The Only Number on the Wall

The daily standup ran in the aisle. Thirty people in hairnets and white coats in a loose line, clipboards and tablets in hand, facing a run of whiteboards on casters that stretched most of the width of the packaging hall. Safety on the far left. A barrier and preventive action log split into short term and long term, half of it filled in by hand in red marker. A performance board with the headline number and the OEE under it. Scheduling notes. Then line after line of charts, one pair of panels per packaging line, with a column of handwritten entries from the night shift. Nobody read from a laptop. By the standards of most plants I walk, it was an excellent wall.

The plant was in crisis anyway. It packaged product for pharmaceutical customers, which meant blister lines, cartoners, serialization, and a customer base that did not care how good the meeting was if the order shipped late. Service had slipped far enough that customers were calling above the plant, and I had been asked to help get it back into good standing. I had the usual three weeks to come back with an answer. I had the answer in about an hour. The three weeks were for something else.

The Number at the Top of the Board

The performance board had one headline number, and it was schedule attainment. Did the lines run what the schedule said they would run. It is a popular number in contract packaging because it sounds like a promise to the customer, and because the planning system produces it every morning without anyone lifting a finger. It is also, in a plant like this one, close to meaningless, and the reason it is meaningless is instructive.

Schedule attainment measures the plant against a plan. The plan is written by a scheduler who knows what the lines can do. When a line struggles, the scheduler does what a good scheduler does and reschedules around it, and the plant hits its schedule. When the customer's order slides a week because the plan slid a week, attainment stays green. The number does not lie, exactly. It measures the plant against a target the plant has already adjusted to itself. An operator on the cartoner cannot move it on Friday, because the thing that moves it lives in a planning office and in the customer's forecast. The floor had been handed a number it could watch but not touch.

Worse, in a plant whose entire business is short runs across many SKUs, schedule attainment quietly rewards the wrong behavior. The easiest way to hit the schedule is to run long and change over rarely, which is exactly what a packaging customer with 40 SKUs does not want. The plant had posted, at the top of its board, a number that told the floor to do the opposite of what the customers were paying for.

What They Took Out of OEE

The number under it was OEE, and here the story got more interesting. Someone had done real work on the OEE. Availability, performance and quality losses were tracked by line, coded by reason, and charted on the wall. It was a good system with a hole in it. Changeover time had been removed from the calculation. The team had decided, at some point before the crisis, that changeover was not controllable by the operators, and so it should not count against them. Line clearance, which in a pharmaceutical plant is the regulated procedure of stripping every trace of the last product and its labels off the line before the next one starts, went with it.

I want to be fair to the people who made that decision, because it was a reasonable one at the time. Changeover in a regulated packaging environment is partly fixed. Quality has to sign off. Documents have to be reconciled. There is a floor below which you cannot go. The team looked at that fixed component and concluded the whole thing was outside their control, and once it was outside their control it came out of the number. Nobody hid anything. They defined the biggest lever in the building out of the data and then spent a year wondering why the data did not move.

Nobody hid anything. They defined the biggest lever in the building out of the data, then spent a year wondering why the data did not move.

Paul W. Swaney III

When I put changeover and line clearance back in and reran the OEE by line, the picture turned over. The lines were not slow when they ran. They were stopped, for hours a shift, between jobs, inside a procedure that had been declared uncontrollable and therefore had never been studied. That was the number. It was obvious in the first hour because it was the one thing on the wall nobody was measuring.

Three Weeks to Earn the Board

I could have put changeover on the wall that afternoon. I had the authority to do it. I did not, and the reason is the one piece of this story that generalizes to every plant I have walked since. If I pick the number and the plant does not believe in it, it is my number, and it lasts exactly as long as I am standing in the aisle. The assessment was never about finding the answer. It was about getting 30 people in hairnets to find it themselves, with their own data, so that when the board changed it was their board.

So we did the three weeks properly. We timed changeovers by line and by shift, and broke each one into the part quality owned and the part the crew owned. The regulated piece turned out to be a minority of the time. The rest was walking, waiting and searching: for the next job's components, for a quality signature, for the right change parts. The night shift on one line had worked out its own sequence and was clearing and restarting in roughly half the time of the day shift on the same equipment, which nobody knew because nobody was counting. The pushback came from the people who had built the original OEE definition, and it was honest pushback. By the end of the second week they were the ones presenting the numbers.

In week three the top of the board changed. Schedule attainment came down. Changeover and line clearance time went up, by line, by shift, with the crew's name beside it. Nothing else on the wall moved. Safety stayed on the left, the action log stayed where it was, the line charts stayed in their frames. The standup still started at the same time. One number.

Swaney Group Capital proudly uses Attio as part of our integrated sourcing, transaction, and operations tech stack

The Friday Test

I have a question I ask now, every time I am looking at a board and trying to decide what belongs at the top of the board. If this number is bad tomorrow morning, who in this building could have made it better today, and what would they have done differently? If the honest answer is a planner, a buyer, a customer or the finance team, the number is a report. It can live on the wall somewhere. It does not go at the top. The top of the board is for a number the people standing in front of it can move by Friday with their own hands, and for which one of them owns the answer.

Quality does not go at the top either, and this matters in a pharmaceutical plant more than anywhere. You cannot run a packaging line for a drug company without quality on the wall. It was on the wall. It stayed on the wall. The difference is that quality and safety are gates. They are the conditions for being allowed to run at all. A gate does not need to be the headline number, because nobody at the standup is debating whether to pass through it. Putting a gate at the top of the board makes it look like a dial, and operators do not need to be told to turn quality up. They need to be told which of the 20 things they could do today is the one that matters.

The failure runs the other way too. Years later, at a paper products company where I was a consultant with no control, the right number was already on the wall. On-time in full had fallen to an all-time low of 87% and everybody could see it. It did not move for months, because nobody standing in front of it owned it, and the people who did own it had declined to come to the meeting. When the leadership shuffled and someone finally owned the number, OEE moved from the mid-teens to the mid-30s. The number had been right the whole time. The owner was what was missing. The board needs both.

What Goes at the Top of the Board

If you are an owner, a buyer in the first month, or an advisor walking a plant, the test is cheap and the fix is cheap and the decision is still hard, because the number already at the top of the board is there for a reason and someone built it. Start with what has been removed. Every plant that tracks OEE has excluded something as uncontrollable, and the excluded category is where I look first, because it is the only loss nobody has studied. Then run the Friday test on whatever is at the top of the board. If the floor cannot move it, take it down, and expect pushback from whoever put it up. Then pick one number, by line and by shift, with a crew's name on it, and give the plant the weeks it needs to arrive at that number on its own, with its own data, so the board belongs to them. Leave the rest of the wall exactly as it is. Safety stays left. Quality stays a gate. The action log stays full.

The daily standup at every business I have been part of since opens with that one number. The plant got back into good standing with its customers. The wall looked almost identical to the day I walked in. One number was different.

Twenty numbers is a report. One number is a decision.

I am the founder of Swaney Group Capital, a fundless sponsor focused in the lower middle market. LeverUp® is published weekly. More if I have something else to say.