Why Your Unit Cost Moves With Volume, and Why Cost-Plus Eats Itself
The most dangerous pricing method is the one that feels the most rigorous.
Summary
- Fixed cost per unit depends on the volume you divide by, so when utilisation falls, cost per unit rises without a rupee of spending moving.
- Cost-plus pricing then asks for a higher price in exactly the quarter demand has softened, which loses more volume and raises unit cost again: the loop closes.
- The way out is contribution. When volume falls, ask whether the work still covers its variable cost, not what full cost says the price should be.
- Absorption is for statutory reporting and for testing whether the business works over a year. It should never be the number that prices a specific piece of work.
A services firm has forty engineers. Loaded cost across the team is ₹9 crore a year.
At 70% utilisation they bill 56,000 of their 80,000 available hours (forty engineers at 2,000 hours each). Cost per billable hour: ₹1,607, which is ₹9 crore ÷ 56,000. Add a margin, price at ₹2,400.
At 50% utilisation they bill 40,000 hours. It is a soft quarter: two projects delayed, one client pausing. The same ₹9 crore now spreads over fewer hours. Cost per billable hour: ₹2,250 (₹9 crore ÷ 40,000).
Nothing has changed. Same people, same skills, same work.
But cost-plus now says the rate should rise by 40%, to ₹3,360. The new cost is 1.4 times the old one (₹2,250 ÷ ₹1,607), and the same mark-up takes ₹2,400 to ₹3,360. In a quarter where demand has already softened.
And what happens if they raise the rate 40% into a weak market? Fewer wins. falls further. Cost per hour rises again. The method calls for another increase.
That is the loop, and it closes.
It is not a manufacturing problem
Factories see it most clearly because the fixed cost is a building. But the mechanism is identical anywhere fixed capacity exists, which is everywhere.
Where the fixed capacity sits in seven kinds of business, and what falling volume does to unit cost
| Business | The fixed capacity | What happens when volume falls |
|---|---|---|
| IT / tech services | Payroll of a bench-able team | Cost per billable hour rises with falling utilisation |
| SaaS | Engineering team, committed infra, support function | Cost per account rises sharply at low account counts |
| Hospitals | Beds, theatres, consultants on retainer | Cost per bed-day at 55% occupancy is far above 80% |
| Diagnostics | Analysers, lab staff, collection network | Cost per test on an underused analyser is brutal |
| D2C | Warehouse, fixed logistics contracts, brand team | Cost per order rises as volume thins |
| Manufacturing | Plant, machines, supervision | The classic absorption case |
| Semiconductors, robotics | Development spend amortised over a volume forecast | NRE per unit explodes if the volume never arrives |
Take the hospital row. If all of its cost were fixed, a hospital running at 55% occupancy would carry a cost per bed-day about 45% higher than the same hospital at 80%. The same cost is spread over 55 occupied beds in every hundred instead of 80, and 80 ÷ 55 = 1.45. Consumables and drugs do move with patients, so the real gap is somewhat smaller, but the direction holds. If the hospital prices packages on full absorbed cost, it will price itself out of exactly the volume it needs.
Why it stays invisible
Because every step looks defensible.
"The cost went up": arithmetically true on an absorption basis. "We can't sell below cost": sounds like discipline. "We must protect margin": nobody argues with that in a meeting.
Each statement is reasonable. Together they describe a business pricing itself out of a market because of an convention.
The convention: fixed cost per unit depends entirely on how many units you divide by. Change the denominator and the cost changes, without a rupee of actual spending moving.
How a shared cost is spread across products or hours in the first place is the subject of cost apportionment. The trouble here is what happens to the answer when the count underneath it moves.
Budgeted versus actual, and where the surprise comes from
Most businesses set a rate at the start of the year on a budgeted volume. Budget 56,000 hours and you absorb at ₹1,607.
Then you bill 40,000. You have absorbed ₹6.43 crore into delivery (40,000 × ₹1,607) and the actual cost was ₹9 crore. The gap of about ₹2.6 crore is under-recovery, and it appears as a lump nobody planned for.
The reverse happens too. Beat the budget and your delivery costs look better than they are, flattering margins until the correction.
Two rules follow.
- Set the rate on normal capacity, not optimistic budget. A rate built on a volume you hope for guarantees under-recovery.
