1.234 inches. Would this be a good instrument for collecting the measurement?
Professor : The answer is decidedly no. Because the ruler does not have fine enough increments, the only items you could accurately measure with this ruler would be those falling at the exact inch marks. In other words, the resolution of the instrument is too low to collect accurate data.

Professor : Time to apply these terms to some real-world problems. For each gage problem shown, indicate whether it is an Accuracy, Precision, or Resolution error.

Professor : Now you know the ways that a measurement system can be responsible for variation, let's get back to the Rockledge assignment. We are going to conduct an evaluation of the measurement system that was used to collect that nut removal data. The Measurement System Analysis Checklist will guide us through our assessment.
Professor : The first couple of items look at the measurement procedure. We need to use the same procedure for our measurement system test that was used in the collection of the original data. Let's double-check the operational definition and collection procedure with the GE Facility Manager.

Professor : I have John on the line. Based on the operational definition he provides, we'll write the procedures for the measurement system test.

Professor : The next few items from the checklist prompt us to consider what we already know about the system. Answers to these questions are found in a variety of ways, depending on the tool used.
Professor : In our case, the instrument is a well-calibrated stopwatch with guaranteed precision and accuracy from the manufacturer. It measures to the hundredth of a second, so the resolution is more than sufficient based on our twenty-minute specification. Now that you know the procedure to follow and some details about the instrument, it's time to conduct the study of the measurement system to confirm that the procedures are correctly interpreted and the gage is indeed reliable.

Professor : Before we talk about the tests themselves, we need to answer the question "What are we testing for?" In two words, the answer is Repeatability and Reproducibility.
Professor : Repeatability means looking at variation within one component of the process. This is also known as equipment variation because it is most often evaluating if one operator measured the same items several different times with the same instrument, did the instrument produce the same measurements.
Professor : Reproducibility means testing variation across the process. This is also known as appraiser variation because it is most often testing if different operators measured the same items several different times, did the operators get the same results. A good way to remember these terms is there is an "e" in Repeatabilty for Equipment and an "o" in Reproducibility for Operator.

Professor : So, the last item on the list refers to the two tests that you can conduct to determine amount of variation caused by a measurement system. They are the Test-Retest study and the Gage Repeatability & Reproducibility study.
Professor : A test-retest study can look only at repeatability, meaning it can tell you how much variation in your data is due to an inappropriate device.
Professor : The advantage of Gage R and R is it can separate the individual effects of repeatability from those of reproducibility. Basically it shows not only variation due to the gauge, but also how much variation is due to the operators. This allows you to take action to fix the problem. For example, if you found a large amount of variation due to operator, you might improve operator training or the procedure descriptions.
Professor : For this reason, we are going to conduct a Gage R and R for the Rockledge case. It takes some time to collect the Gage R and R data, so I'm going to get that process started now.

Professor : There are 3 components to a Gage R and R test: Operators, Parts and Trials.
Professor : The operator, as you know, is the person operating the measuring device. It is recommended that you run the test using a minimum of 3 operators. The more operators you have, the more certainty you will have that your procedures are universally understandable.
Professor : The part is whatever product or process is being measured. It is recommended that you provide at least 10 representative "parts". By representative, I mean that the parts being tested should reflect the range of measurements possible. It is also important that the operators are all measuring the same 10 parts.
Professor : The trial is each time the item is measured. A minimum of 3 trials per part, per operator is recommended and the parts should be presented in random order to avoid any influence in the individual measurements.

Professor : Well, the data is in on the Rockledge study. I am going to conduct some statistical calculations on the test to determine repeatability and reproducibility variation, but I want you to know how the data should be recorded to allow that analysis.
Professor : In the Rockledge case, each part is the process of removing a single nut. In order to provide the SAME parts to each operator, we had to be a bit creative. We videotaped one turbine casing disassembly, and that gave us 88 examples of nut removal. We took 10 representatives from those 88 to use in our study and numbered the video clips from one to ten. That part number appears in this first column in random order.
Professor : We used the recommended 3 operators. The operator number appears in this second column.
Professor : We'll also did the 3 recommended trials per part, per operator, and that trial number appears in the third column.
Professor : In the last column, the actual nut removal time that the operators recorded appears in minutes. So now that the data is ready, I'm going to run the test.

