Where the single rate quietly lies
In the earlier costing guides you learned to attach [[direct-vs-indirect-costs|indirect costs]] — factory overhead — to products with a single [[predetermined-overhead-rate|predetermined overhead rate]], computed before the year begins and applied through one [[allocation-base|allocation base]], usually direct labor hours or machine hours. That machinery is correct and it is not going away. But it carries a hidden assumption that is easy to miss: it assumes that every product consumes overhead *in proportion to* that one base. Double the labor hours, double the overhead charged. Activity-based costing is the response to a single, uncomfortable observation — that assumption is often simply false.
Think about what actually lives inside a modern factory's overhead. Machine setups when you switch from one product to another. Purchase orders and receiving inspections. Engineering changes. Quality tests. Material handling. Notice that almost none of these are driven by *how many labor hours* a unit takes — they are driven by *how many times you do a thing*. A setup costs the same whether the batch that follows is 10 units or 10,000. A purchase order costs the same whether it brings in a tonne of steel or a single rare bolt. Yet the single-rate method, blind to all of this, simply smears the whole overhead pool across products by labor hours, as if a setup-hungry specialty part and a long, untouched production run were the same kind of animal.
The four moves of activity-based costing
[[activity-based-costing|Activity-based costing]] (ABC) replaces the one big smear with a small set of more honest channels. Instead of asking "how do we spread overhead by labor hours?" it asks "what *activities* consume our resources, and which products demand those activities?" The logic flows in two stages: resources cost money, activities consume resources, and products consume activities. ABC simply refuses to skip the middle term. Here is the whole procedure in four moves.
- Identify the activities. Walk the process and list the things that actually consume overhead resources — machine setups, purchasing, inspections, material moves, design changes — rather than lumping everything into one bucket called 'overhead'.
- Pool the costs. Gather the overhead spent on each activity into its own activity cost pool — one pot of money for 'setups', another for 'purchasing', and so on.
- Pick a cost driver for each pool. Choose the measure that truly causes that pool to grow — number of setups, number of purchase orders, number of inspections — not a one-size-fits-all base.
- Compute a rate per driver and apply. Divide each pool by its driver's total volume to get a rate (cost per setup, cost per order), then charge each product by how many driver units it actually used.
The pivot of the whole method is the [[cost-driver|cost driver]] chosen in step three. A cost driver is the thing whose volume genuinely causes a cost pool to rise — and the art of ABC is choosing drivers that reflect *cause*, not mere correlation. Labor hours might correlate loosely with total overhead, but the number of setups *causes* setup cost. By tying each pool to its own cause-bearing driver, ABC stops pretending a single base can speak for a dozen unrelated cost behaviors. Notice this is still the same overhead application idea from before — rate times usage — just run many times, once per activity, instead of once for the whole plant.
A tiny factory, two products, two stories
Numbers make this vivid. A workshop makes two products: Standard, a long high-volume run, and Deluxe, a fiddly low-volume specialty. This year it makes 9,000 Standard and 1,000 Deluxe, and total overhead is 200,000. Both use one direct labor hour per unit, so total labor hours are 10,000 and the single plant-wide rate is 200,000 ÷ 10,000 = 20 of overhead per unit — flat, identical, for *both* products. Under the single rate, Standard and Deluxe each carry exactly 20 of overhead. Clean, simple, and — as we are about to see — quietly wrong.
Now look at what the overhead is really made of. Say it splits into two pools: 80,000 of setup cost and 120,000 of machine-running cost. The driver for setups is the *number of setups*; Standard runs in a few big batches and needs only 20 setups all year, while finicky Deluxe is made in many tiny batches and demands 180 setups. The driver for machine-running is *machine hours*, which both products use in proportion to volume — 9,000 hours for Standard, 1,000 for Deluxe. Setup cost is therefore 80,000 ÷ 200 setups = 400 per setup. Standard absorbs 20 × 400 = 8,000 of setup cost; Deluxe absorbs 180 × 400 = 72,000 — nine times as much, on a tenth of the volume.
Total overhead 200,000 -> Setups 80,000 + Machine-run 120,000
Pool Driver Rate Standard Deluxe
---------- ----------- ------------------ ------------ ------------
Setups # setups 80,000/200 = 400 20x400= 8,000 180x400= 72,000
Machine-run mach. hours 120,000/10,000 = 12 9,000x12=108,000 1,000x12= 12,000
------------ ------------
ABC overhead total 116,000 84,000
Units / 9,000 / 1,000
ABC overhead PER UNIT ~12.9 84.0
Single-rate overhead per unit (both): 200,000 / 10,000 hrs = 20.0Read the punchline carefully, because it is the heart of this guide. Under the single rate, Deluxe looked like it cost 20 of overhead per unit; under ABC, it costs 84. If the company priced Deluxe believing it cost 20, it may have been selling its fanciest, most demanding product *at a loss* the whole time — and making up the difference, unknowingly, on the fat margins of Standard. This is the classic and dangerous pattern: low-volume specialty products quietly consume far more overhead than their price assumes, and high-volume bread-and-butter products subsidize them. ABC is the method that drags this subsidy into the light.
When ABC is worth it — and when it is not
ABC is more accurate, but accuracy is not free. Identifying activities, building pools, and measuring drivers every period costs real time and money, and the system itself is overhead that someone must maintain. So the honest question is never "is ABC better?" — it almost always is, on paper — but "is the improvement in accuracy worth what it costs to obtain?" That trade-off tips toward ABC under specific conditions, and you can learn to spot them.
ABC earns its keep when (1) overhead is a *large* share of total cost — automation has replaced cheap labor, so the pool you are misallocating is huge; (2) the product mix is *diverse* in volume and complexity, so cross-subsidies between simple and fiddly products are large; and (3) the cost of getting it wrong is high — fierce competition where mispricing one product loses you the others. Flip every condition and ABC stops paying: a firm with trivial overhead, a near-identical product line, and comfortable margins gains almost nothing from the extra machinery, and the single rate is a perfectly sensible, cheaper choice. Sophistication you do not need is not virtue; it is waste.
What changed, and what to carry forward
Step back and see the shape of the idea. The single predetermined rate you met earlier asks one question of every product — "how many labor hours?" — and answers all costing with that one lens. ABC asks several questions — "how many setups? how many orders? how many inspections?" — because it accepts that overhead is not one thing with one cause but a crowd of different costs with different causes. Each cost object is then charged for the specific activities it genuinely triggered. The single rate is ABC collapsed to exactly one activity pool and one driver; ABC is the single rate refined to as many pools as the decision is worth.
One last caution against over-claiming. ABC gives a *more refined estimate*, not a perfect truth. The activities you choose, the drivers you pick, and how you trace resources into pools all involve judgment, and two careful accountants can build two defensibly different ABC systems for the same factory. The honest claim is modest but real: by tracing overhead to its causes, ABC usually turns a *known* distortion into a *smaller* one. That is enough to flip which products you call profitable, which prices you defend, and which lines you would have wrongly axed — and in the next guide on absorption versus variable costing, you will see yet another way the same overhead can be made to tell different stories.