Cost Apportionment: The Allocation That Decides Everything After It
Get the apportionment wrong and every number after it is confidently incorrect.
Summary
- Cost apportionment is the choice of basis for spreading costs that cannot be traced to a single product, service line, specialty or channel. Every margin, floor and investment decision after it inherits that choice.
- Revenue share is the default basis almost everywhere and wrong almost everywhere, because revenue does not consume rent, nursing or engineering. Apportion on whatever actually causes the cost to be consumed.
- Allocate directly everything you honestly can first, keep to five or six drivers, estimate where the data does not exist, and write every basis down.
- Costs with no honest driver, such as the founder's salary or group legal, are better reported as a block below contribution than spread across lines.
A mid-sized hospital we worked with was reviewing which specialties to invest in.
Oncology billed ₹22 crore and looked like the strongest department. Orthopaedics billed ₹9 crore and looked marginal. So the plan was to add oncology capacity and leave orthopaedics alone.
Then we asked how the shared costs were being split.
Nursing, housekeeping, biomedical maintenance, administration and the share of building cost, around ₹14 crore in total, were apportioned across specialties by revenue. Oncology billed the most, so oncology carried the most.
But orthopaedics was consuming roughly 40% of operating theatre minutes against 12% of revenue. Oncology day-care cases were using a fraction of the nursing ratio of an orthopaedic inpatient.
Re-apportion nursing on bed-days weighted by acuity, and theatre cost on OT minutes, and cost moves from oncology to orthopaedics. Oncology had been carrying far more shared cost than it used, so it was stronger than the report showed. Orthopaedics had been carrying far less, so the department that looked marginal was in fact being carried by the rest of the hospital. The plan to leave it alone rested on a margin it did not have.
Nothing about the hospital had changed. Only the apportionment. The full set of hospital drivers, from ward nursing to billing and claims, is laid out in costing hospitals and diagnostics.
Why this is the most consequential costing decision you make
Because everything downstream inherits it.
Your unit cost. Your product, service line or specialty margin. Your client and segment profitability. Your contribution margin floor. Your pricing. Your decision about where to invest and what to stop.
All of it sits on one choice: how you spread the costs that cannot be traced to a single thing. Get that wrong and every number afterwards is wrong in the same direction, with no warning on the report that it has happened.
In most businesses this choice is made once, by default, by whoever set up the chart of accounts.
Allocate what you can, apportion only what you must
Two words that get used interchangeably and shouldn't.
Allocation is when a cost belongs entirely to one thing. The engineer who only works on one product module. The nurse assigned solely to the ICU. The warehouse rented only for one channel. No judgement needed.
Apportionment is when a cost is shared and you have to choose a basis. Office rent. The CTO. The plant manager. Central customer support.
Allocate everything you honestly can first. Every rupee you trace directly is a rupee nobody can argue about. Most businesses apportion far more than they need to, because tracing feels like effort and apportioning feels like a decision already made.
Apportion on what consumes the cost, not on revenue
One rule: apportion on whatever actually causes the cost to be consumed.
The default in almost every business is revenue share. It is easy, it feels fair, and it is wrong nearly everywhere, because revenue does not consume rent, floor space does; revenue does not consume engineering, feature work does.
What basis to apportion each shared cost on, by business
| Business | Shared cost usually mis-apportioned | Basis that reflects consumption |
|---|---|---|
| SaaS | Engineering, infrastructure, customer success | Story points or sprint capacity by module; compute and storage by account; CS hours logged |
| IT / tech services | Delivery management, pre-sales, recruitment, L&D | Project hours; pursuit count; hires per practice; billable headcount |
| Hospitals | Nursing, OT, housekeeping, biomedical, admin | Bed-days weighted by acuity; OT minutes; procedure count; case complexity |
| Diagnostics | Equipment, lab staff, logistics, IT | Test volume weighted by run time; sample pickups; reporting volume |
| D2C / ecommerce | Warehousing, customer service, performance marketing | Cubic volume stored; tickets per channel; attributable ad spend by SKU |
| Manufacturing | Rent, power, supervision, maintenance | Floor space; machine hours; direct labour hours; maintenance tickets |
| Pharma | QA, regulatory, plant overhead | Batches; filings by market; campaign days |
| Semiconductors, robotics | R&D, test, applications engineering | Programme effort; test hours; design wins supported |
The harder your business is to see, the more this matters
A factory can at least count machine hours. Somebody wrote them down.
