Product costing: how overhead reaches the product
Materials and labour are easy to trace to a product. Factory overhead is not, and what a business does with it changes the cost it reports for every unit, the value of its inventory and, in some years, its profit. This lesson follows overhead into a bicycle three ways: spread over production with absorption costing, charged to the year with variable costing, and traced to activities with activity-based costing. It ends with a check on how much the choice could matter at Apple and Microsoft.
- Level
- Beginner
- Reading time
- About 24 minutes
- Know this first
- Operating leverage
By the end you will be able to:
- calculate a predetermined overhead rate and the cost of a unit under absorption and under variable costing
- explain why the two methods report different income, and calculate the difference
- calculate activity rates and show how one overhead rate can misstate the cost of different products
- say which method is used for external reporting and which for internal decisions
Putting a cost on a product
Product costing is the work of assigning the costs of production to the units made. The direct costs are easy: materials and labour can be traced to a bicycle. Manufacturing overhead is the difficulty. It covers the rent, supervision, power and depreciation of the factory, which serve every bicycle and belong to none in particular. How overhead is assigned decides what each unit appears to cost, what the inventory on the balance sheet is worth, and which products look profitable.
Two systems collect the costs in the first place. Job order costing traces costs to each individual job or order, and suits custom work such as signs, buildings or print jobs. Process costing accumulates costs by production process or department and divides them over the units that pass through, which suits continuous, mass-produced goods such as paint, cereal or gasoline. Both then face the same question about overhead, and the rest of this lesson is about the answers. The running example is the fictional bicycle maker from the earlier lessons; a second invented plant appears in the section on activity-based costing, and Apple and Microsoft appear near the end.
The overhead rate
The usual way to put overhead on a unit is a rate set before the year begins. The predetermined overhead rate is the estimated overhead for the year divided by the estimated activity, measured on a base such as direct labour hours, machine hours or units. It is set in advance because waiting for the year's actual figures would delay every product cost by months, and because actual overhead is lumpy. Overhead is then applied to each unit as it is made. If the amount applied by the end of the year is below what was actually spent, overhead is underapplied; if above, overapplied, and the difference is cleared against the cost of goods sold.
To keep the arithmetic visible, this lesson uses units made as the activity base and divides by the units actually made. The bicycle maker's factory overhead is $150,000 a year. Spread over 8,000 bicycles it is $18.75 each, so a bicycle that cost $90 in materials and $30 in labour is carried at $138.75 once overhead is included. That per-unit figure depends on production: make fewer bicycles and each one carries more of the overhead.
Two ways to treat fixed overhead
Absorption costing, also called full costing, treats all manufacturing costs as product costs: direct materials, direct labour, and both variable and fixed factory overhead. The fixed overhead sits in inventory with the bicycle and is expensed only when the bicycle is sold. Variable costing, also called direct or marginal costing, treats only the variable manufacturing costs as product costs. Fixed factory overhead is charged in full to the year in which it is incurred, whether or not the bicycles were sold.
Neither method changes the cash spent. They change when the fixed overhead reaches the income statement. For external reporting, absorption costing is the required method under US GAAP, and it is also used for tax in the United States and many other countries. Variable costing is not acceptable for GAAP reporting, but managers use it inside the business, because it keeps fixed and variable costs apart in the way the break-even and leverage lessons needed.
| Cost of one bicycle | Absorption costing | Variable costing |
|---|---|---|
| Direct materials | $90.00 | $90.00 |
| Direct labour | $30.00 | $30.00 |
| Factory overhead ($150,000 ÷ 8,000) | $18.75 | charged to the year |
| Cost held in inventory | $138.75 | $120.00 |
In both methods, the sales commission of $5 a bicycle, the head-office salaries of $50,000 and the advertising of $40,000 are period costs charged to the year. Only the treatment of factory overhead differs.
Make more than you sell
Nothing is in stock at the start of the year. Choose how many bicycles the plant makes, how many it sells (sales cannot exceed what was made) and the factory overhead. The left page shows the income statement under absorption costing, the right page under variable costing.
The plant makes 8,000 bicycles and sells 6,500, leaving 1,500 in stock. Absorption costing spreads the $150,000 of factory overhead over the 8,000 made, so each bicycle carries $18.75 of it, and reports an income of $275,625. Variable costing charges all $150,000 to the year and reports $247,500. The gap of $28,125 is the overhead that absorption costing holds in the 1,500 unsold bicycles, 1,500 × $18.75.
