Improvement Insights Blog

Do you use data to discover problems and innovative solutions? Or do you use data to confirm what you already believe and think you know? This is called confirmation bias.

“Hi, I’m Jay Arthur, author of “Lean Six Sigma Demystified” and QI Macros [software].

“So in Quality Improvement we’re asked to pledge allegiance to science and evidence. A lot of people do that, but very often they get trapped because they like data that supports what they believe about how things work, and they don’t like data that contradicts what they believe. Does this make sense?

“So are you one of those people who selectively goes out and says, “Hey, this data supports me. That data doesn’t support me, so I’m not going to talk about it”? Sometimes the best answers are in the data that you don’t like. The best answers come from the data you don’t like, right? So rather than being stuck, let’s go a different direction.

“I worked on this one project when I was in the phone company, and all the managers thought, “We have these four day lead times to fix phone service,” and I think, “Okay, well let’s fix that.” What they wanted to prove was that they needed more repair people so they spun up a whole test of this thing: an eight-week adventure in Salt Lake City. I flew out there every week to go out and work on this project. I was there about four hours the first morning when I finally understood what they were trying to do and I said, “Oh, they’re trying to use their data to prove the answer they want, not the answer that would benefit all of us.”

“The correct answer was most of these calls were for repair… well, guess what? We need less repair! We have data, we could go figure out why service goes out in Arizona (because it’s hot) and Seattle (because it’s wet). We can fix those things, but they wanted nothing to do with that.

“So after about eight weeks they figured out that no matter how many technicians they threw at it, they could not collapse that four day service window. You know, I think we got a bonus for working on that thing, but it didn’t work, right? And then I couldn’t get anybody interested in “let’s figure out how to reduce the amount of repair we need.”

“But anyway, that’s it. You have to use the data, even if you don’t like it, because that will solve the problem.

“Let’s create a hassle-free America, hassle-free healthcare. You may not like the data you see, but take that data and use it. Let’s go out and improve something this week.”