Shrink the Problem

Do not boil the plant. A problem the size of the site becomes a meeting, and a meeting about the whole site ends in a slogan.

Here is what the big version sounds like: “Quality is down.” Everybody agrees. Nobody can test it. You cannot run a trial on quality. You cannot stand next to it and watch.

Now shrink it. Say the scrap tags are piling up at packing. Sort them by part, and most come from one part. Sort by step, and they are all from trimming. Sort by shift, and second shift has most of them. Look at what is different when they happen, and it is the first hour after a material change.

Before: “Quality is down.” After: “Trimming on second shift makes most of our scrap in the first hour after a material change.”

The second sentence is something you can test before Friday. You can stand at trimming on second shift for the next material change. You can watch that first hour. You can try one change, like a first-piece check right after the swap, and count the tags the next night.

That is the mechanism. A smaller problem has fewer causes hiding inside it, so the one that matters can be seen. It also has an owner, because quality belongs to nobody and trimming on second shift belongs to somebody. And it fails cheap. If your test does not move the tags, you lost one shift, not one quarter.

Shrinking is not ignoring the rest of the plant. It is picking the spot where the problem is loudest and proving you can quiet it there. What you learn usually travels.

If you cannot point at the step, you do not have a problem yet. You have a mood.

This week, take one complaint everybody agrees on. Shrink it to the step, the shift, and the condition. Test that. Leave the rest of the plant alone until you do.

Go deeper: the free White Belt course and The 5 Whys Only Work If You Answer Honestly.