Professor : Here are the results of the Gage R&R. The first number to check is the Total Gage R&R. From it, you can see that the total variation due to the measurement system is one point seven three percent. The rule of thumb you should use when evaluating the total is that less than two percent is desirable, and up to eight percent is marginally acceptable. So the bottom line is that our measurement system is adequate because it is not responsible for much of the variation. However, before we leave this subject, I want to explain the other numbers you see.
Professor : Under total gage R and R, you see the separations by repeatability and reproducibility. In this case the repeatability, or equipment variation, is only point two four percent. And the Reproducibility, or variation due to operator, is only one point four eight.
Professor : The remainder of the total variation in the data is the part-to-part. That is the actual process variation; or in this case the actual differences in the time it took to remove a nut. So, this test completes the Measurement System Analysis, and with these results we can feel confident that the measurement system is adequate to capture the capability of the actual process. Before you leave Measure, we have one more task, that is to make sure the actual nut removal data is in the proper form for analysis, much like we just did with the Gage R and R test data.

Professor : The goal of six sigma training is to give you the necessary skills to develop and improve business processes, not to turn you into a statistician. However, statistical analysis of the process is an important part of identifying and validating the improvements we make, so you need to be familiar with some statistical measures.
Professor : To make this easier, GE has purchased a tool called MiniTab that performs the calculations for you. MiniTab is a strategic software package that we will use to understand and analyze our data. It is part of the core load of General Electric. The Gage R&R analysis we just looked at is an example of what Minitab can do. Minitab was used to run a test on the data and generate those results.You can find more information about Minitab in the Resources section of this course.
Professor : Your first job is to put the data in correctly, so Minitab can do its work. This is much like what we just did with the measurement analysis data. Before you leave Measure, we are going to put the nut removal data into Minitab. Then in the Analyze phase you'll look at some of the reports.
Professor : The Worksheet area of Minitab is where you enter the data. It functions much like an Excel spreadsheet. Numbers can be entered in columns or rows; however, the default set up of Minitab is columns, so we will use that format.
Professor : Column titles are entered here and then the data below.

Professor : GE measured nut removal and installation time on eight scheduled outages during the course of a year. During each outage, time measurements were taken for twenty-two randomly selected bolts of the eighty-eight found on one gas turbine.
Professor : Here is the data. In this case, we have four data columns.
Professor : The first column indicates the sequential position of the measurement over the course of all measurement collection. For example, this is the forty-fourth measurement that was recorded in the year.
Professor : The second and third columns contain the measurements themselves. So nut removal time number forty four was twenty-three minutes and installation was fourteen.
Professor : The last column indicates the maintenance cycle. So measurement number forty four was part of the data recorded during the second outage. This format for the data will allow you to conduct an analysis in step four. If you wish to experiment with this data in Minitab, the project file can be found in the course Resources, accessed from the left-hand menu.

Professor : In this step, you learned the purpose of a data collection plan and determined the data to be used for the Rockledge case.
Professor : You've learned several terms related to measurement system analysis.
Professor : And, can recognize possible sources of variation in a measurement system.
Professor : You have learned how the Measurement System Analysis checklist can guide your validation of a data source.
Professor : And finally, you've become familiar with the role of Minitab in a greenbelt project and have prepared the Rockledge data for analysis in step four.

Professor : In preparation for Analyze, let's summarize where we are in the Rockledge project by taking a more visual look at the data.
Professor : A good starting point for data analysis is generating this histogram in Minitab. It shows you the frequency of occurrences for a given measurement, in this case removal times.
Professor : Based on the results of the Gage R&R, we know this data was collected with a valid measurement system. The graph shows that several measurements are over the upper spec limit of thirty minutes,
Professor : and most of the measurements are over the 15 minute target performance, so it appears that our current nut removal process is not doing too well. In Analyze, you'll verify that problem and look for some reasons why it is occurring.

Professor : Well, we did it, we're at the end of Step 3 and that means the end of Measure. Great job. Before you get into Analyze, I'm going to send you to back to Master. He wants to give you his brief review of Measure and see how you're doing.

Master: Hey, glad to see you back. Marks is a real Measure pro, so I'm sure you learned a lot working with him. I want to ask you a few questions to see how you're doing, but first I'll quickly run through the steps you covered and hit the highlights.
Master: In step one you took that CTQ and drilled down further to something more manageable for a greenbelt project. There are five tools to help with this process, and you used three of them on the Rockledge case: the Quality Function Deployment, Fishbone and Process Map. So by the end you had determined the task that you would focus your project on. You also learned how to complete a standard Failure Modes and Effects Analysis.

Master: Hey, glad to see you back. Marks is a real Measure pro, so I'm sure you learned a lot working with him. I want to ask you a few questions to see how you're doing, but first I'll quickly run through the steps you covered and hit the highlights.
Master: In step one you took that CTQ and drilled down further to something more manageable for a greenbelt project. There are five tools to help with this process, and you used three of them on the Rockledge case: the Quality Function Deployment, Fishbone and Process Map. So by the end you had determined the task that you would focus your project on. You also learned how to complete a standard Failure Modes and Effects Analysis.

