Where Manufacturing Overhead Actually Hides: Inventory Carry, Cost of Poor Quality and What Standard Costing Will Not Show You

TL;DR: Manufacturing overhead is defined as a residual, meaning everything in production that is not direct material or direct labour, and residual categories do not get managed. Three costs hide inside it: inventory carrying cost, which sits on the balance sheet instead of the profit and loss account; cost of poor quality, which the American Society for Quality splits into four categories of which most manufacturers formally track two; and unused capacity, which absorption accounting quietly converts into inventory value. US manufacturers held $966.9 billion of inventory in July 2026 with an inventories-to-shipments ratio of 1.47, per the US Census Bureau M3 survey released on 2 September 2026. Standard costing will report all of this as a variance, which tells you that the number moved but not which decision moved it.

I spent the first part of my career in manufacturing and distribution at Hindustan Unilever, Whirlpool and GE, then a decade running finance at ACT Fibernet and BlackBuck. The pattern repeats. When a plant misses margin, the first meeting is about price and raw material. The second is about labour. The third, if anyone gets that far, is about overhead, and by then everyone has accepted that overhead is a fixed block you allocate rather than a set of decisions you make.

That third meeting is where the money is.

What manufacturing overhead is, and why the definition is the problem

Manufacturing overhead is defined as every production cost other than direct materials and direct labour. It covers indirect labour, indirect materials, factory rent, plant utilities, equipment depreciation, supervision, inspection, maintenance and factory insurance.

Read that again. It is a residual, the bucket for everything that could not be traced cleanly to a unit. That is a sound accounting convention and a poor management category, because a bucket defined by what it excludes gives you no handle on what it contains.

The practical consequence is that overhead gets a rate rather than an owner. Finance computes a predetermined overhead rate, applies it to production volume, and reports the gap between applied and actual at period end. Nobody in that chain is asked which activity created the cost, or whether the activity was necessary.

Cost discipline is not cost cutting. Cost cutting takes a percentage off every line. Cost discipline reviews the operations behind the lines and asks which activities, and therefore which costs, can be deferred or eliminated without damaging the business. You cannot do the second thing to a residual bucket. You have to break it back into causes first.

The three costs that live inside the bucket

The three largest are inventory carrying cost, cost of poor quality, and the cost of capacity you paid for and did not use. Each is invisible for a different structural reason, which is why one fix does not surface all three.

What standard costing will not show you

Standard costing is a method that assigns a predetermined cost per unit for material, labour and overhead, then reports the difference between that standard and actual results as variances. It is efficient, it is auditable, and it is still the backbone of most manufacturing cost systems.

It has three blind spots that matter to anyone making decisions rather than closing books.

The standard is an assumption, not a benchmark. A favourable variance means you beat last year’s assumption. It does not mean the process is efficient. If the standard was set with three percent scrap baked in, three percent scrap reports as zero variance forever. The waste becomes structurally invisible the moment it becomes normal.

Absorption ties overhead to volume, so building inventory improves reported margin. Under absorption costing, fixed overhead attaches to units produced. Produce more than you sell and a portion of this period’s fixed overhead moves onto the balance sheet inside finished goods rather than hitting the profit and loss account. Gross margin looks better. Cash is worse. This is legitimate accounting and a genuinely dangerous incentive when plant managers are measured on absorption.

Variances tell you the number moved, not which decision moved it. An unfavourable overhead spending variance of $400,000 could be energy prices, unplanned maintenance, expedited freight covering a quality escape, or overtime to rework a batch. Four different problems, four different owners, one line on the report.

Inventory carry is a real cost that sits on the balance sheet instead of the profit and loss account

Inventory carrying cost is the total annual cost of holding stock, comprising the financial cost of capital tied up in it, the cost of storing it, and the cost of it being damaged, obsolete or shrunk.

Here is why it disappears. Inventory is an asset, so building it does not reduce reported profit. The cost of holding it is scattered across warehouse rent, insurance, interest, write-offs and handling labour, none of it labelled inventory carrying cost, and several of those lines sitting inside the overhead bucket described above.

The scale is not small. According to the US Census Bureau’s Manufacturers’ Shipments, Inventories and Orders report released on 2 September 2026, US manufacturers held $966.9 billion in inventories in July 2026, a tenth consecutive monthly increase, with an inventories-to-shipments ratio of 1.47 against monthly shipments of $658.8 billion.

Applying a carrying rate to that stock is where the number becomes real. In its 2016 State of Logistics Report, the Council of Supply Chain Management Professionals disclosed that earlier editions had estimated inventory carrying cost by applying a fixed rate of 19 percent to total business inventory value, and its own ten-year table showed the carrying rate moving between 16 and 21 percent across 2006 to 2015. That methodology is a decade old and rate conditions have changed, so treat it as an order of magnitude rather than a current benchmark. Even at the low end, a rate in the high teens applied to a nine-figure inventory position belongs in a board pack. For wider context, CSCMP’s 37th Annual State of Logistics Report, produced by Kearney and released on 16 June 2026, put total US business logistics costs at $2.4 trillion, or 7.8 percent of national GDP, against $2.6 trillion and 8.7 percent in the previous edition.

The finance question is almost never asked in these terms: what is the annual cost of the decision to hold this much stock, and which commercial commitment is it buying us? If the answer is service level, price it. If the answer is that the forecast is unreliable, that is a planning problem being paid for out of the balance sheet. We covered the operational side of this in how GCCs are driving efficiency in logistics supply chains.

