There is a whiteboard in a service center outside of Dallas that the warehouse team calls "Mission Control." It tracks which orders are being processed, which trucks are loading, and which customers called to complain today. It has not been erased in three years. Layers of dry-erase marker have built up into a geological record of daily operations.
That whiteboard is the most honest piece of technology in the building. It exists because the five software systems running the business cannot talk to each other.
The Five-System Reality
A typical mid-size steel service center, 20 to 50 employees, $15 million to $60 million in revenue, runs some version of this technology stack:
Inventory: MetalTrax, INVEX, or an aging on-premise system installed during the Bush administration. It knows what is in the warehouse, most of the time, but cannot tell you whether a specific coil is already promised to someone.
Accounting: QuickBooks, Sage, or maybe a lower-tier ERP. It handles invoicing, AR, AP, and general ledger. It has no idea what the warehouse looks like.
CRM: Outlook contact folders, a shared spreadsheet, or if someone got ambitious, Salesforce with 40% of the fields filled in. Customer history lives in the sales rep's head and their email archive.
Quoting: Excel. Always Excel. Template spreadsheets with formulas that break when someone accidentally deletes a row. Pricing data that was accurate last week. Maybe.
Production scheduling: The whiteboard. Or a paper log book. Or the production manager's memory. Pick one.
The Hidden Cost: Duplicate Data Entry
Every order that flows through a disconnected service center gets typed into at least three systems. The sales rep enters the order details into the quoting spreadsheet. Someone re-enters it into the inventory system to allocate material. Someone else re-enters it into the accounting system to generate an invoice. The shipping coordinator enters it again to create the BOL.
Each entry takes 3 to 8 minutes. Four entries per order, 40 orders per day at a mid-size center. That is 8 to 21 hours of pure re-keying per day. Not value-added work. Not decision-making. Just typing the same information into different screens.
At a loaded labor cost of $35 per hour, that is $280 to $735 per day. Over a year: $70,000 to $185,000 in salary spent on copying data between systems that should be sharing it automatically.
The Hidden Cost: Errors
Every manual transfer is a chance for error. A transposed digit in the gauge: 0.060" becomes 0.006". A wrong customer address because the shipping system has an old entry. A pricing mistake because the sales rep referenced last month's spreadsheet instead of this month's.
The steel industry runs on precision. A 48-inch coil slit to four widths of 11.875" each is not the same as four widths of 12.000". The difference is 0.5 inches of material, but on a 40,000-pound coil, that mismatch produces parts that do not fit in the customer's stamping die. The mis-shipment comes back, you eat the freight both ways, the customer loses a day of production, and the relationship takes a hit that no apology fully repairs.
Service centers we have spoken with estimate that data entry errors cost them between $18,000 and $75,000 per year in direct rework, re-shipment, and credit costs. The indirect cost, lost customers who quietly switch to a competitor after one too many mistakes, is harder to measure and almost certainly larger.
The Hidden Cost: Delayed Decisions
The GM wants to know: what is our margin on the Johnson Controls account this quarter? In a unified system, that is a 10-second query. In a five-system environment, it requires pulling order data from the inventory system, matching it against invoices in the accounting system, factoring in processing costs from production records, and subtracting freight charges from shipping logs. Someone builds a spreadsheet. It takes half a day. By the time the answer arrives, the quarter is almost over.
Real-time visibility is not a luxury. It is the difference between managing a business and discovering what happened to your business after the fact. The service center that knows its margin on every order, every day, makes better pricing decisions than the one that calculates margin once a month from an accounting report.
The Hidden Cost: Customer Experience
A customer calls to check on their order. The sales rep checks the inventory system: yes, the material is allocated. But has it been processed? That information is on the whiteboard, so the rep puts the customer on hold and walks to the warehouse. Is it shipped? The shipping coordinator checks the BOL log. When will it arrive? Nobody knows because the carrier tracking number is in a different email thread.
Three minutes of hold time. Three people interrupted. And the answer is "let me get back to you." The customer hangs up and considers whether the service center across town might be easier to work with.
Now compare that to a single system where the rep types the customer name and sees: order placed Tuesday, processing completed Wednesday at 2 PM, shipped Thursday morning on carrier XYZ with tracking number and estimated delivery Friday by noon. The rep answers the question in 15 seconds without standing up. That is not better software. That is a better customer experience built on connected data.
The Math Nobody Wants to Do
Add it up for a 30-person service center doing $30 million in annual revenue:
Duplicate data entry: $120,000 per year in wasted labor.
Error-related costs: $40,000 per year in rework and credits.
Decision delays: Impossible to quantify precisely, but a conservative 0.5% margin improvement from better pricing data on $30 million in revenue is $150,000.
Software licensing for five separate tools: $40,000 to $80,000 per year.
The total cost of running disconnected systems is not the $40,000 you pay for software licenses. It is the $300,000 to $400,000 you pay in inefficiency, errors, and missed margin that the five-system architecture makes invisible.
The whiteboard in Dallas is still there. It works. But the business it supports is leaving money on the floor every day, and the whiteboard cannot tell them how much.