Small plants: what a system genuinely has to do, and what gets sold instead

The spreadsheet works until several people need it

A well-built spreadsheet maintained by one person who understands the business is a genuinely effective tool, and telling a small manufacturer otherwise is usually wrong. The failure appears when more than one person needs the same answer at the same time: two versions circulating, a promise made from a stale copy, nobody able to reconstruct why a figure changed. Those symptoms, not company size, are the signal. Watch also for the concentration risk, since a plant whose scheduling exists solely in one laptop has a continuity problem regardless of anything else.

What to systemise first, and in what order

One item list that everybody uses comes first, because every later capability references it. Then order status visible without asking, so the office stops interrupting the floor. Then stock, at whatever accuracy the plant can sustain rather than the accuracy a consultant proposes. Job costing follows once bookings are reliable enough to mean something. Scheduling comes last, because it depends on routings and capacity data that a small plant will not have until the earlier stages have been running for a while and generating real numbers.

Capability that will not be used

Advanced scheduling on routings nobody has timed produces a plan the shop floor ignores. A full quality module with no one to administer it becomes a login nobody has. Demand forecasting on a handful of customers who phone their orders in adds nothing. These modules are bought because they are in the package and because they describe a future the business intends to reach, and they carry configuration effort, training and licence cost in the meantime. Buy for the plant you are, with a documented path to what you might become.

The administration a small plant cannot absorb

Every system carries an ongoing load: users, parameters, price updates, new items, corrections, backups, upgrades. In a large plant that is somebody's role. In a small one it lands on a production manager who already has a full day. The practical consequence is that configuration complexity has to be kept low deliberately, even at the cost of some capability, and that the plant needs a supplier who will do the routine administration under a support agreement. Systems that survive here are the ones a busy non-specialist can still operate correctly after a year.

Running the transition without a project team

There will be no dedicated resource, so the plan has to assume part-time effort and tolerate interruption. That argues for staged introduction with a working outcome at each stage rather than a single date on which everything changes. Keep the old method running alongside the first stage until people trust the new numbers, then stop it deliberately, because the real danger is halting halfway and operating two partial systems indefinitely. Setting a firm date to switch off each spreadsheet is a small act of discipline that decides whether the transition finishes.

Frequently asked questions

Do we need a full system or something narrower?

Start from the problem rather than the category. A plant losing money on quoting needs estimating and job costing. A plant missing deliveries needs order visibility and stock accuracy. A plant failing customer audits needs traceability and document control. Narrow tools that do one of those well are often a better first step, provided they can export their data cleanly. The risk of the narrow route is accumulating several disconnected tools, so agree in advance which one will hold the item list.

How do we implement without a project team?

Accept a longer elapsed period and protect a fixed slice of time each week rather than pretending someone is available full time. Pick stages small enough to finish between other commitments, and get one thing genuinely working before starting the next. Use the supplier's people for configuration and data loading where it is cheaper than your own time, but keep the decisions internal, because a configuration chosen by someone who will not be here is how a small plant ends up with a system nobody can explain.

Is there a size below which a system is not worth it?

There is no headcount threshold, but there is a complexity one. A workshop making a few repeat products for a handful of customers can run on well-maintained spreadsheets almost indefinitely. The threshold is crossed when the number of open jobs, part numbers or customer variants exceeds what one person can hold in their head, or when the business needs to answer questions about the past that nobody recorded. Those conditions arrive at very different sizes in different trades.