Cost of poor quality has four categories and most manufacturers formally track two

The American Society for Quality defines cost of quality as a methodology for determining the extent to which an organisation’s resources are used on activities that prevent poor quality, that appraise quality, and that result from internal and external failures. It sets out four categories.

• Prevention costs are incurred before operations begin, to stop quality problems arising. ASQ lists specification setting for incoming materials and finished products, quality and inspection planning, maintenance of the quality system, and training. These are the only quality costs most companies budget deliberately, and usually the first cut when margins tighten.

• Appraisal costs cover measuring and monitoring: checking incoming material, process setup and product against agreed specifications, quality audits, and supplier assessment and approval. Appraisal is visible because it has headcount attached, which is exactly why it gets challenged more often than the failure costs it exists to prevent.

• Internal failure costs are incurred fixing defects found before the customer receives the product. ASQ lists scrap that cannot be repaired or sold, rework and rectification, failure analysis, and waste from unnecessary work or stock held because of errors. Most plants track scrap in kilograms and rework in hours. Very few convert either into a single financial figure a CFO can act on.

• External failure costs are incurred after the customer has the product: repairs and servicing of returned and field units, warranty claims, complaint handling, and the handling, investigation and transport cost of returns and recalls. This is the category that lands across four departments and therefore appears whole in nobody’s report.

The evidence on whether finance functions have this under control is not encouraging. ASQ, citing its 2025 ASQE Insights on Excellence Cost of Quality Report, states that only 31 percent of respondents feel they fully understand the impact of quality costs on the financial performance of their organisations.

One honest note on benchmarking. You will find a wide range of figures circulating for cost of poor quality as a percentage of revenue, commonly 15 to 20 percent, sometimes 10 to 30 percent and occasionally as high as 40 percent, almost all attributed to ASQ. ASQ’s own cost of quality resource does not state any of those ranges. Before quoting a benchmark to your board, check whether the named source actually published it. A number you cannot trace is a number you cannot defend in a capital allocation discussion.

Where each hidden cost sits and what keeps it invisible

The four-step diagnostic a finance function can run in one quarter

This does not require a new system. It requires reorganising data you already hold. Great execution is thinking at macro level and executing at micro level, and this is the micro end.

1. Rebuild the overhead pool by cause rather than by general ledger account. Take twelve months of overhead and reassign every line to the activity that triggered it: setup and changeover, material handling, inspection, rework, maintenance, storage, expediting. Do not attempt a full activity based costing implementation. A directional reallocation done in a spreadsheet in three weeks answers the question that matters, which is where the money actually goes.

2. Price the inventory position at your real cost of capital. Take average inventory value by category, apply your weighted average cost of capital, and add storage, insurance, obsolescence provisions and shrinkage. Present the answer as one annual number per product family. The moment the carry cost of a slow-moving SKU family exceeds the gross margin it earns, the commercial conversation changes on its own.

3. Build a cost of poor quality ledger against the four ASQ categories. Assign every scrap ticket, rework hour, warranty claim, return, complaint handling cost and premium freight event to prevention, appraisal, internal failure or external failure. Report it monthly as a single figure with the four-way split. The split matters more than the total, because prevention and appraisal spend that is falling while failure cost is rising is the clearest early warning signal in a plant.

4. Re-run the make or buy and capacity decisions on the corrected numbers. Most make or buy analyses use fully absorbed standard cost, which overstates the cost of internal production when the plant has idle capacity and understates it when quality escapes are frequent. Rebuilding the numbers usually reverses at least one previous decision, and that reversal is where the payback on this exercise sits.

What changes once the numbers are right

The output is not a cost reduction programme. It is a decision-quality improvement, and the two look very different in practice.

Governance is not a side hustle. A costing system that cannot tell you which decision created a cost is a governance gap, not a reporting inconvenience, and it stays invisible until a fundraise, an acquisition diligence or a margin collapse forces the question. Founders and boards consistently assume growth is their responsibility and that someone else is building the processes underneath it. Nobody is.

Where this work needs outside help is rarely the analysis. It is holding the line when the rebuilt numbers contradict a decision leadership already made. That is the job of a strategic CFO as distinct from a bookkeeping or reporting mandate, a distinction we set out in strategic CFO versus virtual CFO versus bookkeeping, and it sits close to the trade-off territory in what the boardroom rarely admits.

The reporting then has to survive contact with a monthly close, which is a process problem rather than an analytical one. Consolidating the cost of poor quality ledger and the carry calculation into a repeatable close is usually a job for agile shared services, and we have written separately on accelerating the month-end close and on cost containment that does not kill growth.

The point, plainly

Every plant I have worked in had a number nobody owned. At Unilever and GE the discipline was drilled in early, that you think about the business at the macro level and then execute at the micro level, and the overhead bucket is where that discipline usually breaks. Somebody sets a rate, everybody applies it, and for the next five years the company reports variances against an assumption nobody has revisited.

If you take one thing from this, take the inventory carry calculation. It takes a finance analyst a week. It produces a single number per product family. And in my experience it starts a commercial conversation that no amount of cost cutting ever gets to, because for the first time the business is looking at what a decision costs rather than what a line item totals.

The costs are not hiding. They are recorded, audited and signed off. They are simply recorded in a shape that answers the auditor’s question instead of yours.

^[1][2][3]