Metal Service Center and Coil Processing Software: Where Does the Cost and the Heat Number Go When One Coil Becomes Six Mults?

Why a service center is not a distributor with extra steps

A 48 inch wide coil arrives, 21,400 pounds, one heat, one mill test report. It goes on the slitter and comes off as six mults: two at 6 inches for a stamper, three at 10 inches going into stock, one at 4 inches that nobody ordered because the arithmetic left it over. Two of those mults ship this week to different customers, each needing the mill test report for that heat. The third gets pickled at a third party, comes back three weeks later lighter and with a processing invoice attached. The last two sit in stock while the market price of hot rolled moves.

Now answer a simple question: what did the mult that shipped Tuesday cost you, and what did you make on it. In most service centers the answer comes from a standard cost per hundredweight, an assumed processing rate, and an assumption about scrap that was set some years ago. The real cost depends on the purchase price of that specific coil, the actual yield on that specific slit job, the setup time on the specific line, and whether the customer was billed on theoretical or actual weight.

The stack is usually a metals ERP such as Invera STRATIX, Enmark eSTELplan or Compusource, or for smaller houses a general distribution ERP plus a very hardworking spreadsheet. The metals products know this industry, which is exactly why they exist, and they are strong on the fundamentals of coil inventory and heat tracking. Where operators run out of road is elsewhere: dated user interfaces that slow down the sales desk, limited API surfaces when you want a customer portal or a shop floor tablet, and customisation that has to go through the vendor and their release cycle. A general distribution ERP is worse, because it thinks inventory is a quantity of a part number, and yours is not.

Problem 1: one item becomes many, and the accounting has to follow

Standard inventory logic is subtract from a quantity. Metals logic is that a parent item ceases to exist and several children come into being, each with its own dimensions, its own weight, its own tag and its own location, all inheriting one heat number and a share of the parent cost.

Getting that allocation right matters. Allocate cost purely by weight and you understate the value of the wide mult and overstate the narrow one, when the market says otherwise. Allocate by market value and your inventory ledger diverges from purchase reality. Then there is the skeleton and the scrap, which have their own recovery value and must be credited somewhere or your job margins are permanently wrong.

What a custom build does: model the transformation event explicitly. Parent piece in, children out, scrap out, with a documented allocation rule per process that you choose rather than inherit. Every child carries a complete lineage back to the receipt and therefore to the purchase price you actually paid. Reversal is supported, because operators do slit a coil, discover a defect and reprocess, and a system that cannot unwind an event forces staff into corrections that destroy the audit trail.

Problem 2: theoretical weight, actual weight and where your margin quietly lives

You bought that coil on actual weight from the mill. Your customer buys 10 inch by 0.075 inch strip on theoretical weight computed from dimensions and density. Those two numbers are never the same, and the gap changes with thickness tolerance, with how the mill was running, and with the specific grade.

Most systems carry one weight field and a conversion. That is enough to invoice and not enough to manage, because it hides a real and controllable margin component. Some houses are systematically losing on it and have no idea because it never appears as a line on any report.

What a custom build does: carry actual and theoretical weight on every piece at every stage, compute the difference at receipt, after processing and at shipment, and report it by supplier mill, by grade, by thickness and by customer. That reporting alone tends to change purchasing conversations, because the pattern of which mill runs to the high or low side of tolerance is visible for the first time. Invoicing then follows the contract, theoretical for one customer, actual for another, priced per hundredweight or per piece as agreed, without a salesperson recalculating anything.

Problem 3: the mill test report has to arrive with the metal

Customers in automotive, structural, pressure vessel and defence work will not accept material without the certificate for the heat it came from. Some want it attached to the packing list, some want it uploaded to their portal, some want it emailed before the truck arrives. A shipment may contain material from three heats, which means three certificates and a clear mapping of which piece came from which.

The manual version of this is a person in the shipping office searching a folder by heat number, opening PDFs and stapling. It fails in two predictable ways: the wrong certificate goes out, which becomes a quality complaint, or nothing goes out and the truck is held.

What a custom build does: the certificate is stored against the heat at receipt, with the key values extracted so they are searchable rather than trapped in a PDF. Every piece already knows its heat, so the shipment assembles its own certificate pack automatically, in the customer preferred format and delivery method. When a customer calls two years later about a failed part, you produce the certificate and the full processing history for that piece in seconds. That is also the difference between a contained complaint and a general recall of everything you shipped them.

