Is your Rework Tank a storage problem or a profit center?

Introduction

Every scheduler watches the rework tank. You already know that slop represents real margin, and you're actively looking for places to blend it back in. That part isn't the problem.

The problem is that when you're juggling fifty tanks and a week of changeovers in a spreadsheet, you can't evaluate the financial impact of every gallon across the whole horizon. So the sheet pushes you toward a simple rule: if the tank is filling up, put the rework in the next batch that allows it. That keeps the plant running, and it quietly leaks margin, because the difference between a good schedule and an optimal one was never whether you use the rework. It's when you use it, and in what.

"First feasible" versus "most profitable"

Here's the difference in practice. Generic Product A runs tomorrow and its formulation allows a share of rework. The tank is filling, so the rule says use it, clear the tank, move on. And it works. But Product A is built on cheap base stocks, so every gallon of slop you blend in is displacing cheap material. You've solved a storage problem and captured almost nothing.

Now look three days out. Premium Product B is scheduled for Friday, and it's built on expensive synthetic raw materials. The same gallon of rework blended there displaces something far more costly, so it's worth substantially more. The optimization sees both options at once, weighs holding the inventory for three days against what it saves when it lands in the premium formulation, and holds, if the math says holding wins. That's the whole shift: the rework stops being something you dispose of at the first opportunity and becomes something you place where it's worth the most.

The long tail you never scan

There's a second thing a spreadsheet quietly forces on you: a short list. Scanning your whole formulation library by hand isn't practical, so in practice the rework gets matched against the handful of products everyone knows allows it, the usual five or six. That's not a deliberate decision, it's just what's manageable.

But your technical database often holds dozens of low-volume SKUs that are also valid candidates, products nobody thinks of as rework destinations because they're obscure or they run rarely. The optimization doesn't have to maintain a mental shortlist. It evaluates the whole library, so it can find the odd low-volume batch that happens to be a good home for a few thousand gallons, relieving tank pressure without touching the schedule of your primary products. Those opportunities were always there. They just weren't findable by hand.

Storage isn't a rule, it's a price

Most plants manage the tank with a static rule: hold until it's 90% full, then take whatever option is available. It's simple and it's blunt, because it treats storage capacity as a switch, either you have room or you don't.

An optimization treats it as an economic constraint instead, and this is where the behavior gets genuinely useful. When the tank has plenty of room, holding costs you very little, so the plan is selective, it waits for the high-value formulation. As the tank fills, that free room disappears, the cost of holding another gallon rises, and the plan naturally becomes less picky, accepting a lower-value destination rather than risking the sequence. Nobody wrote that rule. It falls out of the math, because the tighter the constraint binds, the more holding actually costs. The result is a sliding scale rather than a cliff: every gallon goes to the best home still available at that level of pressure.

Frequently Asked Questions