How to Find Hidden Constraints in Your Plant

Every operations leader has read The Goal. Or says they have.

Most of us remember the method and forget the hike.

The method is solid. Goldratt taught us to find the constraint, get everything we can out of it, line the rest of the operation up behind it, and only then spend money to expand it. He also gave us the line every plant manager should keep on their desk: an hour lost at the bottleneck is an hour lost for the whole system.

The hike is where he put the bigger lesson. The most famous bottleneck in the book isn’t a machine. It’s Herbie, the slowest kid on a Boy Scout hike. Alex Rogo doesn’t yell at Herbie or buy him better boots. He puts Herbie at the front of the line and takes some of the weight out of his pack. The whole troop speeds up.

The constraint was a person. The fix was leadership.

Start with the obvious

If you’re looking for a constraint, start where everyone starts. Walk the floor and look for piles. Work-in-progress stacks up in front of a constraint and starves the steps after it. Ask which machine everyone waits on, which one gets expedited, and which one the schedule is built around. Check where the overtime goes.

That finds the constraints you can see. It’s worth doing, and sometimes the answer really is a machine.

But the constraints that cost the most have something odd in common. Somebody already knows about them.

In every story below, someone on the floor, in the office or in the room knew exactly what was wrong. The constraint wasn’t hidden from everyone. It was hidden from the people who could act. What sits between those two groups is people and leadership: whether someone feels safe enough to say it, whether anyone owns the gap, and whether a leader is willing to hear it and decide.

So the real question is where the knowledge gets stuck. Here are four places to look, from the floor up.

1. On the floor: what operators know but don’t say

What it looks like: a problem gets handled every shift, quietly, and never reported. From the office the area looks fine. Clean, even.

At a plant I worked at, an operator wiped up an oil leak on his machine every shift. Maintenance had said leaks like that were normal, which was fair for minor ones. It became tribal knowledge. When a new operator started, he was told what everyone had been told: clean it up and keep going.

Then the transfer bed seized. The machine was down for two weeks waiting on parts and a vendor technician. The repairs cost about $5,000, and that was the small number. The plant already ran 24/7, so there was no overtime to make up the gap. The other machines carried the load and wore faster, and late jobs climbed. My estimate of the real cost was around $25,000.

The plant passed its 5S audits every time. The audit measured clean. It couldn’t measure silence.

Amy Edmondson found the same thing in hospitals. She expected the best-led nursing teams to report fewer medication errors. They reported more. They weren’t making more mistakes. They felt safe enough to say so. The teams that looked best on paper were often the ones that had learned to keep quiet.

Nobody in that plant was hiding anything. The operator did exactly what he’d been taught. From the office, a silence people have learned looks exactly like a plant that’s running well.

How to find it this week:

• Ask operators about the work, not about problems. “What would surprise someone new on this machine?” gets you further than “Any issues?” People describe workarounds as just part of the job. The workarounds are the clue.
• Ask when someone last raised a problem, and what happened next. If nobody remembers, or the answer is “nothing,” you’ve found a constraint no audit will show you.

2. Between departments: the work nobody owns

What it looks like: a process takes weeks, and every department says its part only takes a day or two. They’re all telling the truth.

I worked on a process like that. Three months from start to finish.

We didn’t run a textbook value stream map. We ran a simplified one, built with the people who did the work instead of for them. We put the owner of each step in one room, gave each department its own colour of sticky note, and had everyone map their own steps. Then we did what a value stream map is really for. We looked at the lines between the notes.

Every line was a handoff: something sitting in an inbox, waiting for a signature, or waiting for an answer to a question nobody knew to ask. Add up the time inside the boxes and you got days. Add up the time in the arrows and you got the other eleven weeks.

The work was in the boxes. The time was in the arrows.

We got it down to 10 days, with an emergency path that could do it in 3. Nobody worked faster. We ran steps side by side that never needed to wait for each other, and we gave the whole process one owner whose job was to chase anything unfinished.

Goldratt warned that when every department improves its own piece, the whole can get worse. Each department was measured on its own box. Nobody was measured on the arrows, so nobody saw them. Your org chart has the same flaw. It shows the boxes and leaves out the arrows.

The biggest shift came when people saw, for the first time, how their piece touched everyone else’s. People don’t change how they hand off work until they see who’s on the other end of it.

How to find it this week:

• Follow one order from release to shipping. Time the waits, not the work. Wherever it sits longest, ask who it’s waiting for, not what.
• Ask who owns the whole process. If the answer is a committee, or everyone, nobody does.

3. In a person: the manager everyone likes

What it looks like: a busy, likeable team with lots of ideas and very little that changes.

I had an engineering manager who was bright, ambitious, full of ideas and genuinely good with people. His team was always busy. They weren’t effective.

We’d bring the team together, collect ideas and build real energy in the room. Then everything went to him, and he would spin. He’d investigate each approach, debate which one would have the biggest impact, start one, get pulled to another, and never get deep enough to reach the root cause.

John Maxwell calls this the Law of the Lid: a leader’s ability sets the ceiling on how effective their people can be. That team had talent and energy. It also had a lid.

His people loved him, and that was part of the problem. Kim Scott calls it ruinous empathy: caring about someone so much that you stop challenging them. The people around him had let him spin for years because they liked him.

I didn’t want to take over, because that would have taken the job away from him. So I made it smaller. Every time, one question: what’s the next step you can take? Then I held him to that step, and only that step. When the team came up with ideas, I’d ask him questions about them, then ask whether any could be done before the end of the shift.

It built momentum. His people, who had been frustrated by how little they were achieving, started to feel things moving. More ideas came out of the wins, and some of the simplest ideas turned out to be the most effective. Small improvements compound. A perfect plan that never starts doesn’t.

I’ll be honest about the ending. He carried an old resentment toward the company over something that had happened ten years earlier, and coaching couldn’t reach it. I eventually let him go.

Alex Rogo got to keep Herbie. I didn’t get that ending. Coaching can lighten someone’s pack. It can’t make them want to walk.

How to find it this week:

• Count last month’s improvement ideas and how many got done. Find the desk where they stop.
• Then check whether it’s yours.

4. At the top: the decision nobody makes

What it looks like: everyone on the floor knows what the problem is. It’s been raised. There may even be a business case. Nothing happens.

At a shop I worked at, we ran two old lasers. They struggled to cut some of our material, broke down often, and sent rework downstream. Everyone on the floor knew they were done. The business case was there.

The owner wouldn’t approve it.

To be fair to him, it’s different when it’s your own money. To the person who builds it, a business case is a spreadsheet. To an owner, it’s a cheque with his name on it.

But something else was going on, and it isn’t unique to owners. When the same request keeps coming up from the floor, it starts to sound like complaining. From the corner office, it’s easy to decide that people always want something new, and that you know better. So you stop hearing it.

That constraint really was a machine. The bottleneck to fixing it was a person. It took a change of ownership, with new owners willing to invest in the business, before those lasers were replaced. That was about three years after I arrived.

Three years of breakdowns and rework, and everyone knew.

How to find it this week:

• List every request that has come to you more than once and is still waiting for a decision.
• For each one, ask yourself honestly whether you decided “no,” or just filed it under complaining.

Before you buy anything else

Sometimes the answer is a new machine. More often it’s someone who feels safe enough to speak, someone who owns the gap between departments, a manager who needs the first step made smaller, or a leader willing to hear the same request one more time.

The method finds the constraint. Leadership is what lets you see it, and what lets you fix it.

Every plant has a Herbie. The only question is whether you’ve found yours, or whether it’s you.

---