Four experiments. First, at the starting settings the plant makes 8,000 bicycles and sells 6,500. Absorption costing reports $275,625 and variable costing $247,500, a gap of $28,125, which is 1,500 unsold bicycles carrying $18.75 of overhead each. The stock is worth $208,125 under absorption costing and $180,000 under variable costing, and the difference between those is the same $28,125. Second, move sales up to 8,000: nothing is left in stock, both methods report $360,000, which is the profit from the earlier lessons, and the gap disappears. Third, go back to 6,500 sold and raise production to 10,000. Variable costing still reports $247,500, since it does not depend on production, but absorption costing now reports $300,000. The plant has made no more sales, and absorption costing shows a higher profit only because more of its overhead has been parked in unsold bicycles ($52,500 more than variable costing). Fourth, set both production and sales to 6,500 and the gap vanishes again. When a business sells all it makes, the two methods agree.
Over time the difference evens out
The gap is a timing effect. Overhead deferred into inventory in one year is expensed in the year the stock is sold. The rule is exact: absorption income equals variable income plus the fixed overhead in ending inventory less the fixed overhead in beginning inventory. Production stays at 8,000 bicycles for three years while sales vary.
| Three years of the bicycle maker | Made | Sold | In stock at year end | Income: absorption | Income: variable | Overhead added to stock |
|---|---|---|---|---|---|---|
| Year 1 | 8,000 | 8,000 | 0 | $360,000 | $360,000 | +$0 |
| Year 2 | 8,000 | 6,500 | 1,500 | $275,625 | $247,500 | +$28,125 |
| Year 3 | 8,000 | 9,500 | 0 | $444,375 | $472,500 | −$28,125 |
| Three years | 24,000 | 24,000 | $1,080,000 | $1,080,000 | $0 |
In year 2 the stock grows by 1,500 bicycles and absorption costing reports $28,125 more than variable costing. In year 3 the business sells 9,500, using up that stock, and absorption costing reports $28,125 less. Over the three years both methods add up to $1,080,000, because all 24,000 bicycles made were sold. Absorption costing therefore smooths profit when sales swing and production stays steady, while variable costing makes profit follow sales. It also gives managers a reason to be careful: a plant manager rewarded on absorption-costed profit can raise it for a year by producing more than is needed, as the third experiment showed.
When one overhead rate is too crude
A single rate based on labour hours or units treats every product as if it used overhead in proportion to that base. That works when labour dominates production and the products are alike. It breaks down when overhead is large and products use it differently. Activity-based costing (ABC) groups overhead into cost pools by activity, finds the cost driver that causes each pool to grow, and sets an activity rate for each driver, which is the pool's cost divided by the driver's expected quantity. Products then pick up overhead according to how much of each activity they use.
To see the difference, take a second fictional plant that makes 9,000 standard bicycles and 1,000 custom bicycles a year, with $300,000 of overhead in three pools.
| Overhead pool | Cost | Driver | Standard model | Custom model | Rate |
|---|---|---|---|---|---|
| Machine setups | $120,000 | setups | 20 | 100 | $1,000 a setup |
| Machine running time | $100,000 | machine hours | 18,000 | 2,000 | $5.00 an hour |
| Inspection | $80,000 | inspections | 100 | 300 | $200 an inspection |
The standard model needs 2 labour hours and the custom model 3, so a single rate on direct labour hours would be $300,000 ÷ 21,000 hours = $14.29 an hour. The custom bicycle, which is made rarely and set up often, needs far more setups and inspections than the standard one. The table compares the two approaches.
| Overhead for each bicycle | Standard model | Custom model |
|---|---|---|
| Direct materials and labour | $120.00 | $195.00 |
| Overhead by direct labour hours ($14.29 an hour) | $28.57 | $42.86 |
| Overhead by activity | $14.44 | $170.00 |
| Selling price | $200 | $400 |
| Profit for each bicycle, by hours | $51.43 | $162.14 |
| Profit for each bicycle, by activity | $65.56 | $35.00 |
The labour-hours rate gives the custom model an overhead of $42.86 and the standard model $28.57. Activity rates give $170.00 and $14.44, because the custom model uses 100 of the 120 setups and 300 of the 400 inspections. Under the single rate the custom model looks highly profitable, at $162.14 a bicycle; under activity rates it earns only $35.00, and the standard model, which looked less profitable, earns $65.56. The total overhead of $300,000 and the total profit of $625,000 are identical both ways. ABC only changes which product is charged, and with it the decisions on price, mix and which models to push.
