Run One Experiment This Week

Every floor I’ve ever walked has a graveyard. It’s not out back. It’s in the break room, on the bulletin board, in the binder nobody opens. It’s where good ideas go when we launch them like a fire drill and then never check if they did anything.

Here’s how it usually goes. Somebody has a good idea. A new handoff sheet, a new way to stage parts, a new start-up check. We roll it out to everybody on Monday. By Thursday half the crew is doing it, a quarter is doing it wrong, and the rest never heard about it. Two weeks later somebody asks, “Did that handoff sheet help?” and nobody knows. So it dies. Not because it was bad. Because we never gave it a fair test.

I want you to stop launching ideas and start running experiments.

That sounds like a lab coat word. It isn’t. An experiment is just four things: a number before, one change, a number after, and a line that tells you when to quit. That’s it. You don’t need a project charter or a steering meeting. You need a pen, a card and a week.

Start with something you can count. Not “communication is bad.” Count the repeat questions at handoff. Not “the line runs slow.” Count the minutes you wait for parts. If you can’t put a tally mark next to it, you’ll argue about it forever.

Count it before you touch it. Two shifts. This is the part everybody wants to skip, and it’s the part that keeps you honest. Without a before number, every change looks like it worked, because you want it to.

Change one thing. One. If you change the sheet, the meeting time and the staging area all at once and things get better, congratulations, you have no idea why. Next time you’ll do all three again and it won’t work, and you won’t know which piece you lost.

Write your stop line first. This is the one that separates a supervisor who experiments from a supervisor who gambles. Before you start, decide what makes you quit. If it gets worse two days running, you go back. If anybody gets hurt or blamed, you go back that minute. Write it down before you start, because on day three your pride will want to keep going.

Then count every day for a week.

Here’s the part I love. At the end of that week you win either way. Either you’ve got a better way to run your area, with a number to prove it, or you’ve got proof that idea wasn’t it, and you found out in seven days instead of seven months. Failing fast and small is cheap. Failing big and slow is how the graveyard fills up.

And watch what happens to your crew. The first time you stand at the board and say, “We had 14 questions a shift last week, we’re at 6 now, here’s what changed,” people lean in. They stop seeing change as something done to them and start seeing it as something they can test. The next idea comes from them. That’s the real win. You’re not just fixing a handoff. You’re teaching a floor how to think.

So pick one thing this week. Count it for two shifts. Change one thing. Write your stop line. Count again for seven days.

Then tell me what you found. I want to hear it.

The One-Week Floor Experiment card (five boxes and a stop line) is coming to the Tools page. Until then, run the same five boxes on a blank sheet. And get a new way to run your floor every week in the BBG Briefs.

Marc Bradburry