In the Kepner-Tregoe Problem Analysis process, we start with symptoms and then work our way to causes. The first steps involve specifying the details of the issue, in terms of What object is having a problem and What problem it is having; in terms of When that problem shows up; Where it shows up; and the Extent to which it shows up. We also contrast these Is with comparable Is Nots—when might we expect the problem to show up, but it doesn’t, どこで might we expect it to manifest itself, but it doesn’t, and the like.
One area where many folks have trouble involves the third When question: When in the process or life cycle did the deviation occur? This question can be tricky. To begin with, it’s not one question but two: When in the process . . .? When in the life cycle . . .? On the process side, we are looking for events or stages of process flow, such as, after visual inspection, prior to integration, or during final release testing. Whenever the problem appeared, its cause must be either right at that point in the process or shortly before it, not after.
On the life cycle side, when in the life cycle . . .? often leads to responses like, the first batch we ever made, the last batch before we went on vacation, or the first batch after shift change. These often get to the effects of lead operator process adjustments from shift to shift, and they can also point to differences in the first batch or last batch compared to others. In the first batch we may have machinery that is not “warmed up;” in the last batch we may have employees who are thinking ahead to the holidays and paying attention less than usual. But it can be even more sophisticated that this.
For example, I recently ran into a case of craters in the painted finish of some metallic panels. Each panel, after painting, was heat-dried, and after drying, tiny, one-millimeter craters appeared, depressions in the paint that went all the way down to the primer coat. Not many, just a few, but enough to consign the panels to the re-work bin. We happened to ask, “What is in these craters? Do you see paint by-products, primer by-products, or anything else?” As it turns out, they had seen something, and what they had seen was fluorine.
Hmmm . . . so we began to ask, When in the process is fluorine used? But the answer was, never. There is no fluorine, fluoride, or any other composite of that element anywhere in this paint’s formula. That was surprising. What had looked like a suggestive lead led nowhere. But the testing was accurate; if there was fluorine in there, it had to have come from someplace. Initially, we were stumped.
Then we asked the question more broadly, as in, Where is Flourine used in the building? As it turns out, it was used, occasionally, in another process. In fact, it was used on the line right next door to our crater-producing line, on the other side of the wall just to the east.
When we re-asked, When in the life cycle, it turned out that the bad, cratered batches were all manufactured on the same day that a fluorine lot was being produced next door, and that none of the good batches, produced without craters, had been made at the same time as fluorine batches; they had been made when nothing was running next door.
But it’s next door, someone said, through a wall that’s three feet thick. We sent someone to take a look, and yes, there was a three-foot-thick wall between the two lines. But there was also a three-foot-diameter hole in the wall, and a fan blowing from the fluorine side of the wall right across the top of the tank that had produced the paint with craters. This last fact was critical; it’s not enough that the two kinds of batches co-occurred in time. That’s only a coincidence. What the fan gave us was a mechanism of the cause, a way for the fluorine batches to affect the other batches.
Cast your net broadly, when looking at time, look carefully at any cycles that co-occur, and drill deeply enough to notice any key differences. When in the life cycle can be a tricky question to use, but it can be a powerful one as well.