The Lot-Size Decision: As Unit Cost Falls, Which Cost Rises?

When a supplier presents a price list broken into quantity tiers, the conversation across the table follows almost invariably the same trajectory: a quantity calculated for three months of consumption is stretched to six, sometimes to twelve, in order to reach the next tier, and the justification for that decision becomes visible on a single line — the fraction of a cent saved per unit. The burdens the same decision creates, by contrast, do not gather on one line; they disperse across ten or so independent items — warehouse space, insurance premium, counting labour, financing charge, obsolescence provision, engineering change cost — each sitting under a different budget owner. The benefit is concentrated and immediate, the cost dispersed and deferred, and for as long as that asymmetry persists, the direction in which the decision drifts is entirely predictable.

The same pattern appears in internal production planning, often in sharper form. On a press or an extrusion line carrying a long changeover, enlarging the run raises utilization, lifts units per shift, and improves the period performance indicator; within the same week, however, delivery of another variant in the product family begins to slip, work-in-process rises, and planning rebuilds the entire schedule for no reason other than a deferred die change. That these two movements — falling unit cost and narrowing delivery flexibility — are two faces of a single decision becomes apparent only when both are set side by side on one table. In practice the first resides in the production report and the second in the customer complaint log.

The tendency this configuration produces is known in operations literature as lot-size bias — the assumption that a larger batch lowers unit cost, obscuring the quality and flexibility costs it simultaneously creates. The defining characteristic of the tendency is that it is not an error; under specific conditions it is entirely functional. Where setup time is genuinely fixed, demand variance genuinely low, product design genuinely mature, and the supplier relationship genuinely stable, the arithmetic of enlarging the lot holds and the decision is rational. The difficulty lies not in the shortcut itself but in its persistence after the conditions that justified it have shifted; once setup time becomes a decision variable rather than a constant, or once the product family takes several revisions a year, the same formula continues to return the same answer, because nobody has reopened its inputs.

A second layer sustaining the tendency sits in the measurement and incentive architecture. With procurement typically assessed on unit price savings, manufacturing on line utilization or shift productivity, and logistics on cost per shipment, all three functions find their incentives pointing in the same direction, toward the larger lot. The only party structurally positioned to argue the other way is finance, yet finance tends to enter the discussion after the order has been placed, at the point where a cash flow variance is reported. The counter-argument is therefore institutionally unowned; nobody makes it because it is nobody's task, and the institutionalization of a tendency requires nothing more than that.

The third layer is the delay in quality feedback, and it is the mechanism that generates cost most directly. A defect originating in a process contaminates every unit produced up to the moment of detection, which makes the quantity exposed to that defect linearly related to lot size. When the defect surfaces at a downstream process, on an assembly line, or at a customer site, the question raised is not merely the scrap value but how many weeks the defect has circulated through the system, how many shipments it has entered, and how the recall boundary can be drawn. A smaller lot, while raising unit cost modestly, shortens the distance between error and detection and narrows the exposure by something closer to an order of magnitude — a gain that appears as a positive line in no standard cost report.

On the balance sheet, the consequence of the decision reads less in the absolute level of the inventory line than in that level's relationship to turnover and its seasonal behaviour. In a business where the large-lot habit has settled, inventory turns fall below the sector band, the cash conversion cycle lengthens, and working capital requirement grows faster than revenue rather than in step with it. This is among the more common structural explanations for why a company that looks profitable experiences persistent cash tightness; on the credit side, the same pattern inflates the net debt component, quietly compressing the headroom under covenant thresholds and raising the buffer required at periodic measurement dates.

The consequence of the same habit within an acquisition or investment process is harsher still, since the question asked at the diligence table is the question the company has never put to itself: what does the ageing distribution of the inventory line look like, which items have shown no movement in the last twelve months, and does the existing provisioning policy genuinely cover that distribution. Where the provision against aged stock proves insufficient, the finding converts directly into a price adjustment and, on occasion, into a higher escrow percentage. Beyond that, because the closing working capital peg is negotiated off historical averages, a large-lot habit pulls that peg upward and increases the cash the seller must leave in the business at closing; the unit price savings of prior years are, in effect, repaid at once and with interest at the moment of transaction.

The price paid on the flexibility side accumulates in the contract text. Tiered pricing secured against an annual volume commitment appears at the outset to strengthen the buyer's negotiating position, yet it inverts the relationship in the second half of the commitment period: unable to reduce volume, the buyer can no longer test the price against an alternative supplier, single-source dependency becomes formalized, and when quality performance deteriorates, the only leverage remaining is a volume the buyer has already promised to consume. Add to this the prospect of a product revision — an engineering change, a customer specification update, or a regulatory restriction on a material — and the entire quantity in the warehouse carries the risk of becoming unusable, with the burden falling not on procurement but on the finance function that books the provision at period end.

The mechanism that neutralizes the tendency is decision architecture rather than individual attention, and in practice it separates into three components. The first is that the lot-size decision be recorded in writing at the moment of proposal rather than at the moment of approval: a single paragraph stating which setup time assumption, which demand variance estimate, and which product revision expectation produced this quantity, so that a year later the accuracy of each assumption becomes auditable. The second is a threshold requiring that orders above a defined value, or above a defined number of months of consumption, be evaluated on total cost of ownership rather than unit price; absent a threshold, the exception becomes the rule. The third is that setup time be treated as a decision variable and remeasured jointly by engineering and manufacturing at least twice a year — because as setup time falls, the optimal lot shrinks on its own and the debate largely dissolves.

In capital-intensive and financed projects, BEIREK approaches this decision as a recorded item at the interface between engineering management and procurement. Rather than leaving order quantity to the supplier's price tiers alone, the delivery schedule for equipment and consumable packages is attached to the contract as a staged release programme; the committed volume is preserved with the supplier while the rhythm at which material — and therefore cash — lands on site is set against physical progress. Beyond flattening the project's cash curve, this structure keeps the prospect of specification change manageable, since at the moment of revision the quantity affected is not the full contract but only the releases already opened.

Its complement is a review discipline running on the monthly project rhythm and holding one question constant: in which items does the quantity held materially exceed what the next three months of work requires, and does the justification for that excess still hold. Because the justifications were recorded at the moment of proposal, the question functions as a reconciliation check rather than an accusation; ownership of the decision does not change, only the shelf life of the assumption is bounded. The credibility of the working capital projection reported to an investment committee or a lender depends directly on the existence of that record, since projection quality derives less from forecasting technique than from the traceability of the assumption underneath it.

A firm's lot-size policy is, in the end, the most concrete expression of its posture toward uncertainty: a large lot is a position taken on the proposition that the future will resemble the present, and the premium on that position is paid up front while the risk matures later. So long as the decision goes unexamined, the position simply rolls forward; the moment it is examined, the discussion moves off unit price and onto how much confidence the organization places in its own demand forecast — and whether that confidence can be substantiated is, ultimately, what the valuation turns on.

One question is worth putting: when did an order quantity last change not because of a price tier, but because setup time had been reduced?