Improvement Insights Blog

Like a lot of you, I’ve gotten really good at some things I wasn’t very good at a couple of months ago. From a personal perspective, my cooking has gotten much better (thanks to online cooks I found such as Chef John). From a business standpoint, I’ve gone from meeting clients face-to-face to becoming good at conducting most of my business via online video meetings with clients. The online video meeting has its ups and downs, but it sure is nice to be able to meet with someone on the other side of the country without having to spend time traveling.

It was during one such meeting with a client in New Hampshire that I discovered an eye-opening analogy showing the difference between Run Charts and Control Charts to the uninitiated.

I had been working on and off with Jacques Bernard for years. Jacques is now the owner of LolaMae Manufacturing in Wisconsin, but I met him years ago when he and I were both delivery drivers for a small restaurant in New Hampshire. Though he had hired someone locally to help him out with Quality Improvement, he still occasionally brought me in to advise on things… and to get a different perspective. Generally, I agreed with the advice his local QI person was giving, but on this occasion, the conversation came to Run Charts.

“Run Charts?” I asked.

“Yeah, Run Charts,” he replied. “Now, I know that you recommended those other charts a while back…”

“Control Charts,” I offered.

“Yep, Control Charts,” he acknowledged. “Anyway, Vincent says that it doesn’t really matter, because one chart is just like another. They all show the same thing, and Run Charts are easier and quicker to create.”

I held my face in my hands for a moment and growled into my palms in frustration. “Grrrrrghhh!!!”

Jacques laughed; as I said, we’ve known each other a long time. “So I take it that you don’t agree with Vincent?” he chuckled.

I opened up my fingers and looked at the webcam (and the image of Jacques) though them. “Oh, you might say that…”

Jacques folded his arms. “So what’s the difference? I mean, I don’t understand this stuff. He does, and you do. So how am I to know which one of you is right?”

“Hmmm… where to start…” I pondered. “Okay, I think I’ve got a way to explain that will put things perfectly in perspective. Give me just one second…”

I began typing on my laptop, searching for the data I needed. In a minute or so I shared my screen.:

“Okay,” I asked. “Let’s say you had two production lines, and they’re both producing the same product. Let’s say I generated these two graphs for you. What could you tell me about your production lines just by looking at these two graphs?”

Jacques looked at the two graphs I’d shared. “Well, not knowing much about this stuff, it sure looks like Production Line 1 is running consistently. Production Line 2 is all over the place! I guess I’d start by finding out what the problem on Production Line 2 is.”

“Okay,” I continued. “That’s a reasonable reaction… except for one thing. Take a look at the Y axis on both graphs.” I paused while he could study the screen. “Would you believe that both of these graphs show the exact same data?”

“Wait, what?” Jacques paused. “Oh, I see. Line 1 goes from 0 to 12000, while Line 2 just shows from 9700 to 1200.”

“Exactly. It’s hard when things are displayed like this because it’s tough to put things into perspective. Now neither of these are a Control Chart, and neither are a Run Chart because they’re missing the Center Line. If I add the Center line to show the average on the 2nd graph, now it’s a Run Chart.” I did so:

“Now what can you tell me about your process?” I asked.

“Well… I suppose I can tell that we run, on average, between 10,000 and 10,100 units per batch,” he replied.

“Very good!” I encouraged him. “Now it looks like you have some variance from that average. What can you tell me about that variance? Is that good? Is that too much variance?”

“Um…” Jacques studied the Run Chart. “I really don’t know… not good, I suppose?”

“The truth is, you can’t tell what’s good or not good because there’s nothing that tells you on a Run Chart whether your variance is within expectations or beyond them. So now I want you to remember Vinnie’s Restaurant.” I said.

“Vinnie’s? The place where you and I were delivery drivers? How could I forget?” Jacques replied. “But what about it?”

“Remember those nights when we got those huge rainstorms? We still had to go out and deliver orders, and sometimes it would be raining so hard it was hard to drive.” I changed my screen to display a photo of just such a storm:

“Boy, do I ever remember those storms. It was impossible to drive! You couldn’t see the lines on the road.”

“Exactly,” I explained. “You couldn’t tell if you were driving in your lane or not, and you couldn’t tell if the cars coming toward you were driving in their lane or not. You just had some vague sense of where you were, but you couldn’t tell if you or the other drivers were  driving in the correct lane or not. Are you about to hit a parked car on your right, about to hit an oncoming car on your left, or are you fine? You can only guess, because you couldn’t see the lines on the road.”

“So what does that have to do with what we’re talking about?” Jacques asked.

“Because one big difference between a Run Chart and a Control Chart is the Control Chart puts those lines on the road so you can see how you’re doing.” Using QI Macros, I changed the Run Chart into a Control Chart and showed him:

“In order to understand the situation you’re in, you need to see the lines on the road. In a control chart, those lines are the Center Line (CL), the Upper Control Limit (UCL) and the Lower Control Limit (LCL). Those values are calculated from your data, and give an idea of what range of variation is normal and what variation is unusual.” I explained.

“Think of the UCL and LCL as the edge of your lane. You don’t want to go over those or you risk driving in the opposite lane or hitting a parked car. Those are spaced at +/- 3 sigma from your center line, and you can even show greater detail if you want. You can click on the chart itself in QI Macros and select “Show/Hide 1-2 Sigma lines”, and get a result that looks like this:”

“Aha! That makes sense!” Jacques exclaimed. “But aren’t the other charts easier to draw?”

“Well… if you’re drawing them by hand. Is Vincent drawing them by hand, or is he drawing the charts with software?” I asked.

“With software; that Q-something that you recommended.”

“QI Macros.” I answered.

“Yeah, that one.”

“So here’s how to create a Run Chart using QI Macros: Highlight your data, click a button, and there’s your Run Chart. Want to take a guess at how you create a Control Chart with QI Macros?”

“Highlight the data, click a different button?” Jacques offered.

“Yep.” I replied. “It takes exactly the same amount of time.”

“Wow! I’m so glad I called you. That makes so much sense, and now I can go back to Vincent and tell him why these are better.” Jacques sounded excited to learn something new.

“Vincent’s not a bad guy, he’s probably just repeating what he was taught. People taught students run charts for years back in the day. Now with computers doing the calculations and the graphing, it only makes sense to use Run Charts in certain specific situations. For everything else, a Control Chart tells you way more.”

“Thank you! You know what, just for that, I’m going to send you out some honey. My wife Lynn started a new honey bee colony recently, and she’s out collecting the honeycomb frames now.”

And that, my friends, is how I discovered both a great analogy and got the key ingredient for a fantastic Crispy Honey Sriracha Chicken Wings recipe. Not a bad day at all.

If you’re interested in learning more about Run Charts and turning them into Control Charts with QI Macros, click HERE to read about it on the QI Macros website.