Good morning, and thanks for reading Part IV of Highlights from a Keynote - Achieving Successful DMAIC. If you've been following my previous posts, you're now aware of two of the six key points I have found in common with successful DMAIC projects. Today's post focuses on the third one - Develop Trust In Your Measurements.
To illustrate this point, I'm going to shift gears from the automotive industry to the bowling industry, where I serve the United States Bowling Congress (USBC) as a Technical Adviser. As a side point, it is amazing how DMAIC quality tools can be applied to sport governance and the setting and enforcement of specifications for the equipment that makes the sport so great.
For this story, I remember the day I received a call from USBC regarding a manufacturer who had recently submitted samples for approval. The manufacturer had measured the samples before shipment to USBC and was confident there would be no issue in passing the certification requirement. However, when USBC measured the samples, the expected distribution of manufacturing was found to not pass the specification. Naturally, the manufacturer was not expecting this decision and insisted that USBC's measurements were not correct. Asking for my advice, I asked USBC to make sure that the manufacturer had measured with a gauge that was calibrated and had passed a gauge R&R assessment (I knew USBC had, because I was there when they performed it). When the manufacturer responded that they had indeed conducted and passed a gauge R&R and all their calibrations were up to date, I suggested an additional step. Since the manufacturer was located reasonably close to USBC headquarters, I suggested that the manufacturer pack up their gauge and bring it to USBC headquarters. The study that would follow had to be followed precisely.
First, the manufacturer would use their gauge to measure the samples, but with the USBC employees watching--and not saying anything. Next, the USBC employees would measure the same samples on its equipment, with the manufacturer watching--and saying nothing. Once the results were compared, there was a clear shift in the results--with the manufacturer's results always being the same amount lower than USBC's results. This explained why the manufacturer was anticipating a pass, while USBC deemed a fail--since the equipment performance was very close to the specification limit. (For this particular specification, all manufacturers produce some equipment as close as possible to this specification, as it produces a desired performance effect.)
Next, I had the manufacturer and USBC speak to each other about the differences in how the measurements were taken, and it turned out that the manufacturer was skipping one small but important step in the test procedure, which had been published by USBC. The skipping of this step led to the lower reading (we were able to explain it through physics). The manufacturer was fortunately able to make a small modification to their product, and the product then was determined legal for certified competition.
I've said it many times that differences in standard operating procedures are often the source of gauge and measurement problems, and this was no exception. So be sure to develop trust in your measurements! It's not only about calibration, but also about precision--repeatability and reproducibility.
In a few weeks, I will continue with part V of this series--Find and Eliminate the Contrast. Until then, I wish you successful problem solving.