Issue link: https://iconnect007.uberflip.com/i/1259453
24 PCB007 MAGAZINE I JUNE 2020 controlling the equipment to minimally inte- grating automated tracking of defects, etc. You look at it now, and the real transition we're going through over the last 10 years and prob- ably the next 10–15 years is starting to auto- mate more of this. We're gathering a lot more data, and you can have lots of data and not do anything with it. If you have too much data, it's too much for a human to keep up with. Now, we're start- ing to get a lot of better-automated tools that run both within a process and within a piece of equipment. Two processes help present the data, analyze the data, and, in some cases, au- tomatically correct based on the data. We've seen a lot more of that in our industry. Johnson: What are some of the example tools? Korf: The number one tool that's used is Excel. People gathering data or a piece of equipment will come with embedded software or part of their user interface that will collect the data in a spreadsheet form. Then, you pull that da- ta out and analyze it. Almost all the software tools, in effect, take that data and present it in various ways. They put it on a screen in front of you, show you whether your process is in control or not, and have the system react to that control when it goes out of control and adjusts the process automatically like dosing systems in plating lines. Johnson: It doesn't necessarily mean that you must have an enterprise management system or sophisticated quality management soft- ware. A two-dimensional tabular database program like Excel can go a long way. Korf: I've been through lots of those specialized systems. They run for a while, then the new type of sys- tem comes in the next year, and they switch to another system. The next year, an- other system comes in, so you switch to another system, and they don't tend to gather traction. People lose interest. They start gath- ering data and putting it on charts to present at a meeting, and it dies. The intentions are all well. In most companies, the quality depart- ment will manage the process and push people to follow the process and react to the data that you're getting. But fundamentally, you have to take people off the line to react to the data or have them focus on improving versus pushing product out the door. That has always gener- ated conflict. How do you get enough people assigned to react to the data? Along with that, you tend to shotgun the da- ta. You get data from every process, even ones that don't need to be improved as much. You get data overload instead of having the teams focus on one particular area until that gets bet- ter. A good example from many years ago was we had companies that merged together, and our front-end processes were all competing against each other to have the lowest defect rate, which was great. It worked out well, and we all had great pride in our processes. Every month, we got together, compared the metrics, looked at the defects, how we reacted to it, and we all learned from each other. But then we got down to a point talking about scrap produced. We were down to hav- ing people chasing on the floor, trying to get $5 of scrap off of our list because another site was going to have $3. We were a one-billion-dollar company at that time. We all looked at each other and said, "This is nonsense. We're wast- ing money trying to save, in effect, nothing." We drew a baseline and said, "If we're below this defect level, we'll call that zero and focus on something else that's bigger." That was a huge transition point of getting us to im- prove our processes so we could focus on the big- ger problems versus the chart data, per se. Johnson: The moral of the story is don't take "total" too literally in TQM.