Master: Finally in step 3 you determined your data collection plan by deciding between collecting new data and using historical data. You then learned the potential for error in a measurement system and some methods for evaluating a system. Finally, you learned how to set up data in the Minitab strategic software package for analysis in Step 4.

Master: That was a quick summary of the key steps in Measure. If you want further review, you can access all of the steps from the Measure Menu at the top of the screen, and the Resources, Glossary and Tools can be found in the left-hand menu. Once you are ready, I'm going to ask you some questions about the Measure phase.

Master : In step one you learned about five tools that can be used to help refine the focus of your project. Each tool has its own specific strengths for that process. Match the tools with their description by placing the letter of the tool in the field next to its description.

Master:Do you know how to use complete a QFD? An empty matrix is shown below. Imagine you are involved in a greenbelt project for Finance group. You are beginning to drill down on the customer CTQs. The items labeled a, b, c and d include a CTQ, the importance rating of the CTQ, one internal process that impacts the CTQ, and the relationship rating between the CTQ and the process. Drag label to the highlighted area on the QFD where that item belongs.

(Illustration) Measure Step-2

Measure step - 2

Professor : In step one of Measure, you determined the specific sub-process that is now the subject of your greenbelt project.
Professor : In step 2, you'll create a standard for the performance of that process. This is the point when you determine the best way to turn what the customer wants into a numeric measurement. This measurement will later be compared to the measurement of your current process in order to see how well you're meeting the customer's need.

Professor : In Step 2, you'll become familiar with the components of a good Performance Standard, including the operational definition, specification limit, target performance, and defect definition.
Professor : You'll also learn about the two types of measurements that you can make-discrete or continuous-and recognize examples of each.
Professor : Once you understand all those concepts, you'll be able to write a performance standard for our Rockledge Plant case.
Professor : Why write a performance standard? We already know that reducing the nut removal time will support meeting the customer's C T Q of increasing plant availability, so why don't we just get started?
Professor : This is where the numbers become important. For one thing, what does "reduce" mean? If we reduced it by twenty seconds would that be good enough, or do we need to reduce it by twenty minutes? The performance standard translates the customer need into a clearly defined, measurable characteristic.

Professor : A good performance standard includes: an operational definition of the process, a target performance, specification limits, and a defect definition.

Professor : Let's start with the Operational Definition. It's purpose is to remove ambiguity so that everyone, including the customer and all G E staff involved, has the same understanding of the process being investigated and agree on how to measure it.First, let's tackle the question, "what is the process?" by looking at an example from the airline industry.
Professor : Airlines know one of their primary customer C T Qs is "on-time flights."Here is a sample of how a customer defines "on-time."
Professor : Here is how the airline defines it. Can you see a problem?
Professor : The customers are concerned with getting in the air, but the airline is only focused on getting them onto the plane. The result is that the airline is advertising ninety-nine percent on-time departures, yet the customers still are unsatisfied because they're sitting on the runway for two hours before the plane takes off! Now you see how important it is to get customer agreement to the definition of a process.
Professor : In this example, we'll adopt the customer's definition of on-time departure.

Professor : Let's start with the Operational Definition. It's purpose is to remove ambiguity so that everyone, including the customer and all G E staff involved, has the same understanding of the process being investigated and agree on how to measure it.First, let's tackle the question, "what is the process?" by looking at an example from the airline industry.
Professor : Airlines know one of their primary customer C T Qs is "on-time flights."Here is a sample of how a customer defines "on-time."
Professor : Here is how the airline defines it. Can you see a problem?
Professor : The customers are concerned with getting in the air, but the airline is only focused on getting them onto the plane. The result is that the airline is advertising ninety-nine percent on-time departures, yet the customers still are unsatisfied because they're sitting on the runway for two hours before the plane takes off! Now you see how important it is to get customer agreement to the definition of a process.
Professor : In this example, we'll adopt the customer's definition of on-time departure.

Professor : So, once you know what it is, the next step is determining "how do I measure it?" Let's begin by talking about data types.
Professor : You need to be aware of two types of data, Discrete and Continuous.

Professor : Discrete values can only vary by a finite amount, which cannot be further subdivided.
Professor : For example, counting the number of people that walk into a room in an hour is collecting discreet data because one is the only unit of measure for a person . A half-person couldn't walk in, right?

Professor : Continuous data, on the other hand, can be refined to any degree of exactness. Time is a common example of continuous data. We often put convenient constraints on time, measuring it in years, months, days, hours, etc. But, in reality, you can refine those measurements to infinity.
Professor : Think of the Olympics where runners win by one thousandth of a second - now that's detailed! And, if you had the right tool to capture the measurement,
Professor : you could refine that number even further. How far we refine a measurement depends on the needs of a given process.Oh, there's the master again.

