In a production planning meeting, the least contested number on the table is usually the batch size. Demand forecasts are challenged, delivery dates are negotiated, shift patterns are rebuilt from scratch, yet the question of how many units a line runs in a single pass rarely surfaces, because that figure is perceived not as a decision but as a physical property of the equipment. When it is raised, the answer offered tends to be historical: the batch was sized large because the die change took a long time. That is a defensible rationale, and it was almost certainly correct when it was formed. Whether anyone has measured how long the die change takes today is a separate question. In the intervening years a portion of the line may have been automated, fixtures standardized, operator experience deepened; the parameter, meanwhile, stays exactly where it was placed.
The same pattern is more visible on the purchasing side. When a supplier's price schedule offers a meaningful per-unit improvement above a given quantity, the order rounds up to that tier; the discount captured appears as a measurable line in the procurement performance report, while the carrying cost of the additional inventory created appears nowhere in the same report. The asymmetry here is not in the pricing but in the approval architecture. In most organizations, increasing an order quantity requires no supplementary justification, whereas reducing it — particularly moving to more frequent, smaller replenishment — requires an explanation, a calculation and, more often than not, an additional layer of sign-off. What pushes the decision-maker in a predictable direction is not preference; it is the fact that the two directions do not cost the same to choose.
The pattern has a name — **large-batch inertia** — the persistence of a batch-sizing decision after the conditions that produced it have dissolved. The mechanism is embedded in the economic order quantity logic itself: batch size is derived from the trade-off between setup cost and carrying cost, and once setup cost has been assumed high, the formula validates the large batch mathematically. The defect is not in the formula but in the currency of its inputs. To the extent that changeover time goes unmeasured, a value observed at some earlier point hardens into a fixed property of the process. Inertia consequently does not live in anyone's attention span; it lives inside a system parameter that no one experiences as a choice, and that therefore no one is required to defend.
A second layer is physical and organizational. As the batch grows, the infrastructure carrying it grows with it — rack configuration, pallet dimensions, work-in-process staging area, handling equipment, and the minimum order and reorder point fields inside the ERP are all calibrated to that magnitude. After some period, a proposal to reduce batch size ceases to be a planning preference and becomes an intervention that opens warehouse layout, supplier contracts and system parameters simultaneously. The threshold cost of change therefore rises incrementally with every quarter that passes. None of this indicates that the original decision was wrong; it indicates only that the decision has become progressively more competent at defending itself, which is a different property altogether and one that operates independently of whether it remains correct.
A third layer is the timing of feedback. The benefit of the large batch materializes within the current period, in directly monitored indicators such as unit cost and machine utilization. The cost distributes itself across subsequent periods, surfacing as decelerating turnover, aging stock-keeping units and diminished line flexibility. The unit that measures the benefit is frequently not the unit that carries the cost: unit cost is booked to manufacturing, carrying cost to finance. So long as that separation holds, what makes the choice rational stops being the operating condition and becomes the location of an internal accounting boundary — a distinction that rarely appears in any document, yet governs which of two defensible positions prevails when the two functions disagree.
The first place the institutional cost becomes visible is not the inventory balance on the balance sheet but the turnover of that balance. The cash conversion cycle can lengthen without total inventory growing at all, because the same monetary value migrates toward a composition that moves more slowly and concentrates in fewer line items. Each additional day in the working capital cycle represents, beyond a financing cost, a loss of optionality; in a structure running on large batches, when the demand mix shifts, the company's cash is committed not to the right product but to the product that was right last quarter. That consequence is not legible in a short-term liquidity statement. It is legible in how long it takes to answer a new customer request.
The second cost item is quantity at risk. When a quality deviation, a raw material lot inconsistency or an engineering change occurs, the volume affected is precisely the batch size; the large batch does not make the error cheaper, it delays detection and consolidates the remediation cost into a single event. The same logic governs product revisions: stock sitting on the rack becomes, at the moment a revision takes effect, an item requiring either rework or a provision. On the supply side, running large batches with a single-source supplier typically produces a specific outcome — bargaining power gained on price and surrendered on delivery flexibility, a trade that looks favorable in the contract and unfavorable the first time a schedule moves.
The third and frequently most expensive cost emerges when the company enters a sale, partnership or credit process. The party on the diligence side does not examine the headline inventory figure; it examines the aging schedule and turnover at the item level, provisions for anything above a defined age threshold, and flows that provision directly into the net working capital calculation. To the extent that the normalized working capital peg is constructed from trailing period averages, inflated inventory may appear to favor the seller, yet the buyer's disaggregation of the slow-moving portion generally produces a closing adjustment, an increased escrow ratio, or an earn-out structure anchored to an operational threshold such as inventory turns. An operating habit thus converts into a component of transaction price.
What neutralizes this tendency is decision architecture rather than individual vigilance, and the intervention decomposes into four separable components. The first is elevating setup and changeover time into an independently tracked performance measure; once batch size is re-derived as the output of a measured changeover rather than accepted as an input, the inertia is removed from inside the formula. The second is converting batch and reorder parameters from ownerless system fields into fields attached to a named owner and a defined review cadence. The third is inverting the approval asymmetry, so that increasing a quantity requires at least as much justification as reducing one. The fourth is moving the inventory aging schedule out of the financial reporting appendix and into the opening document of the planning meeting.
BEIREK typically begins such engagements by constructing a parameter inventory: which batch size applies to which item, when that value was last set, under what changeover assumption, and whether the person who set the assumption and the measurement behind it still exist. That this inventory has never been maintained in most companies is itself the finding, since it identifies precisely where institutional memory broke. Two registers follow — a changeover measurement protocol and a batch-size change log — attached to a monthly review cadence. The objective is not to make batches smaller. The objective is to make batch size a decision that must be justified again each period, which is a structurally different condition from being correct.
In capital investment decisions the same discipline is applied one step earlier. In a portion of the investment requests presented on capacity-constraint grounds, what generates the constraint is not demand itself but the way the prevailing batch logic occupies the line; an additional line or additional warehouse bay, in that configuration, finances a parametric constraint rather than a real one. Opening this distinction as a discrete heading in the technical review of the investment — placing maximum achievable output under the current batch structure alongside maximum achievable output under a revised one — frequently changes not the size of the investment but its timing, and that alone is material enough to affect the cost of the financing structure surrounding it.
A company's flexibility is not observable in the breadth of its product range or the vintage of its equipment. It is observable in when the numbers it runs on were last changed. Batch size is the quietest of those numbers — defended by no one, owned by no one, and precisely for that reason never examined — sitting inside the balance sheet as an assumption that has outlived the measurement that produced it.