A metals service center is one of the strangest businesses in the industrial economy. Your input is a commodity, your output is a commodity plus work, and the market prices both without asking you. Hot-rolled coil went from under $500 a ton to over $1,900 and back inside three years. Your costs, meanwhile, are boringly fixed: the building, the slitter, the cranes, the people.
That mismatch is the whole game. A service center earns money three ways, and the operators who survive price cycles are the ones who manage all three as separate businesses.
The three revenue streams, and what each one really is
Toll processing fees are the charge for working someone else's metal. This is the only revenue you fully control, and it should be doing more work than most operators let it. A processing fee is not a courtesy add-on; it is the insurance premium against commodity risk. When the metal market is strong, competitive pressure pushes processing spreads down, and operators give away the insurance exactly when they will need it next.
Material margin is the book: buy coil at one price, sell slit, cut, or blanked product at another. This is a portfolio with real exposure. The spread looks like operating margin in a rising market and reveals itself as inventory risk in a falling one. A warehouse full of coil bought at the top is a position, whether or not anyone called it that.
Yield loss is negative revenue. Every ton of side trim, off-gauge ends, and rejected material was bought at coil price and leaves at scrap price. A yield rate creeping from 94% to 90% is a quiet double loss: you paid full price for tons you now sell for a fraction, and you paid to process them on the way through.
The five numbers that run the plant
- Blended margin per processed ton. Processing fee plus material spread minus yield loss, per ton across the whole book. This is the number that should be on the wall.
- Processing cost per ton, split fixed and variable. Because the fixed half is why throughput matters so much: a slitter running 60% of available hours is spreading the same mortgage over fewer tons. The throughput logic applies to a service center with full force.
- Revenue per line-hour. The line is the constraint. Downtime, coil changes, and changeovers between width campaigns are the service center version of the bottleneck problem, and an hour lost at the line is an hour of margin lost for the whole plant.
- Yield, by product and by customer spec. Yield loss is not evenly earned. One tight-tolerance contract can carry double the trim loss of another. If you cannot see yield by customer and spec, you cannot price by customer and spec, and the forgiving jobs subsidize the demanding ones.
- Book exposure vs index. What share of your sales book is firm-priced against inventory bought at a different level, what share floats with the index, and how many weeks of coil you are holding at today's replacement cost.
What the durable operators do differently
They price the yield into the quote. Tight-tolerance work carries its own trim loss, measured by job costing, charged in the price. This converts your worst cost driver into a pricing input and, more usefully, changes which work the sales team chases.
They match the book. Index-linked purchases against index-linked sales where customers will trade for it, firm-price commitments covered by firm-price inventory, and coil measured in weeks of supply rather than months. Hedging the liquid products caps the tail. Giving up some upside in a rising market to cap the downside in a falling one is not conservatism. It is matching a fixed-cost plant to a volatile top line.
They treat the escalators as revenue. Fuel, freight, and material surcharges written into supply agreements have to actually get billed. In our experience the gap between contracted and invoiced terms runs one to three points of margin, and in a thin-spread cycle that is the difference between black and red.
They know their numbers weekly, not at month-end. Mill prices move weekly. A service center managed on monthly statements is being steered by the rearview mirror through a market that turns faster than the reporting.
Why this is CFO work, not just plant work
Every lever above lives where the operating data meets the financials: pricing by yield requires cost-by-spec data; matching the book requires modeling the downside; escalators require contract-to-invoice discipline. That intersection is exactly the seat we fill for industrial distributors and processors, as both the finance function and the operating discipline around it.
If commodity swings keep repricing your business and the P&L only explains it a month later, start a conversation. We have sat in that seat.