The Labor Savings Nobody Scheduled Time to Claim
A plant leader at a Midwest protein processor had been sitting on the same opportunity for four years.
The savings had a number and an age
A plant leader at a Midwest protein processor had been sitting on the same opportunity for four years. Eight to twelve heads at one of the two plants, no capital required, low-hanging fruit that any line-tracking review would surface inside a week. Call it $300K a year. The number was not in dispute. What was in dispute was whether pulling it out was worth the trouble.
His read, roughly translated: if you tell me there is $300K sitting here, I think you are full of it, because if that were true I would have to admit I have been sitting on $300K for four years. So the data pulls stalled. The system access requests waited. The opportunity kept its number and kept aging, and everyone involved treated that as normal.
It is not normal. It is the most common way real money stays on a floor.
Found is not claimed
Labor is the easiest opportunity in a plant to see and the easiest to price. The question "what is the ideal headcount for this line at this volume" has a clean answer, and the answer converts straight into a gain-share number. Throughput improvement is harder to monetize and harder to defend. Headcount is not. That is exactly why a headcount opportunity that sits for years is so strange on its face: the value is legible, the math is simple, and still it does not move.
The reason it does not move has nothing to do with finding it. It is the capture cost. Claiming an 8 to 12 head reduction is not a single decision; it is a chain of small unglamorous tasks. Someone has to get login access to the OEE tracking system for both plants. Someone has to determine whether the bad numbers at the second plant are a configuration problem, a late-start habit, or something deeper. Someone has to own the seam between the line data and the schedule. Each task is a couple of hours. None of them is on anyone's plan. And they all land on a plant leader whose day was already full before the opportunity showed up.
So the opportunity gets silently reclassified. On paper it is $300K in savings. On the floor it is a stack of data requests and config fixes with no owner and no budgeted time, which reads as burden. Burden loses to the day job every time. The mole here is not a missing insight. It is an unowned capture cost.
The automation version is the same shape from the other direction. At a lunchmeat plant, the high-volume line was heavily automated and carried the bulk of the run; the low-volume lines stayed manual because their volume was too infrequent and unpredictable to justify the capital. Perfectly rational. But it means the labor that is hardest to reach with capital is exactly the labor that sits, year after year, because no one wants to spend the effort to claim a savings that capital cannot claim for them. The value that is hardest to reach is the value that stays.
Put the capture cost on the board
The move that finally shifted the four-year opportunity was not a better savings estimate. It was making the capture cost visible next to it.
The fix was two columns on every project plan: hours expected from the client's team, and hours expected from the outside team. Milestone level, not task level, so nobody drowns in detail. Recurring work called out separately from one-time work. And, critically, plant-leader time broken out from corporate time, because fifteen minutes of a plant leader's day is not the same asset as fifteen minutes of a corporate analyst's. Written out, the ask stops being nebulous. Initial data pull, two hours. Recurring, thirty minutes a week for a stretch of weeks. First-pass analysis, one day. Now the $300K has a price in hours, and about 20 hours of internal time against $300K a year is not a burden. It is the best trade on the floor.
You can run this on your own plant this week without hiring anyone.
- List every known headcount or throughput opportunity you can name, each with its dollar figure. If you cannot write the dollar figure, it is not yet an opportunity, it is a hunch.
- Next to each, write the hours required to validate it and the name of the person who owns those hours.
- Wherever the owner is a plant leader whose time was never budgeted for it, stop. That is your constraint. Not the opportunity, the ownership of the capture work.
The opportunities that have aged the longest will be the ones whose capture hours were never assigned to anyone. That is not a coincidence; it is the whole pattern.
What a claimed-savings floor looks like
Every named savings line carries its capture hours right beside the dollar figure, split by whose time it takes. A known opportunity moves from identified to validated inside one quarter, not four years. The seam between the line-tracking data and the schedule has one owner, and that person's capture tasks sit on a calendar with dates on them. Nothing labeled "opportunity" sits older than a quarter without a validation task assigned. When a plant leader pushes back, the pushback is about the hours, which are written down and arguable, not about the number, which is settled.
The closing
The number was never the hard part. The plant leader was not wrong that claiming it felt like all his work; the savings were real, and so was the cost to pull them out. Nobody had put the two in the same column, so the cheap trade looked like an expensive chore for four years running.