Problem 4: processing cost per job is not a rate card

Slitting a soft grade at full width on a good line is not the same job as slitting a high strength grade in narrow mults with six knife changes. Cut to length with a stretcher leveller on 0.5 inch plate is not the same as blanking thin gauge. Yet most service centers cost processing with a per hundredweight rate that averages all of it, so the difficult jobs subsidise the easy ones and the sales desk quotes accordingly.

Then there is outside processing. Material leaves for pickling, galvanising, painting or heat treatment, sits on somebody else's floor, comes back with a different weight and an invoice. Many systems handle this by writing the material off and receiving it back as new, which destroys lineage and makes the true cost unknowable.

What a custom build does: routings with setup and run time per line and per grade family, so job cost reflects the actual work, plus a proper outside processing model where the material stays yours, keeps its lineage and accrues the vendor cost when it returns. Then margin by job, by customer and by process becomes real. The frequent discovery is that a long standing account with high volume and constant narrow mults is the least profitable relationship in the building.

Problem 5: metal price movement makes stale quotes dangerous

You hold inventory bought at one price and quote against replacement cost, or you should. When hot rolled moves quickly, a quote issued three weeks ago at yesterday's cost is a loss waiting to be accepted. Contract accounts with quarterly pricing add another layer, as do consignment and just in time programmes where you hold stock for a customer against a forecast that they will miss.

What a custom build does: quotes carry an explicit cost basis and a validity period, with visibility of both current inventory cost and current replacement cost at quote time. Contract price lists, index linked adjustments and customer specific extras live as data, and consignment stock is tracked as your inventory at their location with its own ageing report.

What this costs and how long it takes

Across the 2,000 plus projects Digital Heroes has delivered, a service center first release runs $80,000 to $170,000 in 14 to 20 weeks. That covers multi dimensional inventory with parent to child transformation, heat traceability and certificate handling, order entry with theoretical and actual weight, and shipment. It is a system the sales desk and the shop use on day one. The full platform, adding processing routings and job costing, outside processing, quoting with price basis, contract and consignment programmes, purchasing with mill claims, and a customer portal, runs $200,000 to $500,000 phased over 8 to 15 months.

Cost drivers particular to metal service centers:

• The number of distinct processes, since slitting, cut to length, blanking, plate burning, sawing and tube cutting each have their own yield and scrap behaviour.
• Whether you run multiple locations with transfers, which doubles the inventory model and adds transfer costing.
• Line and equipment integration, such as pulling actual weights from a floor scale or a coil car rather than typing them.
• Certificate extraction, if you want incoming mill test reports read into searchable data rather than filed as PDFs.
• EDI, since automotive and large industrial customers will send releases and expect advance shipping notices, and each trading partner is its own piece of work.

What keeps cost down: start with one location and your two highest volume processes. Plate burning and tube can wait.

Build versus buy, and when the metals ERPs win

Buy Invera, Enmark or Compusource if you are a conventional service center whose processes match the industry model and whose main need is a solid transactional backbone. They understand coil, heat numbers and hundredweight pricing in a way general ERP never will, and replacing them with a custom build purely for a nicer interface is not a good use of capital.

Build when two or more of these are true. You run processes or programmes the package does not model, and you are maintaining spreadsheets alongside it to make the business work. You need a customer portal, mobile shop floor entry or EDI depth that the package cannot expose because its API surface is thin. You cannot get job level margin including actual yield and real processing time, so pricing is instinct. You operate several locations with transfers and toll processing and reconciliation is manual. Your quoting ignores replacement cost in a moving market.

The honest middle path is common. Keep the metals ERP as the financial and transactional core, and build the layer around it: portal, shop floor capture, job costing analytics, certificate automation. That is often a $60,000 to $120,000 project rather than a replacement programme, and it addresses most of what actually hurts.

How to choose a developer for service center software

Ask them to model a coil being slit on a whiteboard, including scrap, skeleton and cost allocation to children. If they reach for a bill of materials with a quantity, they are thinking in manufacturing assembly terms and will produce something that cannot represent your inventory.

Ask how they carry theoretical and actual weight. If there is one weight field, walk. That single design decision determines whether you can ever see the margin component that lives in the gap.

Ask how outside processing keeps lineage. A build that writes material off to a vendor and receives it back as a new item has destroyed your traceability and your costing in one step, and you will not notice until a customer asks for a certificate.

Ask what they have integrated. Pulling data from Invera or eSTELplan, receiving an automotive EDI release, and reading a floor scale over a serial connection are three different problems. Ask for the specific system and the specific document type.

Ask who owns the code and settle it in writing before kickoff. You should own the repository, the infrastructure accounts and the right to hire anyone else. At Digital Heroes the code is yours from the first commit. Your inventory and certificate history is the record of your business, and it should never sit in a system you cannot leave.