The Invoices Are Fine. The Costs Are Gone. | Manufacturing Accounting Software And Job Costing

If you run a small manufacturing business, your accounting system is probably working perfectly. Invoices go out. They are correct. Payables are current. The bank reconciles. Nobody is complaining about it. And you still cannot tell me what last month's gross margin was by product line. Neither can your controller, not with any confidence, and not until the accountant closes the year. That is not a bookkeeping failure. It is the specific way an accounting system runs out of road inside a manufacturer, and it is worth understanding, because the part that breaks first is not the part anyone is watching. Everything on screen is an illustrative example. Invented company, invented numbers, nobody's real books. Here is the thing to hold onto. Your accounting system is not wrong. It is uninformed. An invoice is easy. Price times quantity, terms, tax. Every number an invoice needs is already inside the system, because somebody typed it there when they took the order. So invoicing keeps working, correctly, right up until the last day you own the business. A cost is a different kind of number. To know what a job cost, the system needs to know what actually left the stockroom, how many hours actually went into it, and what got scrapped and remade. Not one of those three is born in accounting. All three happen on the floor. So the ledger has one side of the story in real time, and the other side never. And it is perfectly happy about that. Follow one job through. A customer orders four hundred units. Somebody builds a quote, and the quote uses a standard cost that came from a spreadsheet a previous operations manager maintained. Production runs it. They pull more material than the bill of materials called for, because the first run had a problem. Two people work a Saturday. Eleven units fail inspection and get remade with new material. The job ships. The invoice goes out, four hundred units at the quoted price, and it is entirely correct. Now, in the accounting system, what happened? A revenue line. A materials purchase, dated whenever the pallet arrived, which was three weeks earlier and attached to no job at all. And payroll, in a lump, for the period. The margin on that job exists. It happened. It is simply not written down anywhere you can read it. So there are exactly three numbers missing, and they are worth naming, because everything else is built on them. Material actually consumed by this job, rather than material purchased that month. Labour hours actually recorded against this job, rather than payroll for the period. And scrap and rework, which is the one almost nobody captures, and which is where the surprises live. Scrap is worth a sentence on its own. It is not missing because it is hard to record. It is missing because recording it means somebody writing down that a thing went wrong, on a form, with their name on it. That is a management problem wearing a data problem's clothes. Get those three and you can compute a real job cost, a real margin, and eventually a standard cost that is not a rumour. Miss them and every number downstream is an average of your good jobs and your bad jobs mixed together. Which is a number that describes nothing, and which you are currently using to price next year's work. Here is why you find out late. Nothing about the missing numbers makes the accounting system complain. It does not throw an error. There is no red figure on any report. Everything reconciles, because everything it was given has been accounted for. It was just never given the rest. The gap becomes visible in exactly two ways. Somebody counts the physical stock and it does not match the books. Or the accountant sets closing inventory and the cost of goods sold moves in a direction nobody predicted. That is once a year in most businesses. Quarterly if you are disciplined. And there is a name for the number nobody can read: work in progress. The value of what is on the floor right now, part-finished. In a shop without job-level capture, that figure is an estimate somebody makes once a year, and everything above it in the accounts inherits the error. By the time it surfaces, the jobs that lost money shipped nine months ago, and you have spent those nine months quoting new work off an average those jobs are hiding inside. If you recognise yourself in this, the first move is not buying an ERP. The first move is capturing those three numbers against a work order, wherever you can capture them. A work-order record with material issued, hours recorded and scrap counted is worth more than a suite nobody on the floor will use. That is also the honest test of whether you are ready to buy anything larger. A business already capturing job-level material, labour and scrap can evaluate systems on their merits, because it knows what it needs them to do. A business that is not will buy an ERP, spend nine months implementing it, and rebuild exactly the same blind spot inside it. So, plainly. An accounting system fails in a particular order inside a manufacturer. Costing first, invoicing last, and it will not tell you which stage you are at. If you cannot answer what a job cost, this month, without waiting for your accountant, that is the signal. Not the invoices. Not the bank. We are CodeStringers. We build and connect the systems operationally complex businesses run on, and in a shop this usually starts with getting the work order to talk to the ledger rather than replacing either of them. Finding where your own three numbers currently die is a short, scoped piece of work, and it is where we normally start.