- Track it monthly, not annually. Under-recovery found in month eleven is a surprise. Found in month three it is a decision. Telling a volume gap from an efficiency gap is part of reading planned against actual.
The way out is contribution, not a higher price
The escape is contribution margin, which keeps the two kinds of cost apart:
Price − variable cost = contribution
Variable cost per unit does not move with volume. For a services firm the variable cost of an hour already being paid for is close to nil. For a hospital, an extra case on an existing bed consumes consumables and drugs, not the building. For a SaaS account, it is incremental compute and support.
So when volume falls, the question is not what cost-plus says the price should be. It is: does this still contribute? If yes, taking it is better than not taking it, because the fixed cost exists either way.
This is why the floor is the only floor worth having. Full cost is a reporting number. Contribution is a decision number.
One necessary caution. Contribution logic says take the marginal work. It does not say take all work at marginal prices forever. If every deal is priced at contribution, nothing covers the fixed cost and you go out of business profitably. Use it for the marginal decision at the margin, not as a pricing policy. A rate conceded to fill a quiet quarter also tends to become next quarter's expectation, which is the anchoring risk a services rate card runs into.
When absorption is right
Two things.
- Statutory reporting and inventory or WIP valuation. Required, not optional.
- Checking whether the business as a whole works. Over a year, at normal volume, total contribution must exceed total fixed cost. Absorption is one way of testing that.
What it is bad for is a specific decision about a specific piece of work at a specific price, because it loads a cost that will be incurred regardless.
The rule: absorption for reporting, contribution for decisions.
Trouble starts when the reporting number walks into a pricing meeting.
Four habits that keep volume out of the price
- Know both numbers and label them. Variable cost per unit, clearly marked. Fixed cost per period, clearly separate. Never one blended figure that moves silently with volume. Building the unit cost is its own exercise; the point here is to keep its two halves apart.
- Price from the market, not from the cost. Cost gives you the floor. What the buyer values gives you the ceiling. Only one of those comes from your accounts.
- When volume falls, do the opposite of what cost-plus suggests. Lower volume usually means a harder market, which is the worst possible moment to raise prices. If contribution is still healthy, defend volume.
- Report under-recovery separately. Not hidden in unit cost. A line that says: this is the capacity we paid for and did not use. That is a capacity conversation, not a pricing one, and it leads somewhere useful, like whether to resize, redeploy or sell harder.
Why cost-plus survives, and the check that exposes it
Cost-plus survives because it feels rigorous and because it transfers responsibility. If the price came from a calculation, nobody has to defend a judgement.
But the calculation contains a judgement, the volume you divided by, and that judgement is doing all the work while looking like arithmetic.
The check to run this week: recalculate your unit cost at 70% and at 130% of current volume or utilisation. If the two numbers are far apart, and they will be, then any pricing decision you made on full cost was a decision about volume rather than about cost. Where cost is entirely fixed, the first is 1.43 times today's figure (1 ÷ 0.7) and the second 0.77 times (1 ÷ 1.3).
Our Pricing Maturity Assessment gives you a short read across your pricing and the cost floor underneath it, including whether full cost is quietly setting prices it should not.
Frequently asked questions
Why is cost-plus pricing a problem when volume falls?
Because the cost it adds a margin to is not a fixed fact. Full unit cost includes fixed cost divided by volume, so a fall in volume raises it. Cost-plus turns that rise into a higher price at the moment demand is weakest. The higher price loses more volume, which raises unit cost again.
What is under-recovery of fixed cost?
The part of fixed cost a business planned to absorb into its products or hours and did not. It happens because actual volume fell short of the volume the rate was set on. In the example above it is about ₹2.6 crore: ₹9 crore of cost against ₹6.43 crore absorbed at ₹1,607 an hour over 40,000 hours.
Should a business ever price below full cost?
For a specific piece of marginal work, yes, if it covers its variable cost and the capacity would otherwise sit idle. As a policy, no. Over a year, at normal volume, total contribution has to exceed total fixed cost, or every deal can look acceptable while the business as a whole loses money.
When should a business use absorption costing?
For statutory accounts and for valuing inventory or work in progress, where it is required. And use it as a yearly test of whether the business covers its fixed cost at normal volume. It is the wrong basis for pricing an individual job, because it charges that job with cost that will be incurred whether or not the job is taken.
How useful was this article?
One tap. It tells us what to write more of.
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
