July 25, 2017
Don’t Just Collect SPC Data: Interrogate It!
Your Statistical Process Control (SPC) data is dying to tell you something important.
There’s been a crime. You saw the perpetrator! Okay, well, maybe not a crime, per se, but…
There’s data from production line 3 that is showing something is out of spec. Money has been lost. Quality has suffered…and that’s criminal.
What to do? Well, to get to the bottom of the crime, you’ve got to interrogate the data. Put that data under the bright spotlight, sweat it, ask it a thousand questions, and listen closely to the answers. What part were you pressing on the night of the 14th? What were the other machines in your plant doing at the time of the incident in question? What shift did it happen on and what raw material were you using? Do you own a weapon? (Okay, maybe that last question went a bit too far. But you get the point.)
Your organization collects tons of quality management data every minute, every hour, every day. Your operators can see the data that’s important to them on a real-time control chart, pinpoint the issues they are having, and act accordingly–fix it or submit a corrective action to fix it. That’s the front line of data consumption, out on the shop floor.
But you haven’t really asked the big questions yet. You haven’t dug in there and forced the data to spit out the salient points of information you really need to keep all your manufacturing processes running as smoothly as possible. What you need to do is… wait for it…
Interrogate the data.
What is Data Interrogation?
Data interrogation is quality geek heaven. It’s beyond what the operator on the shop floor sees and reacts to. It’s about stepping back to obtain the broader view of your operation–seeing the entire scope of what all your production machines are doing, but in a context that enables you to maximize product quality and minimize downtime and overall costs.
When we interrogate the data, we’re looking at a lot of data—a week’s worth, a month, or a quarter—across several or even all of your production lines. Essentially, it’s about summarizing a bunch of data by placing it all in one big bucket. Imagine that bucket sitting on a chair in a small, windowless interrogation room. You’re there, shining the bright spotlight on it, making it sweat, and trying to decipher what you see in the bucket. What’s in there that you didn’t know before? Let’s ask some questions of the data, and let’s learn something new about our operations.
The exciting stuff (if you’re one of those aforementioned quality geeks) is what we find out. All the machines that run the Acme-supplied raw materials are running heavy compared to target, but line 3 is running light. Line 1 has the least variability, regardless of supplier, while line 8 generates the most variation despite running multiple different products.
Because all the data is in one bucket, we can see it all and we can ask any question we want. And we can dig deeper, if we want to. We can assess how a single mold performs on different production lines. And we can review performance differences between mold cavities, how they differ line-to-line or from weekday to weekend days, if that’s what we want.
And because we step back frequently, we can see that during week 3 of every month, the parts run a little heavier than the other weeks. We’re able to spot trends, see potential issues develop before our eyes, and anticipate problems before they even occur. We’re finding out who the perp was hanging out with before the alleged incident, what led to the crime, and how we might prevent it from happening in the future.
How Often Do You Analyze Quality Data?
Well, that’s entirely up to you. Weekly? Monthly? Quarterly? Whatever works for you and your organization. And it may change according to your specific situation. But the important thing is to regularly and consistently step back, roll up lots of SPC data, toss it into a single bucket, and interrogate it.
Is Interrogating SPC Data Worth Your Time?
You can’t afford not
to interrogate your data…if you want to save money and run at optimum efficiency. Speaking from experience, the most successful organizations I have worked with regularly interrogate large amounts of data. If your company mantra is, “our quality is no worse than our competitors,” then you may not be motivated to do so. However, if you want to see big improvements in quality across your organization while enhancing your bottom line, you need to interrogate data on a regular schedule.
When you interrogate your quality data, you have detailed information available to you that you just didn’t have before; much more than the operator sees in a control chart. And that’s invaluable. Well, let’s say that it’s really, really valuable.
That perp in the chair under the bright light has information that can save your organization money. And he knows how your operation can be more profitable and competitive. That’s valuable. That’s the result of interrogating data.