ABC is more work. Each driver needs its own measurement, so it is worth using where overhead is a large share of cost, processes are complex, and products differ in how they use activities. It is used outside factories too, in services where overhead dominates. Where overhead is small or products are similar, a single rate gives much the same answer for far less effort.
Does the choice matter at Apple and Microsoft?
The effect of absorption versus variable costing on income is limited by the amount of fixed overhead sitting in inventory, and that cannot exceed inventory itself. Companies do not report how much of their inventory is fixed overhead, but the inventory balances put a ceiling on the effect. The change in income between the methods is the change in fixed overhead held in stock, so it cannot be more than the larger of the opening and closing inventory, even if every dollar of inventory were fixed factory overhead, which is far from the case since inventory also holds materials and bought-in goods.
| Inventory, $ million | Apple | Microsoft |
|---|---|---|
| Inventories, end of fiscal 2024 | 7,286 | 1,246 |
| Inventories, end of fiscal 2025 | 5,718 | 938 |
| Cost of sales, fiscal 2025 | 220,960 | 87,831 |
| Inventory in days of cost of sales | 9.4 | 3.9 |
| Operating income, fiscal 2025 | 133,050 | 128,528 |
| Larger inventory balance as a share of operating income | 5.5% | 1.0% |
Apple's larger balance, $7,286 million, is 5.5% of its fiscal 2025 operating income, and at the end of that year its inventory of $5,718 million was about 9.4 days of cost of sales. Microsoft's larger balance, $1,246 million, is 1.0% of operating income, and its year-end inventory of $938 million was about 3.9 days. These percentages are ceilings on the effect, and the true figures are lower. For companies whose stock turns over in days, the question of how fixed overhead is treated is of little consequence for profit. It matters most for a business that holds large stocks for long periods, such as a maker of aircraft or wine, or that builds up stock ahead of sales, as the bicycle maker did.
The same table shows why the earlier lessons were framed in terms of cost behaviour. Neither company's published statements split costs into fixed and variable, and since external statements under US GAAP use absorption costing, an outside analyst cannot rebuild a variable-costing profit from them. What can be done is what the earlier lessons did: estimate behaviour from several periods of results, and treat the answers as estimates.
What these methods cannot tell you
- The cost of a unit depends on the method and on volume. There is no single true cost per unit once fixed overhead is included, only a figure for a chosen method and production level.
- Allocation is a convention. Fixed overhead does not change with the number of bicycles made, so dividing it among them does not show what one more bicycle costs.
- The activity base matters. Labour hours suit labour-intensive work. As machinery replaces labour, a rate based on hours becomes less representative.
- ABC needs good data. The drivers have to be measured, and the pools chosen with judgement. Too many pools make the system costly, too few bring back the problem it was meant to solve.
- Inventory values are not market values. Under absorption costing, unsold stock carries overhead at cost, and only a sale shows whether it can be recovered.
Common mistakes
Thinking variable costing saves money
It does not change what is spent. It changes which year bears the fixed overhead.
Reading absorption profit as a sign of sales
Income can rise because production rose. Compare it with sales, and look at the change in inventory.
Treating a unit cost as a decision cost
The absorption cost of $138.75 includes $18.75 of overhead that would be incurred anyway. For deciding whether to take an extra order, the relevant cost is the one that changes, which the lesson on decisions returns to.
Using the wrong base for overhead
A rate based on labour hours will charge a product for overhead it does not cause if the overhead is driven by something else, such as setups.
Assuming ABC changes total cost
It redistributes the same overhead between products. Total cost and total profit stay the same.
Try it yourself
1. A factory makes 5,000 units with direct materials of $8, direct labour of $10 and variable overhead of $2 a unit. Fixed overhead is $40,000. What is the cost of a unit under variable costing and under absorption costing?
Show the working
Variable costing: 8 + 10 + 2 = $20. Absorption costing adds fixed overhead per unit: 40,000 ÷ 5,000 = $8, so the cost is 20 + 8 = $28.
