Let’s begin our discussion of control charts—the tried and true method for studying processes over time—with a simple statement: Control charts are useful tools
. No doubt about it. But I’m here to spread the word that viewing control charts do not have to be the first stop for operators. It just doesn’t. And I’ll explain why.
A control chart, also known as a Shewhart chart (after Walter A. Shewhart—a founding father of statistical process control, or SPC) or “process-behavior chart
,” is a tool used to determine if a manufacturing or business process is in a state of control. There it is. Thanks, Wikipedia!
In SPC, as we all know, there are statistical rules in place that state if a data point falls a certain distance from your mean, then something beyond what you expected probably happened. Or if you start exhibiting a trend, then you need to pay attention. The reason control charts were created was to give a visual indication of trends and abnormal patterns, to let you know when something may have changed from an expected baseline.
Control charts were designed to be easy to plot on paper. You take your measurements, you calculate a mean and standard deviation, and you plot those two statistics as points on their respective charts. Then you can ask yourself some questions. For example:
- “Did these most recent plot points fall too far away from the expected mean?”
If yes, then something strange happened, and someone needs to take action on the process.
- “Did two out of three of the last points fall between two and three standard deviations from the expected mean?”
If yes, then (again) someone needs to take action on the process.
So, the real purpose of a control chart is to answer, "Do I need to take action—Yes or No?" This was originally done when we were all using pencil and paper. Everything is done in small, simple steps. And after you draw that dot, then you can take a step back and say, "Is this dot too far away? Yes or No?" Or, “Is this dot, or two out of the three, in the wrong zone? Is there a trend beginning?” That’s why all these prescribed steps exist: so that a human being (without a degree in statistics) could benefit from powerful statistical analysis methods though simple calculations and plotting dots on a piece of paper…with the ultimate goal being “Yes/No, do I need to adjust my process?”
Is this making sense? Is it clear that somewhere along the line in manufacturing, there was a shift from art form to science? From intuition to educated decisions.
The Story…and More
When you’re using control charts you are looking at a story…where you’ve been, where things are trending, where things might be going. And you make necessary adjustments along the way. Control charts tell the story of your process. That may be the biggest reason that people still love them (everybody loves a good story).
And since we were all so entrenched in paper and pencil, to go along with the control charts, it was easy to fall in love with the tactile experience of holding up a control chart and drawing a line and saying something like, “At this time, between this point and this point, we switched to a new roll.” It’s another piece of the story that control charts tell; but it’s more because you’re hands-on, you’re making notes in the margin, you’re holding it up or pinning it to the wall, so everyone can see it.
Paper and pencil and control charts made a great combination over the years. There’s no denying it. But…
So, suffice it to say that paper control charts were all we had for a quite a while. And they served their purpose, and then some. When times change, and people move on to newer and often better technologies or methodologies, it’s important to note that it does not have to mean the end of the other. There is no end here for the analysis
that control charts provide.
What I’d like to discuss now is that you, the quality pro or the shop floor operator, should still use them, but emphasize the point that you do not need to put your eyes on them nearly as frequently as you’re used to
Control Chart Challenges
Let’s talk about some challenges when relying on users to create and/or analyze control charts. And there are some, believe me.
- I can put the “dot in the wrong place” on the chart. I can transpose numbers. I can be off by one grid line.
- I can make the process worse by trying to help. It’s human nature to want to fix things. Your shift just took over and you spot five ascending dots in a row (statistical process control (SPC) rule: six in a row means there’s an issue, and you should intervene); so, of course, you are inclined to want to fix it. That’s tampering with the process. And tampering makes your variation worse.
(I know, it sounds harsh; but that’s what it’s really called—you’re not supposed to touch anything until you hit that magic number “6.”)
- I can make the wrong decision. Having to review too many charts is overwhelming. What happens if your eye doesn’t pick up that trend? You might skip over a control chart that needs attention because you’re trying to wade your way through a big stack of charts.
So, here’s my point: as an operator, is it really your job to be an SPC chart interpretation expert? Or, is it your job to run the equipment and make product?
The Past is the Past
So, for years, operators have been asked to look at control charts because that was the only way they could get their immediate “Yes/No,” which was very necessary. Should I do something or not?
If you are
the person responsible for making an adjustment to a machine when something goes bad, when something is out of control, or out-of-spec, then you do
need that story—you need that control chart.
For us here at InfinityQS, it's not about ripping control charts out of people’s hands; we do not want to take away the benefits of that beloved tool. We’re saying, "Look, you don't need to worry about control charts…until you need to worry about them." We want to offer you something different.
A New Way
Let’s think about what we’ve talked about thus far: operators look at control charts to see when/if something is trending in the wrong direction, to see if something is off the rails, and then they have that magical “Yes/No.” They can decide whether or not they need to act.
Well, since that is the main objective for using the control charts, then we can help you put them away. InfinityQS Quality Intelligence solutions are here to monitor your data streams using your control chart rules—but do so in the background so you can focus on the many many other tasks you’ve been assigned and stop worrying about peeking at all your control charts every few minutes.
Our Quality Intelligence solutions can help you in several ways. They can:
- Alert you when you need to “do something.” When you receive that notification that informs you that some statistical rule has been violated, then you can turn your attention to the story (control chart) and see what action you need to take.
- Prioritize your control charts for you. If you have a stack of charts to look at, wouldn’t you want to start with only viewing charts that have violations and when time permits turn your attention to the other charts?
- Find your control charts for you. If you’re in an industry where you take many measurements (perhaps dozens) on each part, it can be very helpful to say things like:
- “Show me all of the control charts for OD – Loc A that have statistical violations.”
- “Show me all of the control charts from Lathe 167 for Part 123 and Part 124 for OD – Loc A, OD – Loc B and OC – Loc C.” (See an example of filtering control charts in the Stream Summary image below.)
, the InfinityQS Quality Intelligence platform, has a Stream Summary dashboard tile to do just these things for you. This gives you all the power of control chart analysis without having to constantly look at each and every assigned chart.
Adding to the Story
We’ve talked about how control charts tell a story, but they aren’t the only way to tell those stories. There are other tools that you should consider. Since we have tools that are constantly monitoring the data for this control chart analysis, we can focus our attention on things like exceptions and summaries. There are some great tools for this: None of these are designed to replace a control chart, but if you’ve got the right tools you can gain even more insight into your operations.
You’re now able to look beyond individual data streams (like you’d find on a control chart) and start looking at the bigger patterns and trends of your manufacturing processes.
- Process Event Paretos summarize events to allow users to see where problems are happening and “slice and dice” those by different criteria (e.g. site, feature, part, shift, etc.).
- Box-and-Whisker Charts are another tool that allows comparing different data in various ways that a control chart just can’t accomplish.
- Stream Grading is a way in which Enact can really simplify looking at process performance. You can easily zoom from a site level view to an individual data stream view to understand where process improvements can have the most impact on product yield. (The Stream Grading tiles are highlighted in the upper left corner in the image below.)
Closing the Loop
Let’s close all of this out by going back to our paper and pencil example. Imagine you’re creating control charts by hand. And you have to make 100 of them. Now, imagine you’ve got a new helper that said they will find you and tell you each time one of those pieces of paper has a violation. Further, they’ll divide all those 100 control charts into a pile of charts with violations, sorted so the charts with most violations are at the top of the stack, and a different pile for charts without violations. And they will instantly reshuffle and reorganize all the charts based on whatever criteria you have. Would you welcome that kind of help?
That’s exactly what you’ve got with a Quality Intelligence platform like Enact. Isn’t that a future worth embracing?
Please read more about Enact, and our other Quality Intelligence solutions, on our website!