HKSQ Seminar "Using DOE in Service Quality"

We were honor to invite Dr. Scott Kowalski (Technical Training Specialist, Minitab Inc.) to give us a seminar named "Using Design of Experiment (DOE) in Service Quality. The seminar was organized by Hong Kong Society for Quality (HKSQ) and co-organized by ASQ Hong Kong. DOE is a powerful tool for exploring and optimizing processes. With the advancement of Lean Six Sigma, many of the statistical analyses that have been utilized in manufacturing have migrated to the service sector. This talk aimed to provide some real world examples of using designed experiments for Service Quality.

Dr. Albert Tsang (Former Chairman, HKSQ) gave welcome remark and introduced our speaker Dr. Scott Kowalski.

In the beginning, Dr. Scott Kowalski briefed the seminar outline included Minitab, Six Sigma & DMAIC, Industries, DOE and Examples. Dr. Kowalski said Minitab had developed at Penn State since 1972 and TechMax HK was one of Minitab Partners in the world. They had three key products included MINITAB, Quality Companion and Quality Trainer.

Then he briefed the real value of Six Sigma was orchestrating process improvement through the implementation of statistical techniques. DMAIC was then introduced. Scott found that DOE would be useful in A and I phases of Six Sigma. Moreover, he pointed out if error reduction was warranted within manufacturing industries, it should be considered even more crucial to service quality (e.g. Financial areas, Government sections and the delivery of Healthcare).

Scott mentioned his customer in different inductries such as bank, coffee, hospital, mining, etc. The following diagram showed some improvement processes in healthcare industry.

After that he introduced DOE which was not only for process optimization but it helped us to understand the impact of parameters change. One of his example was a Hospital which wanted to minimize the percentage of insured patients whose discharge was longer than 50 minutes leading to increased customer satisfaction, increasing the number of admissions and turnovers, and increasing profit.

There were four factors to be considered. They were Resident Time (ResTime), Porter Time (PortTime), Insurance Time (InsTime) and Pharmacists. The design was a 2^4 full factorial with 10 replicates to understand the variation.

Main effects and Interaction plots were showed.

Finally, Scott introduced the DOE Logistic Regression and summarized that it was not usually necessary to use all of the high power DOE tools but use main effects and interaction plots

Q&A session.
Scott replied that DOE was not design specifically for Big Data but Minitab was able to handle it. After DOE analysis, lean project would be one of follow-up actions.

At the end, Dr. Albert Tsang presented a souvenir to Dr. Scott Kowalski.

Reference:
HKSQ - www.hksq.org
Minitab Inc. - http://www.minitab.com/en-TW/default.aspx
P.S. I had attended Scott's Training in Hong Kong from 16 to 19 May 2007.