In my last blog, I boiled quality management down to two overarching goals – consistency of product and process, and continuous improvement. Today, I’d like to use what I call the Quality beehive to demonstrate how all our quality programs interact and how each relies on the others to achieve the overarching goals.
I like looking at a beehive and noting all the hexagonal ‘boxes’. The hexagon has been proven to be a superiorly strong shape. I’m not sure how bees figured this out but they have a strong hive with a beautiful appearance. I insert a quality program into each ‘box’ just to help see that all our programs are interconnected. I thought today we could choose a program to see how data analysis helps us in achieving consistency and continuous improvement.
How to eliminate inconsistency: a practical example
Let’s look at laboratory testing of finished packaged products. In this example, we test each batch/lot for a variety of parameters – moisture, fat, pH, color, flavor, and others. Let’s presume that we are making a product that has a legal maximum for moisture. If we are above that maximum, our product cannot be sold in the marketplace. Being a good and shrewd business, we also recognize that water is our least expensive ingredient, so we want to be as close as we can to that maximum, without going over, so that we have more product to sell.
Can we analyze our moisture test results data to see how often we are exceeding the legal maximum? Certainly, but then what? Should we also look at each incident to find a root cause? Can we look at batch-making records and in-process batch test results to see if the out-of-spec finished product result could have been predicted? This may seem like a lot of work – and it surely is if the work is done manually.
Having collected all the data using software, this can be an easier task. Software can help to sort the in-process data and set it up for side-by-side comparison with finished product results. Software can quickly complete statistical calculations to help us understand how precise and predictable our in-process results can be in telling us if the finished product will be in spec (‘legal’ in our moisture example).
We may find that our typical in-process result is 0.1% higher than the finished product result with a standard deviation of 0.05. With this data, we can then set about creating an in-process specification that will result in the finished product always (with 99.5% confidence) being within the legal requirement. In addition, we can set our specification to achieve the maximum amount of added water (remember, this is our least expensive ingredient), which results in the maximum amount of finished goods available for sale in the marketplace.
It’s a win-win for quality (consistency) and operations (more saleable product). Use and share this data with your operations teams to help build rapport and teamwork between departments.
The beehive is a complex structure: strong and future-proof. And quality management can be the same when manufacturers realize how goals and programs are interconnected. Next time we’ll talk about other uses for software to help us improve consistency and drive the company to be ‘Bigger, Better, Faster, and make More’.
After a life-long career managing quality and food safety programs and working as a consultant, trainer and auditor, Bruce is now owner of Insight Food Safety Consulting. He is a member of the Institute of Food Scientists (IFT), the International Association of Food Protection (IAFP) and the American Society for Quality (ASQ). In 2016, Bruce was selected as an IFT Fellow – a designation recognizing outstanding contributions to the science of food and IFT. He is a published author and has served as a member of the editorial advisory board for QA and Food Safety Magazine since 2011. He has served as a Trustee for Feeding Tomorrow, the Foundation of the IFT, and currently serves as the Vice-Chair of the International Food Scientist Certification Commission.
