Cost Structure Analysis 2026: Same Profit, Two Shapes

Three cost structure splits, and what each one is for

There are three useful ways to cut a cost base and they answer three different questions. Choosing the wrong one is the most common way this analysis stalls.

1. Fixed against variable. Answers what happens when volume moves. This is the split that predicts.
2. Direct against indirect. Answers which costs belong to a product, site or customer. This is the split that decides whether something is worth keeping.
3. Controllable against committed. Answers what a manager can actually change this year. Leases, depreciation and long contracts are committed even when they are variable in principle.

Deloitte's published five-step approach asks for the underlying assumptions to be identified before the analysis runs, checked 30 September 2026 at deloitte.com. On a cost base the assumption that carries everything is which lines move with volume, and it is worth stating out loud because it is rarely in the data pack.

A practical test for the first split: ask what happened to each line in the last downturn the company had. Lines that fell with volume are variable, whatever the chart of accounts calls them.

Two bakeries, same revenue, same profit

A regional group owns two industrial bakeries. Both turn over 40 million dollars a year and both earn 6 million dollars of operating profit, so both report a 15 percent margin. Their cost bases are built completely differently.

Bakery A runs modern automated lines. Variable cost is 60 percent of revenue, which is 24 million dollars, and fixed cost is 10 million. Contribution is 40 minus 24, which is 16 million, so contribution margin is 40 percent, and profit is 16 minus 10, which is 6 million.

Bakery B outsources most of its baking and runs a small finishing plant. Variable cost is 80 percent of revenue, which is 32 million dollars, and fixed cost is 2 million. Contribution is 40 minus 32, which is 8 million, so contribution margin is 20 percent, and profit is 8 minus 2, which is 6 million.

Every summary line a chief executive normally looks at is identical. Revenue, profit and margin match to the dollar, and the two businesses are not remotely alike.

Operating leverage predicts the fall exactly

Degree of operating leverage is contribution divided by operating profit. For A that is 16 divided by 6, which is 2.67. For B it is 8 divided by 6, which is 1.33.

That ratio is a multiplier on percentage changes in revenue, and it works in both directions. Test it.

Bakery A at 36 million: variable cost is 60 percent of 36, which is 21.6 million, so contribution is 14.4 million, and profit is 14.4 minus 10, which is 4.4 million. That is a fall of 1.6 on 6, which is 26.7 percent, and 2.67 times 10 percent is 26.7 percent.

Bakery B at 36 million: variable cost is 80 percent of 36, which is 28.8 million, so contribution is 7.2 million, and profit is 7.2 minus 2, which is 5.2 million. That is a fall of 0.8 on 6, which is 13.3 percent, and 1.33 times 10 percent is 13.3 percent.

The leverage figure predicted both numbers before either was computed, which is why it is worth carrying in your head. On the upside it works the same way: A gains 26.7 percent and B gains 13.3 percent from the same 10 percent of extra revenue. Neither structure is better. A is better in a growing market and worse in a falling one, and the recommendation depends entirely on which one you think you are in.

Breakeven and the margin of safety

Breakeven revenue is fixed cost divided by contribution margin. For A that is 10 divided by 0.40, which is 25 million dollars. For B it is 2 divided by 0.20, which is 10 million dollars.

Margin of safety is how far revenue can fall before profit disappears. A can lose 40 minus 25, which is 15 million, so 37.5 percent of its revenue. B can lose 40 minus 10, which is 30 million, so 75 percent.

That is the sentence worth taking into a board meeting. The two bakeries look identical and one of them fails at a 38 percent revenue decline while the other survives a 75 percent one. Oliver Wyman's published method asks candidates to pick an approach and show why it beats the alternatives, checked 30 September 2026 at oliverwyman.com. Margin of safety is the number that makes the comparison, and it takes two divisions.

Two cautions on the figure. It assumes contribution margin holds as revenue falls, which is generous: a bakery losing volume usually loses its best-priced volume first, so contribution margin slips as well and the real safety margin is narrower. And it assumes the fixed cost stays fixed all the way down, which is also generous, because supervision and shifts come out in steps once volume drops far enough.

Both cautions push in the same direction for Bakery A and neither changes the ranking. A is the fragile site under any reasonable version of the assumptions, and the reason is 10 million dollars of fixed cost against 16 million of contribution.

A published case where the cost structures turn out identical

Oliver Wyman's Dairy Farm case publishes two farms and invites a comparison. Farm one holds 10,000 cows at 100 units of milk each, sells standard milk at a government-set 20 dollars a unit, pays 10 dollars a unit for feed and carries 5 million dollars of fixed cost. Farm two holds 5,000 cows at 80 units each, sells organic milk at 30 dollars, pays 15 dollars for feed and carries 3 million dollars of fixed cost.

The published conclusion reads: "Farm one and farm two have the same profit margin. On absolute dollar terms, farm one has $2 million in profit more than farm two", checked 30 September 2026 at oliverwyman.com.

We recalculated it by hand on 30 September 2026. Farm one: 1,000,000 units at 20 dollars is 20 million of revenue, feed at 10 dollars is 10 million, less 5 million of fixed cost leaves 5 million of profit, which is 25 percent of revenue. Farm two: 400,000 units at 30 dollars is 12 million, feed at 15 dollars is 6 million, less 3 million leaves 3 million, which is also 25 percent. Both of the firm's statements land.

Go one step further than the case does. Contribution margin for farm one is 10 million on 20 million, which is 50 percent. For farm two it is 6 million on 12 million, which is also 50 percent. Operating leverage is 10 divided by 5 for farm one, which is 2.0, and 6 divided by 3 for farm two, which is also 2.0. Breakeven volume is 5,000,000 divided by 10 dollars, or 500,000 units for farm one, and 3,000,000 divided by 15 dollars, or 200,000 units for farm two, and each is exactly half of its own output.

So the two farms are structurally the same business at two scales, despite a 50 percent price difference and a 2.5 times volume difference. The case does not say that, and it is the answer to the question a private equity buyer would ask next: neither farm is the safer asset, so the choice rests on price and on the market dynamics the case puts in its second branch.

Allocated cost is where the cost structure analysis goes wrong

The bakery comparison used site-level fixed cost. Reported fixed cost usually contains something else, and it changes closure decisions.

Say Bakery A's 10 million dollars of fixed cost breaks into 3.4 million of building and oven depreciation, 4.1 million of salaried production supervision and 2.5 million of group overhead allocated by revenue. Group profit across both sites is 12 million dollars.

Close Bakery A and the arithmetic looks bearable: lose 6 million of profit, keep the group. Then the 2.5 million of allocation lands on Bakery B, because group overhead did not leave with the site. Group revenue is 40 million, variable cost 32 million, site fixed cost 2 million and stranded overhead 2.5 million, so group profit is 3.5 million dollars.

Group profit fell from 12 million to 3.5 million, which is 8.5 million, against a site that was earning 6 million. The extra 2.5 million is the stranded allocation, and it appears nowhere in the closure business case unless somebody separates split two from split one.

The test is one question per line: if this site closed tomorrow, would this cost leave the group. Charles River Associates asks candidates to "show your math", checked 30 September 2026 at crai.com, and on a closure case the arithmetic that matters is the arithmetic of what stays.

Where cost structure work is scored

The weights below are ours, read from the rubric code our own grader runs. No consulting firm publishes a marking scheme, so none of this is Oliver Wyman's reading of its own dairy comparison.