Manufacturing ERP for a Small Business: Fix the Item Master Before You Buy Costing.
When a small manufacturer starts looking for an ERP, the shopping list is usually the same. Inventory, purchasing, sales orders, production and costing. Costing sits at the end of the list and at the top of the wish list, because it is the module that promises to answer a question the owner has had for years: which jobs actually make money?
It is also the module that depends most on something nobody puts on the shopping list: the item master. A costing engine does not discover the true cost of a product. It multiplies what the item master and the bills of material say by what the stock transactions say, and it does so precisely. A precise calculation on records that are wrong gives a wrong answer with decimal places.
The item master is the list of every thing the business buys, makes or sells: raw materials, components, finished goods, packaging and services. Each has a code, a description, a unit of measure, a standard cost and a handful of attributes. In most small manufacturers it has grown for years, in a spreadsheet or in whatever the accounting package allowed, and it shows.
The same part under two codes, one from the supplier's catalogue and one the shop invented. A description that says "bracket, small" on one line and "small bracket" on another. A unit of measure that is each in purchasing, boxes of twelve on the shelf and metres in the production notes. A component that was replaced two years ago and is still in four bills of material. A finished good with no bill of material at all, because the first one was easy and nobody wrote it down.
None of that stops the business from running. A person who knows the shop knows which "small bracket" is meant. A system does not. It treats two codes as two items, two stock counts and two costs, and the average cost on each is wrong in a different direction.
A cost is a sum of lines: material at its cost per unit times its quantity, labour at a rate times a time, and overhead at an allocation. Every term in it comes from a record.
If the unit of measure on a purchase is different from the unit on the bill of material, the quantity is out by a factor. If an obsolete component is still on the bill of material, it is costed and consumed on paper. If the same item has two codes, receipts go against one and consumption against the other, so one shows rising stock and a falling cost per unit, and the other goes negative. If the bill of material has no scrap allowance, the cost is lower than the real cost on every run.
The owner then looks at a job cost report and sees a margin. It looks authoritative, because it comes from a system with a costing module. It is wrong by an amount nobody can see, and the pricing decisions it informs are wrong by the same amount. We described the symptom, invoices that are fine and costs that are gone, in job costing for light manufacturers. This is the cause one step earlier.
First, the item master. One code per item, a naming convention a stranger could follow, one unit of measure for storage and a conversion table for each other unit, and a status field that says active or obsolete. This is a clean-up, not a project: for a few thousand items it is a few weeks of part-time attention from someone who knows the parts, ending in a list they sign.
Second, the bills of material. One current version per finished item, components by code and quantity in the storage unit, a scrap allowance where scrap is real, and a retirement date on each old version. The same discipline as the item master, applied to the next table, and we looked at what happens when the bill lives in three places in bill of materials management.
Third, the stock transactions. Receipts, issues to production, finished goods in and shipments out, each recorded when it happens and against the item master's code. This is where discipline on the floor matters, and it is much easier when the first two are right, because each transaction has an unambiguous code and unit.
Last, costing. Only now does the module have honest inputs. The first cost run will still be imperfect, and the differences between the cost run and the actual cost will show you which records need attention. But the differences will be small enough to read, and each one will point to a specific item.
Export the item list and sort it by description. Duplicates sit next to each other, and a person who knows the parts can mark them in an afternoon. Add a column for the unit the item is bought in and the unit it is stored in; wherever they differ, write the conversion. Filter for items with no movement in eighteen months and mark them obsolete rather than deleting them, so history keeps its meaning. Then list the finished goods and check each has exactly one current bill of material.
The output is a signed list and a short exceptions file of things only the owner can decide. It is unglamorous and it is the cheapest accuracy a small manufacturer can buy.
Ask vendors how their system handles units of measure and item versions, and look at how they answer. A system that enforces a storage unit and a conversion table is protecting the thing you need protected. A system that lets any unit on any line is saving you set-up time and costing you accuracy.
Ask yourself two questions before the demo. Could someone who did not build the item master tell, from the code and description, exactly what each item is? And do you know how many items have no bill of material, or more than one? If the answer to the first is no and the second is a guess, the first project is the clean-up, whichever system you choose.
Small manufacturers tend to assume that data discipline is a large-company problem, and that a small business can run on knowledge. That is true until the business adds a second person who orders parts, or a second shift, or a second sales channel. At that point the knowledge in one head stops being enough, and the item master is what is left to hold it.
We run the item master and bill of material clean-up as the first phase of an ERP implementation, before any module is configured, and we connect inventory, accounting and costing on the cleaned records. Discovery is no-risk: we profile your item master, count the duplicates, unit conflicts and orphaned bills, and tell you what costing would produce today, and you pay only if you go ahead. The build carries a guaranteed estimate, so if we estimate low, we absorb the difference.
The module that tells you what a job really cost is only as honest as the list of things it is costing. Fix the list first.
In a no-risk discovery we follow an order from sale to shipment to invoice across your systems and show where it breaks and what one connected system would change. You pay only if you proceed. Or see how we approach it.
More on the same problem:
"Something isn't working." "We're outgrowing our tools." "We're afraid to make the wrong move." No-Risk Discovery is a short, practical conversation that gets you clarity before you commit to anything big. We'll tell you if we're a fit. If we're not, we'll tell you that too.
Say who you are and ask. A person replies within one business day.