In this guide
Low Value, High Leakage
Consumables are individually trivial and collectively substantial, and that combination is exactly why they leak. A single cutting insert, a drill bit, a pair of gloves or a litre of coolant is worth too little to justify a requisition somebody signs, so most engineering stores issue them freely and record them loosely. Multiply an unrecorded issue by a shift, by a year, by every consumable line, and the aggregate reaches a figure that would never be tolerated in any other category. Consumables attract the least control precisely because control looks disproportionate at the unit level, and the same reasoning is applied at every unit until nothing is controlled at all. What the aggregate looks like is a consumption figure that rises steadily without any corresponding rise in output, and a stores balance that is always approximately right and never exactly right. The workable answer is control at the aggregate rather than the item: consumption per machine hour, tracked and questioned when it moves.
Why Consumables Escape Control
Consumables are uncontrolled almost everywhere, and the reasons are structural rather than a matter of discipline. They sit below the threshold anyone watches. Every control system has an implicit or explicit value floor beneath which requisitions, approvals and reconciliations are not required, because applying them would cost more than the items are worth, and consumables live permanently under that floor. The judgement is correct item by item and wrong in aggregate. Issued freely to the shop floor is the operational consequence. Cutting inserts, drill bits, gloves, abrasives, coolant and fasteners are handed out on request because stopping production to raise paperwork for an item worth very little is obviously the wrong trade, and once issue is informal there is no record of who took what or why. Rarely counted and rarely reconciled completes the picture. A physical count of consumables is laborious relative to the value it verifies, so it is deferred, and where it is performed the differences are adjusted rather than investigated because nobody can trace an issue that was never recorded. The result is a category with no issue discipline, no count discipline and no reconciliation, which is a considerable amount of money.
What the Aggregate Actually Is
The case for controlling consumables rests entirely on adding them up, which most businesses have never done. Annual consumption against stock held is the first figure worth producing. It frequently shows a consumption several times the value of the stock carried at any moment, which means the annual expenditure is far larger than any balance sheet figure that ever drew attention. That is precisely why the category escapes notice: it is an expense line rather than an asset of any size. Consumption per unit of output is the control that actually works, because it converts an untrackable population into a ratio that can be watched. Inserts per thousand components machined, coolant per machine hour, or consumables cost per unit of production are all measurable without recording a single individual issue, and a movement in the ratio is informative even when nobody knows which issue caused it. Where the number surprises management is a common outcome of the first exercise. A category nobody considered material turns out to represent a meaningful proportion of conversion cost, and the ratio turns out to have drifted upward over several years without anybody having a reason for it.
Counting Items Nobody Counts
Verifying consumables requires a different method from verifying anything else, because a full count is neither affordable nor useful. Sampling by value and movement is the sensible approach: the higher-value consumables and the fast-moving ones are counted properly, and the long tail of trivial items is either counted in aggregate or accepted at book value with the basis disclosed. Counting every washer to the same standard as every carbide insert is effort spent in the wrong place. Weight and measure for bulk consumables replaces counting altogether where the item is not discrete. Coolant, oils, adhesives, welding consumables and abrasive media are measured by weight or volume, using the container's tare and a calibrated scale, and the conversion to a stock quantity is documented rather than assumed. Handling part-used items is the practical difficulty nobody anticipates. A partly used drum, a tool with half its life remaining, or an opened box are all real value and none of them counts as a whole unit, so a convention has to be agreed before the count: either they are recorded as fractions on a stated basis, or they are excluded and the exclusion is disclosed.
Evidence That Consumption Is Reasonable
Consumables are controlled at the aggregate because the unit is too small to control individually, so the evidence is a ratio rather than a count. Consumption is measured against the production volume that consumed it: inserts per thousand components machined, coolant per machine hour, gloves per shift worked. The ratio is the control, and it works precisely because it does not depend on anybody recording an individual issue. The second element is the trend across periods. A single period's ratio establishes nothing, since it might reflect a difficult batch, a new operator or a change in material. The same ratio drifting upward across several periods, on stable output and unchanged process, is evidence that something has changed which nobody has recorded. Third is the point where the ratio moves without an operational reason. Every genuine cause leaves a trace elsewhere: harder material, a new supplier, a machine running out of tolerance, a change in the product mix. A movement with no corresponding trace in any of those is the finding, and it is reached by elimination rather than by counting anything.
Bringing Consumables Under Control
Apply issue authorisation selectively rather than universally, because universal control on consumables costs more than it saves and is abandoned within months. Draw a value threshold, and above it require a requisition naming the machine or job the item is for. Below it, control at the aggregate. That split keeps the paperwork proportionate while capturing the lines that actually carry money, and it tends to survive because the shop floor can see the logic. Consumption norms are the control below the threshold. Establish what consumption per machine hour or per thousand units of output should be, from your own measured history rather than from a supplier's claim, and review variance against it monthly. The review is the control; the norm is only the reference point. A physical count is the only way to establish the baseline, because consumables stores are rarely accurate enough for the opening position to be taken from the system. Counting once, properly, is what makes every subsequent variance measurement meaningful, and stock audit for engineering plants covers stores alongside production material.
