SKU Math the P&L Never Runs

The changeover the dashboard couldn't see

A multi-line sausage co-packer runs roughly 35,000 pounds per shift across four packaging lines. Its downtime tracking system reports average changeover times between 6 and 13 minutes, varying by line. The floor reports something different. Operators routinely recode portions of changeovers as lunch or other categories. The MES retriggers a new downtime code every time the line stops and starts, which means a single multi-step changeover splinters into three or four entries that never add back to a real number.

The reported figures are physically implausible. A real washdown changeover at this facility runs 30 to 40 minutes. Operators run two to three changeovers per shift, sometimes four to six on a single line. The plant has been carrying this gap for years.

Nobody flagged it because no individual SKU added enough complexity to break the schedule on its own. Each addition cost a few minutes, hid those minutes behind fragmented entries in the tracking system, and absorbed the loss into generic underutilization rather than changeover.

SKU proliferation is not a scheduling problem

Most ops leaders treat SKU growth as a sequencing challenge. Better order, fewer changeovers, smarter Friday plans. That framing assumes the cost of complexity is visible. It usually isn't.

The cost of complexity is the part of capacity loss that nobody attributes to any specific SKU. Allergen sequencing alone forces GAP and organic items to run first and cheese-bearing items to run last because the line has no intra-day sanitation. That single rule shapes the entire week's schedule before anyone touches a sequencer. Add a 0.91-inch link variant next to a 0.95-inch link and the schedule has to absorb a knife move and a diameter check that nobody priced into the SKU.

A different co-packer just had a major QSR chain approve a transition to a pre-marinated steak format across both production facilities. First market trial: April 20. Expected line run-rate reduction: 12 to 15 percent through 2027. Capital spend for the transition is approved. Line layouts and equipment specifications are not. The 12 to 15 percent hit is real and dated. The layout that pays it back is not.

This is the same mechanism as the sausage line. A new SKU type prints a tax on every shift before the operations side has been given the equipment to neutralize it. The P&L treats the tax as residual variance. The floor treats it as overtime, scrap, and a 1.78 percent leaker rate that sits just above the threshold where anybody is willing to fight for the root cause.

What to actually do this quarter

Three steps that pay back fast.

First, audit every line for changeover events that retrigger the downtime code. The fix is mechanical. Recode the changeover itself as a "product" so all stop and start segments roll up into a single bulk run. The sausage co-packer above ran exactly this play. The result was not faster changeovers. The result was a number you could actually compare to the standard, which surfaced a 15 to 25 percent capacity gap that the dashboard had been smoothing over for a decade.

Second, before any new SKU lands on the schedule, force operations to attribute the changeover cost. Not the average changeover. The actual added minutes for this specific transition. A Tier-1 packaging optimization project recently narrowed Phase 1 to 14 high-impact items. Three required spec changes (die cut to RSC, ECT reduction). The savings model showed an 8 percent box-cost reduction. The Georgia plant blocked the three items on headcount and durability grounds. The 11 items that ran cleanly delivered savings. The 3 that did not exposed a labor-allocation problem nobody had budgeted for. Both sides were right. They just weren't in the same model.

Third, separate capacity reduction from new SKUs as its own line in schedule-adherence variance. Right now most plants bury it inside generic downtime or operator underperformance. It belongs in its own column with a named owner. The QSR-steak co-packer's right move is to model the 12 to 15 percent reduction directly into the next two quarters of throughput targets and let demand planning see the gap before they discover it. Capacity that you have already lost on paper is not a capacity problem. It is a planning problem.

The savings are real. So are the losses.

The savings model for SKU additions is real on paper. So is the operational cost on the floor. Nobody puts the two in the same model. The capex committee approves a transition based on customer pull. Operations absorbs the run-rate hit. The downtime tracking system fragments the evidence into entries that don't reconcile. By the time the P&L narrative is written, the SKU has paid itself in schedule minutes that nobody assigned to it.

The plant that sees this first writes a different schedule. The plant that doesn't writes a capex request for a new line.