Brushing Up on the Fundamentals of Statistical Process Control
Returning often to the fundamentals is essential to a host of endeavors, from athletics to arithmetic to business. Basketball players shoot endless free throws, mathematicians remain grounded in the logic and laws of mathematics, and businesses keep vigilant focus on their P&Ls. Why? Because remembering the foundations of your pursuit is necessary to move to the next level, to win the championship, break the Riemann hypothesis, or become a Fortune 500 operation. Statistical Process Control (SPC) should be no different. Quality sat down with Doug Fair, Chief Operating Officer at InfinityQS, to talk about the keys, trends, and fundamentals of SPC.
Quality: What trends have you noticed with SPC?
Fair: These days, consumers expect more. Lots more. They expect higher quality in every product, no matter how inexpensive. Low cost is not necessarily correlated with low quality as in decades past, and this has helped to dramatically increase consumer expectations of quality. As a result, more manufacturers have become serious about improving process and product quality, and doing so on a larger scale. These companies want to expose more information about quality challenges, where they reside and what to do about them, and they recognize that statistical process control can provide them a great strategic advantage in that regard.
These companies look at real-time SPC for timely, actionable information about process performance and real-time insights needed to immediately solve issues with product quality, safety, and consistency. While the challenges aren’t necessarily new, the external forces created by ever-escalating quality expectations from consumers is motivating companies to turn to the tried-and-true advantages of SPC.
No matter what product a company makes, product quality issues can present a huge risk. Those risks include expensive regulatory noncompliance, threats to brand reputation, and the potential to cause consumer injury or death. In today’s hyperconnected world, a single incident or recall can quickly ignite a viral firestorm, making it difficult to regain consumer trust and business.
Real-time SPC presents an ideal solution to mitigate these risks, leveraging data collected off production lines to preemptively catch issues, reduce variation, and improve product quality and consistency. Beyond manufacturers, some regulatory bodies have also recognized the benefits of real-time SPC and even call out the need for the methodology. One example is the Safe Quality Food Institute (SQFI), whose highest level of certification for a food quality and safety program requires the use of SPC.
Quality: How would you describe SPC to a newcomer to the industry?
Fair: SPC is an industry-standard tool that helps operators control quality during the manufacturing process. Operators regularly measure product dimensions in real time, then plot those values on a graph with control limits. Control limits identify how a process normally operates, and they have no relation to fitness-for-use criteria such as engineering specification limits.
If a measurement value falls outside of control limits, it indicates that something “abnormal” has occurred. Plot points that fall outside of control limits communicate to an operator that a significant change has happened in the manufacturing process. These “out of control” events provide information to operators that can help them better understand the manufacturing process and prevent similar events from occurring in the future.
As a result, the control chart helps operators to ensure product consistency at the time of manufacture. Control limits are different from engineering specification limits, which are gates that identify “pass-or-fail” criteria and do not provide critical information about how a manufacturing process is currently running.
Data that fall within control limits indicate that the process is operating in a consistent fashion, and that nothing unusual has occurred. When an out-of-control event is indicated, operators check their machinery, materials, and various settings to determine what changed, so that corrective actions can be taken to prevent defects. Therefore, SPC not only helps generate information about how manufacturing lines are running, but SPC serves also as a tool for preventing quality problems.
Real-time SPC systems—like the Quality Intelligence solutions offered by InfinityQS—help manufacturers do more than point out possible plant-floor problems. We provide analytic tools that manufacturers use to mine their quality data for meaningful insights into ways to improve product consistency, reduce scrap and waste, increase productivity, and drive down costs. In the end, SPC enables you to better understand process behavior so you can continuously improve results.
Quality: What applications areas have been popular with customers?
Fair: One of the most popular applications of SPC in the food and beverage industries is controlling net contents—the weight or volume of a product placed in a package. On a high-volume, high-speed filling line, it can be difficult to ensure that the actual contents of every package match what’s stated on the package label. To avoid underfilling—which can lead to regulatory issues and unhappy customers—manufacturers will often resort to some degree of overfilling. Imagine regular overfilling on multiple lines and across several plants. Even the smallest amount of overfill can quickly add up to significant product giveaway, production inefficiencies, and wasted resources.
For food and beverage companies, SPC helps better control net contents and refine the filling process. But SPC can be used in any manufacturing industry. An effective SPC solution automatically presents insightful, yet easy-to-understand summaries of performance to operators so they can make informed decisions and take appropriate actions. Built-in alerts notify relevant team members (managers, engineers, and other support personnel) if data show troubling statistical trends. Responsible staff can then investigate further, drilling into the data to identify root causes of issues. Remedial steps can be logged to support continuous improvement initiatives and mitigate future costs and risks. This same data-driven methodology is applicable to virtually any production process.