Nobody counts engineering effort by module unless they decide to. Nobody counts customer success hours by account unless somebody sets up the tracking. Nobody counts nursing time by specialty unless the hospital chooses to.
Which produces an uncomfortable pattern: the businesses where apportionment matters most are the ones least likely to be doing it. That is because the data does not fall out of an operational system the way machine hours do.
That is not a reason to skip it. It is a reason to estimate. A sprint-capacity split agreed between engineering and finance, reviewed quarterly, is enormously better than apportioning engineering by revenue; costing a SaaS business sets out how to run one.
The test to run on your own numbers
An afternoon, and it tells you whether you have a problem.
- List your shared costs. Everything in the P&L not directly traceable to a product, service line, specialty or channel.
- Write down the basis you use today. Most businesses find it is the same basis for everything.
- For each, write down what actually drives it. If this cost doubled, what would have caused that? The answer is your driver.
- Re-run your profitability on the better bases.
- Compare the rankings.
If the order changes, if the line you thought was weakest moves up or your star moves down, you have been running the business on the wrong information. And now you know it.
In our experience the ranking changes in a clear majority of businesses that do this for the first time.
Three rules that keep apportionment practical
- Roughly right beats precisely wrong. You do not need a time-and-motion study. If one module consumes most of the engineering, splitting 70:30 on an informed estimate beats 30:70 on revenue by a long way.
- Five or six drivers, not thirty. Activity-based costing in its full form is excellent in theory and defeats most businesses in practice. A handful of sensible drivers gets you most of the benefit. Thirty gets you abandoned in month three.
- Write the basis down. One line per cost: what basis, and why. It makes the number defensible when challenged, and it reminds you next year that a judgement was made.
Undocumented assumptions calcify into facts.
What to do with the costs that genuinely won't apportion
Some will not. The founder's salary. Group legal. Brand spend that benefits everything and nothing specifically.
Two options, and the second is usually better.
Apportion on something arbitrary and accept it. Revenue is fine here, because no better driver exists.
Or leave them out of line-level costing entirely and report them as a block below contribution, between contribution and net margin. This is cleaner. Your line margins then reflect costs those lines actually cause, and corporate costs sit visibly where they belong instead of being smeared across in a way nobody can act on.
For pricing decisions in particular, the second is far more useful. You cannot manage a cost by apportioning it. You can only manage it where it is incurred.
What changes once the basis reflects consumption
What you act on changes, sometimes sharply, as in the hospital above, where the department everyone meant to leave alone was the one that needed attention.
Your floor becomes real. A margin floor built on a bad apportionment is a confident number with no foundation.
And the conversation changes. When shared cost is apportioned on a driver people recognise, such as OT minutes, sprint capacity or cubic volume, the person running that line can actually influence their share. Apportion on revenue and there is nothing anyone can do except sell less.
A cost apportioned on a basis somebody controls becomes a cost somebody manages.
The question for your next review: what basis are your shared costs apportioned on, and what actually consumes them? If those two answers differ, start there. The Pricing Maturity Assessment shows where your cost floor sits among the other layers of your pricing.
Frequently asked questions
What is the difference between cost allocation and cost apportionment?
Allocation assigns a cost that belongs wholly to one product, service line, specialty or channel, so no judgement is involved. Apportionment spreads a shared cost, such as rent or a central support team, across several of them, and requires a choice of basis. Allocate first; apportion only what is left.
Which overhead allocation method should a smaller business use?
Whichever basis reflects what consumes each cost: floor space for rent, machine hours for maintenance, tickets for support, OT minutes for theatre. Use revenue only for costs with no better driver, and keep the total to five or six drivers so the method survives past the first quarter.
Is full activity-based costing worth it for an MSME?
Rarely in its full form. A handful of drivers chosen for the largest shared costs captures most of the benefit. An estimate agreed between operations and finance is better than a precise system nobody maintains.
Where should costs with no sensible driver go?
Below contribution, reported as a single block of corporate cost. That keeps line margins limited to costs each line actually causes, which is what a pricing decision needs.
How useful was this article?
One tap. It tells us what to write more of.
About the author
SRF Capital Studio
The next one
Get what we publish next, by email.
Working notes on raising, borrowing, protecting, growing and structuring capital in India. One email a week at most, and you can leave any time.
We use your address only to send this. See our privacy policy.
We store your address to send you these emails and nothing else. See our privacy policy.
Related reading
