Choosing Wrong Methods  

Measurement System Analysis projects are required for almost any organization that is collecting data for critical and high risk operations. This could include the building of airplanes, automobiles, pharmaceuticals, or product safety testing laboratories. The biggest problem is there are so many different standards to follow and every industry has their preferred methodology. For example, the American manufacturing industry relies heavily on the traditional Gauge R&R ANOVA study. However the European automotive industry uses Guide to Uncertainty of Measurement with a Quality of Measurement Process Decision Rule. The 17025 testing and calibration standard for measurement analysis uses the Guide to Uncertainty of Measurement as well.

The old fashioned Gauge R&R test was a great method back in the 1980s when the American auto industry was desperate to reduce waste and costs.

During the 1970s, there was an oil crisis in the USA that created a big demand for smaller more fuel efficient vehicles. The Japanese automotive industry was well positioned to handle the demand, but the US automakers weren't ready. The Japanese automakers had also developed more efficient manufacturing processes to eliminate waste and boost profitability. In an effort to compete against the Japanese, the American automotive industry had to introduce lean manufacturing methods to stay alive. This is where the Gauge R&R study came into the scene. 

Picture yourself in a machine shop in the late 1970s. There was a good chance that you had very poor lighting, the place was dirty, and the machine shop operators all came from different age groups and had their own opinions about everything. You probably had guys who were World War 2 veterans running presses alongside young men early in their careers. The single most used measurement instrument inside a machine shop, is something called a vernier micrometer. Vernier micrometers can have significant variation from one measurement to the next They're hand cranked by the operator and one operator can easily apply different amounts of pressure. They also require manual adjustments on a regular basis. Micrometers usually come with a small tool that allows the operator to change the "Zero" position of the micrometer. 

Usually these machine shops were custom building tools that were going to be used to build other parts. So at the machine shop level, precision would be extremely important. For example, if a machine shop rolled out a new tooling system which had a bad measurement, that new tooling system could have created countless defects in the final assembly from 1 single bad measurement. This can create giant expensive disasters for the entire supply chain. Not only could they have created defective parts, they could have also flagged out perfectly good parts as defective as well. 

The Gauge R&R was designed to create a standard statistical method to identify the difference between Operator Variation, Equipment Variation, and Part Variation by taking a sample size of 10 parts and have 3 different operators measure each part 3 times. The results had a built in decision rule of 10-30% 

Auditors Are Driving Up Costs With Dinosaur Methods 

As stated those Gauge R&R studies were never designed to be implemented in modern day highly-automated manufacturing processes. Lean Six Sigma gurus on YouTube are still pushing this method. Based on our personal experience, we've never seen someone who earns a living from measurement science ever use the Gauge R&R. Not any of the NIST scientists who study measurement science nor any of the measurement professionals at conferences would EVER go anywhere near this method. I've seen Gurus on Youtube teach Gauge R&R lessons and then can't explain why statistically perfect measurement processes fail the decision rules. They stumble over themselves trying to explain it. To make matters worse, you have auditors who never have done serious measurement science work in their life, going from plant to plant lecturing everyone on how to analyze measurement systems based on a audit checklist and a handbook they've never read. 

GUM Method Is Way More Compliant and Costs Much Less

The recognized international standard for statistically analyzing a measurement process is the Guide to Uncertainty of Measurement. This method was voted in by the top measurement scientists in the world who had a wide variety of backgrounds. It is fully compliant with the Automotive industry's measurement system analysis handbook, the VDA5 German automotive industry, and ISO 17025. 

The Gauge R&R Study can only calculate variation between operators and equipment, however there could be many other components of variation in the process. This could include temperature, vibration, humidity, calibration tolerances, calibration uncertainty, calibration bias,  and long-term stability. GUM Method allows for more components of variation that can help fine tune a measurement process. For more complex measurement systems, Gauge R&R is completely impossible to use effectively. 

Minitab MSA Tool Is Extremely Limited 

Most of the people we see performing Gauge R&Rs for MSA studies, are using Minitab's MSA suite to perform the calculations. A Minitab subscription is over $1800 a year per user and they don't even have GUM Method. If you have 5 people on your team, that will run you somewhere close to $1800 to $9000 a year just for software licensing. GUM Method spreadsheets are available for free or at extremely low cost and can be used without any licensing fees. 

The decision rules cannot be modified either, so that means all the users performing analysis on their tools are using extremely outdated decision rules. Minitab's software runs a serious risk of flagging perfectly good measurement processes as defective. You're going to confuse your management and customers trying to explain why your measurement process is failing even though you think its probably effective. The Gauge R&R MSA gurus on YouTube can't explain it, how are you going to explain it?

Number of Distinct Categories Is The Leading Cause of False Negatives in Measurement System Analysis

If anyone has ever performed a Gauge R&R study, you know it can be extremely time consuming and difficult to collect accurate data from the operators. Most production teams have absolutely no free time at all and collecting part samples can be a huge headache. Depending on the size of the tooling systems and size of the parts, it can actually create giant safety issues. More measurements are being conducted inside lock out/tag out zone, so you may have to comply with complex automated material handling and safety protocols. 

Imagine organizing this extremely complex test method, going through the process of collecting the data, and then inputting the data into Minitab, just to find out the entire test fails for not having enough distinct categories. Once you finish the Gauge R&R, you're then going to have to explain to everyone that your measurement process failed the test, NOT because its a bad measurement process, but because the data collected didn't have enough distinct categories. Nobody is going to understand the finished results and you'll have to scrap all of your data and start completely over. Minitab advises their customers to purchase more expensive measurement equipment in order to mitigate a possible false negative due to a lack of distinct categories or scrap the entire data set and start over.  Why would you modify your production processes on the basis of a possible false negative test result? 

GUM Method does not have any requirements for distinct categories, therefore resulting in less false negative test results. With GUM Method,  you can actually build a solid uncertainty budget of your production environment with just 1 or 2 parts. Also with GUM Method, you could simulate different equipment resolutions to determine if the gage resolution was actually the cause of the measurement variation. Therefore you could quantify the possible variation before even purchasing the equipment. You simply cannot do this with Minitab's software. Its not robust enough. 

Wrong Decision Rules Will Scrap Good Parts and Processes

Decision rules are the rules that an operation sets to determine if the measurement process passes or fails. Using the outdated Gauge R&R Method, the decision rules from the machine shops in the 1970s and 1980s are then going to be applied to completely unrelated measurement processes such as optical sensors and robotics. You cannot rely on these outdated decision rules and you need to calculate the rules on your own in order to avoid scrapping out perfectly good product. These rules need to be documented and standardized for your specific customer requirements. 

We Provide Measurement System Analysis Coaching

To help organizations improve the success of their measurement system analysis projects, we provide affordable coaching services that includes the following: 

• GUM Method Software Package 
• Analysis and Reports 
• Document Package That Includes Training and Reference Material 
• Coaching Calls
• Recorded Training Videos So Your Team Can Learn At Their Own Pace

We Visualize Uncertainty Error Compared to Production Tolerance

3 Step Process For A Successful MSA Project

Step 1.  Download a risk-free sample PDF report of our Measurement System Analysis Process. 

Step 2. We'll send you the free sample report and then reach out to you to schedule a 30 minute free consultation. We'll draft a proposal and send it. 

Step 3. We execute the proposal in a 30 day timeline sized project with an estimated budget of $2000.

Download The Free Sample Report Below