Professor : The C T Q does not dictate the data type; it is up to you to determine the type of measurement you collect. Here's a table showing some C T Q types and examples of discrete and continuous data for each.
Professor : You should always aim for collecting continuous data because with it you can more accurately assess your process.

Professor : What about our airline example? The goal is to assess the on-time performance of an airline and use the information to seek possible improvement. Which of the following approaches would be better suited for that task?

Professor : So we've talked about how an operational definition answers the questions "What is it?" and "How do I measure it?"
Professor : To determine the other three components of a performance standard, you need to first ask the question, "How much variation in performance will the customer tolerate?" Is a fifteen-minute late departure too late to satisfy their need? The answer to this question will set your specification limits.
Professor : The answer can come from a variety of sources, but should always reflect the voice of the customer. In this case, let's say that a customer survey determined that customers would tolerate a flight leaving the ground up to ten minutes late. Your upper specification limit then becomes the scheduled time plus ten minutes.
Professor : In many cases you will have a lower specification limit that mirrors the upper limit. But in this case, a flight leaving before the scheduled time would not make customers happy because they could miss it, so the lower specification would default to the scheduled departure time.
Professor : The Target Performance is where we will aim our process. In a perfect world there would be no variation in the performance of a process or creation of a product. In other words, it would be done exactly the same way every time.
Professor : So, in that perfect world we could say our target performance for the airline is departure at the scheduled time and we would get that each time.
Professor : But, in the real world, there is variation in performance, sometimes you're a few minutes early, sometimes a few minutes late.
Professor : The goal of six sigma is to reduce that variation in performance so that you are as close to the target performance as possible, thus providing a high degree of customer satisfaction. You'll learn more about the statistical concept of variation in the Analyze section of this course.
Professor : The last component of the performance standard is a defect definition. A defect is any nonconformance or any item outside the specification limit. Another way is to think of a defect as anything that results in customer dissatisfaction.
Professor : In this case, a defect would be any departure time before the scheduled time or more than ten minutes late. In Analyze, you'll hear about Defects per Million Opportunities-or D P M O. Each creation of a product or performance of a task is an opportunity. In this case, each flight that takes off is an opportunity for on-time departure. Whew! I know that was a lot to cover, but I wanted to make sure you understood what a performance standard is all about before we work on master's Rockledge case.

Professor : Based on what we know, what question do we need to answer first to begin writing the performance standard for the Rockledge Plant?

Professor : Coming back to the Rockledge case, we know that the high level goal is to reduce the scheduled maintenance cycle time, so our project goal is reducing nut removal time. For the performance standard, we need to turn that goal into a numeric measurement. Select the best measurement for the Rockledge case.

Professor : Master has reviewed existing GE procedures on this process and talked with Mr. Alberti, Mr. Frank, and some GE Millwrights about the nut removal process. Here is the official GE operational definition, which the customer has also agreed to.
Professor : The nut removal process is the actual time measured from when the millwright places the wrench on the nut, until the nut comes out of the socket. The customers have also provided some other useful information.

Professor : So, here's what we know from the customer. Let's wrap up this performance standard.
Professor : First, I want you to try it yourself. Use the information we've gathered to select the specification limits, target performance, and defect opportunities per year from the drop-down lists.
Professor : These are your answers. Let's talk about each component and you can see how you did.
Professor : The upper specification limit is 30 minutes because after 30 minutes the nuts are being cut off, and that is costing money and impacting the length of the outage.
Professor : The lower specification in this case is not applicable because the faster they get them off, the better.
Professor : That makes a defect any time over 30 minutes.
Professor : You also heard that an expert can do this in 15 minutes, so the customer is telling you that based on historical data of best practice the target should be 15.
Professor : Finally, there are 88 bolts per turbine and you've got 2 turbines in the plant, so that gives you a total of 176 opportunities each maintenance cycle.

Professor : In Step 2, you became familiar with the components of a good Performance Standard, including the operational definition, specification limit, target performance, and defect definition.
Professor : You learned to recognize the two types of measurements that you can make-discrete or continuous-and to aim for continuous data as a measure of performance.
Professor : Once you understood those concepts, we were able to write a performance standard for our Rockledge Plant case.

Professor : So, what we've learned about the nut removal process in step 2 of Measure is this:
Professor : Our operational definition of the process is the actual time measured from the wrench being put on the nut to the removal of the nut.
Professor : And, our standard for performance of this process is a time of 30 minutes or less for removal, with a defect being any time over 30 minutes and an opportunity being each time a nut needs to be removed, and with a target performance of fifteen minutes.

Professor : We're at the end of Step 2 and now we know our project focus and have set a standard for its performance.