Custom ERP or off-the-shelf.

Nobody wants an ERP. What a factory wants is to stop reconciling — to stop having a stock figure in one place, a costing in another, and a number the owner actually believes in a third. The software is the means.

That reframes the decision. It is not really ready-made against custom. It is a question of who absorbs the mismatch between how the software expects you to work and how your plant actually works. Off-the-shelf asks the plant to bend. Custom bends the software. Both cost money; they just bill it differently.

A ready-made system is the right answer more often than software companies like to admit, and it is worth being specific about why.

It starts fast — weeks, not months. Somebody else maintains it, patches it and keeps it current with tax rules that change without asking you. The parts of a business that are the same everywhere, general ledger, payroll, VAT returns, are already built and already tested by thousands of companies. You can hire someone who has used it before. And when it breaks at 11pm there is a support line that is not one developer's phone.

If your processes are close to standard, buying is not the cheap option. It is the correct one.

The mismatch tends to show up in the same places. Costing a job when the inputs are a board grade, a machine hour and a setup, rather than a purchase price. Units that change down the line, so material arrives by weight, is consumed by area and ships by piece. Partial deliveries against one order. Tooling that belongs to a customer but sits in your building. Scrap that has to be attributed to a job rather than written off in a lump at month end.

None of these is exotic. They are simply specific, and specific is exactly what a general system flattens.

There is a reliable tell for when the flattening has gone too far. If the numbers people actually trust live in a spreadsheet beside the system rather than inside it, you already have a custom system — an undocumented one, maintained by whoever built the sheet, with no backup and no successor. You are paying for custom software already. You are just paying for it badly.

Custom is worth it when the process is the advantage. If a plant wins work because it quotes in a day where competitors take a week, that quoting process is not overhead to be standardised — it is the reason customers call. Software that forces it into a generic RFQ workflow removes the thing being sold.

It is also worth it when the mismatch is recurring rather than one-off. A workaround that costs two people an hour a day is roughly a full-time salary a year, paid forever, and it never appears on the software line of the budget.

And it is worth it when systems have to meet. A plant with a website that takes specifications, a floor that schedules against them and a finance function that costs them is describing one flow, not three products. Joining ready-made systems at the seams is frequently more work than building the thing that spans them.

Four questions, answered honestly, decide this faster than any feature comparison.

First: write down the five things your plant does differently from a competitor. If none of them touch the system, buy off-the-shelf. Second: count the spreadsheets that exist because the system cannot do something, and price the hours in them. Third: ask what happens when the person who maintains those spreadsheets leaves. Fourth: ask whether you would still want this software if you doubled in size — ready-made systems usually scale in volume, but not always in complexity.

A sensible answer is often neither extreme: a ready-made core for finance, with a custom layer for the part of the plant that is genuinely yours. That is unglamorous, and it is frequently right.

Vetta builds ERP systems for manufacturers in Egypt and worldwide, and we build our own, so the tradeoffs above are ones we pay for too. We will tell you when a ready-made system is the better buy — the projects that go wrong are the ones that should never have been custom.