Focus on Food Safety: Making sense of big data in food safety

You can’t control what you can’t measure” is an adage. Whether quoted from Peter Drucker or W. Edwards Deming, the principle is sound. Food safety and quality managers need to integrate controls into their system, and to control the system takes measuring and collecting data. But what do you need to measure? How much data do you need to collect?

In the era of automated sensors, PLCs, and interconnected enterprise management systems, it’s tempting to collect every bit of data. But why collect them all? I’ve worked with companies interested in finding all the things they can measure, and they end up with spreadsheets of millions of numbers that merely fill up storage instead of creating action for quality. Observational data, when used strategically, should turn into visual tools that help inform decision-making. Meaningful and actionable measurements will have a positive impact on the product and help reduce errors and waste.

Meaningful measurements

Every measure should have a purpose for improving the product’s safety or quality. Each product and process have unique characteristics, so it’s important to have product developers, operations, machine operators, quality and food safety specialists, packaging, shipping, and consumer feedback teams communicate to identify where and how quality is achieved. Each of these members of the manufacturing team have different insights into how the system maintains or loses quality. It’s the food safety and quality manager’s role to make sure those conversations happen.

While it’s tempting to take as many measures as possible with automation, frequency must be coupled with ability to act. Real-time measurements are great if they are linked to some sort of control mechanism such as a SCADA controller that automatically adjusts the system, or a mechanism like optical sorter that removes non-conforming product. If a human needs to interpret and intervene in the quality control process, make sure the frequency of measurement corresponds to the ability to trace and control the affected product, and intervene appropriately.

Making sense of measurements

Statistical tools and spreadsheets, such as those shared publicly by ASQ (www.asq.org/quality-resources/seven-basic-quality-tools), can help visualize data. Statistical process control charts are particularly useful. They allow for observation of the natural variation that occurs in food products and processes.

No system is perfect. There is always some variation, which is described as “common cause variation”. Special cause variation refers to a system that’s losing control. These terms were championed by statistician, Lloyd Nelson. The Nelson Rules help interpret when a process is losing control. Control charts can be used for all sorts of measures, from evaluation of ATP results in sanitation and monitoring microbiology total plate counts, frequency of different attributes or defects to tracking moisture and energy efficiency of operations. Some organizations will create spreadsheet templates or macro tools to help catalogue this data efficiently. Many ERP systems are building this type of analysis into their base programs.

Measuring in the qualitative space

Qualitative measures, such as colour, flavour, consumer attitudes, and workforce morale, contribute to quality. There are creative ways to convert qualitative measurements into numerical systems, such as frequency tables, colour space charts, or Likert scales.

Creating solutions

Automation can collect data for quality and food safety, and statistical tools can help visualize the information, but competent humans are the decision-makers. It’s important to use root cause methodology, take time to investigate and see the big picture of all the interconnected systems causing quality and food safety issues.

In root cause analysis, factors related to measurement, materials, personnel, environment, methods, and machines are investigated for their contributions to quality. Many quality practitioners also add finance and management to the root cause. Each of these factors is questioned and investigated objectively to find opportunities for improvement.

Humans are best suited to seeing the big picture and interconnected aspect of quality factors. Don’t let numbers overwhelm the system. Make sure your organization stays in control by using the right numbers and analyzing them well.  

Dr. Amy Proulx is professor and academic program co-ordinator for the Culinary Innovation and Food Technology programs at Niagara College, Ont. She can be reached at aproulx@niagaracollege.ca.

This column was originally published in the April/May 2023 issue of Food in Canada.