Quality: Any misconceptions about SPC?
Fair: A large misconception about SPC is that it is solely a plant-floor analysis tool. While history has shown that SPC is a powerful tool when in the hands of an operator, I believe that only 15% of the benefit of SPC is realized using control charts. The remaining 85% is realized by summarizing, rolling-up, and analyzing shop-floor data.
The same real-time data that highlight issues to plant-floor operators can serve a higher purpose when aggregated. Data aggregation allows C-suite executives, managers, directors, engineers, Six Sigma teams, and quality professionals to have visibility over quality throughout their entire enterprise. They can run performance analyses between different lines, products, plants, and regions to find opportunities for process improvement and determine where to prioritize their improvement efforts and focus Six Sigma resources. Such operational insights support a big-picture, profit-positive business model with lower production costs, optimal and consistent product quality, and fewer defects.
Another common misconception in terms of SPC is ignoring data that fall within specifications. This seems counter-intuitive to most people. Generally, managers ask me, “Why should I look at data that meets engineering specifications? If they’re in-spec, that means the product is good.” My answer is yes, it does indicate that the product meets the spec, but data within specifications can still reveal extraordinary information that can greatly improve quality and costs. My experience has been that big process improvement opportunities are hidden within in-spec data. Ignoring data that meet engineering specification limits, and you run the risk of missing out on enormous benefits.
One beverage manufacturer that InfinityQS worked with discovered this when looking at container fill volumes on its bottled drinks. While all the net contents were within specification limits, detailed analysis revealed that the manufacturer was overfilling every bottle—and not by a small amount.
By analyzing the in-spec net contents data, the client identified enormous discrepancies in fill-head performance, product-to-product differences, and the effect of speed on filling. But because historically they had concluded, “it’s in-spec, so no need to evaluate the data,” they missed out on big opportunities for improvement.
Through the insights they gained by analyzing their in-spec net contents data, the beverage manufacturer made improvements and production line changes that generated more than a million dollars in yearly savings—on just one filling line. The company then used similar analyses on 20 additional fill lines, and the savings became exponential. Even though data may be in-specification, analysis can prove profitable for any manufacturer looking for opportunities to better understand and improve operations and their bottom line.
Quality: What should customers know before adopting SPC?
Fair: There are a few things to keep in mind before adopting SPC:
- Drop the pencil and paper: In one study conducted by InfinityQS, we found that 75% of manufacturers today still rely on manual data collection, with half of those polled reporting that they still use paper checklists. Unfortunately, pencil-and-paper data collection is not conducive to uncovering meaningful improvements. Besides being highly prone to error, it’s hard to take timely or proactive action based on data written on paper—or even those transcribed into digital spreadsheets. There are also considerable costs associated with maintaining records for audit purposes, including costs for labor, paper products, and storage space.
- Consider using automated or semi-automated data collection methods—as well as mobile-based data entry—for a far more efficient, accurate, and cost-effective data collection process. The data become more actionable, as well as easily accessible for analysis or for auditor-requested reports.
- Standardize across sites: Standardization is essential to take full advantage of SPC and look at quality and performance on an enterprise-wide level. First, there is standardization to make sure naming conventions for parts, features, processes, products, and lines are consistent from site to site. This eliminates confusion between teams and supports cross-plant analysis. It’s also important to standardize on a centralized data repository, making sure everyone saves to, and works from, the same data pool.
- Leveraging the cloud, a centralized data repository breaks down silos between manufacturing sites, enabling true enterprise visibility at the executive level. There are also other considerations to standardization, including metrics, analysis methods, and overall work practices. Ultimately, you need consistency across the board to be able to get the big picture of what each segment of your enterprise is doing, how they’re doing, and where the issues are.
- Obtain operator buy-in: Plant-floor operators are key to the adoption of SPC. As the primary individuals working on the production lines and gathering the data needed for enterprise-wide quality analysis, operators need to fully embrace the SPC system that management has selected. Thus, you want to make sure you select and deploy a system that is user-friendly and intuitive. Perhaps most importantly for system adoption, the selected SPC software needs to make operators’ jobs easier, and automatically generate the information they need to run their processes more efficiently and more effectively. Operators should be better equipped to identify issues in real time and thereby drive continuous improvement efforts that contribute to the organization and its bottom line.
To learn more about how real-time SPC can help you improve quality, reduce variations, and optimize processes, visit www.infinityqs.com/spc.