2. The factory sells only 4,000 of the 5,000 units, with nothing in stock at the start. How do the two methods compare?
Show the working
1,000 units remain in stock, each carrying $8 of fixed overhead under absorption costing, so $8,000 is deferred. Absorption income is $8,000 higher than variable income. Reconciliation: absorption income = variable income + 8,000 (ending) − 0 (beginning).
3. The next year it makes 4,000 units and sells 5,000, using up the 1,000 units in stock. Which method shows the higher income, and by how much?
Show the working
The stock carried $8,000 of overhead from last year, released into cost of goods sold this year, and no stock is left at the end. Absorption income = variable income + 0 − 8,000, so variable costing is higher by $8,000. Across both years the two methods add up to the same total.
4. An overhead pool of $60,000 is driven by setups, and 1,200 setups are expected. What is the activity rate, and how much overhead does a product needing 8 setups receive?
Show the working
Activity rate = 60,000 ÷ 1,200 = $50 a setup. A product with 8 setups receives 8 × $50 = $400, however many units it contains.
5. Why can a manager raise profit under absorption costing by producing more, and why is that not a real gain?
Show the working
More units share the same fixed overhead, so each unit carries less of it and more of the year's overhead is parked in unsold stock instead of being expensed. Reported profit rises, but the cash is spent, the stock has still to be sold, and the overhead will reach the income statement later.
Where this is used
- A factory sets its overhead rate before the year starts, so that every job can be costed as it is completed.
- An accountant values closing inventory under absorption costing for the published statements and the tax return.
- A manager uses variable-costing reports for pricing and break-even work, because they separate the costs that change with volume.
- A business with many varied products reviews its pricing after an activity-based study shows which products use the most support activity.
Key terms
- Manufacturing overhead
- Factory costs other than direct materials and direct labour, such as rent, supervision, power and depreciation.
- Job order costing
- A system that traces costs to each individual job or order.
- Process costing
- A system that accumulates costs by production process or department and spreads them over the units produced.
- Predetermined overhead rate
- Estimated overhead divided by an estimated activity base, set before the period begins.
- Absorption costing
- Treating all manufacturing costs, fixed overhead included, as product costs. Also called full costing.
- Variable costing
- Treating only variable manufacturing costs as product costs and charging fixed overhead to the period. Also called direct or marginal costing.
- Activity-based costing
- Assigning overhead through cost pools and activity rates, according to the activities each product uses.
- Cost driver
- The activity that causes the cost of a pool to rise, such as setups or inspections.
Questions people ask
Which method do companies use in their published accounts?
Absorption costing, which US GAAP requires for external statements. Variable costing is a management tool and is not acceptable for GAAP reporting.
Why does the profit differ if the same cash is spent?
Because the methods disagree about when fixed overhead becomes an expense: in the year it is spent, or in the year the goods it helped to make are sold.
Is activity-based costing always better?
It is more accurate where overhead is large and products use it differently, and it costs more to run. Where overhead is small or products are alike, a single rate is usually enough.
Does the choice change how much tax is paid?
The choice used for tax follows the rules of each country. Inside a business, the management method does not change the tax bill, only the internal reports.
Sources and review
- Apple Inc., Form 8-K, Exhibit 99.1: fourth quarter and full-year fiscal 2025 results, filed with the US Securities and Exchange Commission. Source of Apple's statements.
- Microsoft Corporation, Form 8-K, Exhibit 99.1: fourth quarter and fiscal year 2025 results, filed with the US Securities and Exchange Commission. Source of Microsoft's statements.
- OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 6.5: Compare and contrast variable and absorption costing, for the definitions, the reconciliation between the two methods, and their use in external reporting.
- OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 6.1: Calculate predetermined overhead and total cost under the traditional allocation method, for the predetermined overhead rate and the limits of a single rate.
- OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 6.3: Calculate activity-based product costs, for cost pools, cost drivers and activity rates.
- OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 4.1: Distinguish between job order costing and process costing, for the two costing systems.
How the numbers were checked: the unit costs, both income statements and the inventory values were recomputed in plain code, the reconciliation between the two methods was tested in each of the three years, and the three-year totals and the activity-based tables were rebuilt from the underlying volumes and pool costs. The calculator, which runs separately, was compared with that code across a grid of inputs. Inventory and operating income for Apple and Microsoft come from the filed results.