Improvement Insights Blog

I think it was Will Rogers who once said, “A half truth is like a half brick – it carries better.” Being a Freelance Quality Improvement Specialist, I’ve had to overcome more than my share of half truths in my career. An example of one of the most frequent half truths I encounter happened just this month.

While consulting with a plant that produced industrial cardboard, I asked them to send over their data. They submitted the data they were tracking: each tab in the spreadsheet had almost a dozen metrics, and each metric had almost 6,000 data points… some dating back over 4 years ago. I contacted the client on a video call and and asked about the size of the data set.

“Charlie, I just got the data file,” I opened.

“Yeah? Impressive, isn’t it?” asked Charlie, the foreman. He looked proud. “I bet none of your other clients have as much data as we do, do they? I told you: We’re serious about improvement. That’s why we keep so much data.”

“Okay, that’s fine,” I offered. “However, I’m not going to need to use much of this in order to move forward. We need to get a plan in place to limit the amount of data you’re using to track your processes.”

“What do you mean?” asked Charlie. “We need to track all of this data. We have this data; why wouldn’t we track it all?”

“Because you have too much data.” I replied.

“How can you possibly have too much data?” was the response.

Charlie had fallen into one of the classic half-truths: “You can never have too much of a good thing.” Everybody knows this isn’t true. If the recipe calls for a tablespoon of salt and you put in 2 cups of salt, it won’t taste better. Still, it’s tough for people to know when the “rule of thumb” they learned doesn’t apply any longer. I thought quickly and looked up some figures while I had Charlie on the video call.

“Charlie, I’ve seen you wearing your Green Bay Packers jersey around at Pareto’s Big Bar watching the games, right?” I asked.

“You sure have. I was born and raised there in Green Bay, Wisconsin. It’s in my blood.” he replied.

“So you know Green Bay weather, right?” I asked.

“Of course! It’s a beautiful place. The summers are pleasant, and the winters drive away everyone but the die-hards,” he chuckled.

“Charlie, I’m going to share my screen,” I said. When I clicked the button, this appeared on his screen:

“Charlie, I found a listing of a year’s worth of daily high temperature data online. This is from a couple years back, but would you say that this is accurate?” I asked.

“Sure… that looks about right. What about it?” Charlie asked.

“Well, if we used yearly data, the average daily high temperature is 54.8 degrees, right?

“That’s about right,” replied Charlie.

“And if I used the 3-sigma control limits calculated from this yearly data to make my decisions, I should be comfortable every day of the year as long as I dress for weather that’s between about 40 degrees and 69 degrees… is that right?” I asked.

“Well, some of the time, sure… but you’re gonna be awful cold in the winter, and you’re gonna be awful sweaty for most of the summer if you do that!” Charlie chuckled.

“Why is that?” I asked.

“Well, you can’t just use yearly temperature averages! You’ve got to…” Charlie’s voice trailed off as he realized my point.

“You’ve got to… what?” I pressed.

“You’ve got to only take the current month into consideration,” he finished, his voice dropping low.

“Exactly. When you’re preparing for this week’s temperatures, it’s more useful to know what happened last week and the week before than knowing what happened six months ago, right?” I prodded.

“Yeah, that’s right.”

I continued. “And so you see why taking into account data that happened on a production run four years ago isn’t exactly as useful as data that happened more recently?”

I could see Charlie nodding, and I switched by screen to a focus on a single month.

“So let’s say I was visiting Green Bay in January. Would I be better prepared than the previous example if I instead dressed for a day that’s between 14 degrees and 40 degrees?”

Charlie kept nodding. “Yep. Gotcha.”

“Now,” I said, moving forward. “Fortunately, QI Macros has rolling templates where you can continue to have all that data, but it only takes the most recent X days into account when determining your average and control limits. You can set that to track 30 days, 80 days… whatever makes sense for you. It’ll update automatically when you add new data.”

“Really?” Charlie replied. “That sounds great.”

“You bet. I’ll set it up for you and show you how to continue to use it. It’ll be as easy as…”

“…as easy as beating the Bears at Lambeau Field?” he asked.

I laughed. “Well, considering the current streak between those two teams, I’m not sure it will be quite that easy… but it’ll be close.”

“That’s good enough for me,” Charlie said. “Thanks a lot!”

If you’re interested in learning more about QI Macros Rolling chart templates, click HERE and HERE to learn more.