Retail Case Interview 2026: Three Ratios, One Call

Three ratios every retail case interview turns on

The equation under a store is simple enough to write from memory. Units sold times the price actually realized, minus units sold times unit cost, minus the cost of the space and the people. Everything else is a ratio built on top of it.

That matters because retail cases are usually lost on the realized price. A chain reports list prices and sells at a mix of list and markdown, so the average ticket is a consequence of mix rather than a decision anyone made. Ask for the markdown share of units before asking anything about pricing strategy.

The three ratios then sit in different places in that equation.

1. Sales per square foot. Revenue divided by selling area. It tests location and productivity, and it moves when traffic or the ticket moves.
2. Gross margin. Revenue minus cost of goods, over revenue. It tests buying, mix and markdown together, which is why it needs unpacking rather than quoting.
3. Stock cover. Inventory against sales. It tests whether capital is trapped, and it predicts next season's markdown, since stock that does not sell gets discounted eventually.

One more question belongs early, and it is about the denominator of growth. Like-for-like sales strip out new stores. A chain whose total sales rise while existing stores decline is a chain buying growth with capital.

The retail case prompt Alder and Co brought in

Alder and Co runs 120 mall-based apparel stores in the United States. Store count did not move across the two seasons in question, which means total sales and like-for-like sales are the same number here. Where they are not the same, ask which one you are being given.

The chief merchant reports that profit halved while units sold stayed flat, and wants to know what to change before the next buying cycle.

The data, as the interviewer releases it:

1. Selling area. 120 stores at 5,000 square feet each is 600,000 square feet.
2. Units sold. 10.0 million last season and 10.0 million this season.
3. List price. 50 dollars, unchanged.
4. Markdown price. 35 dollars, a 30 percent cut off list, unchanged.
5. Markdown mix. 30 percent of units last season, 40 percent this season.
6. Unit cost. 24 dollars, unchanged.
7. Store and corporate costs. 180 million dollars last season, 186 million this season.
8. Inventory at cost. 80 million dollars at season end.

Read that list again and notice what did not change. Price, cost, units and store count are all flat. Only the mix of what sold at what price moved, and that alone will carry the whole result.

A case interview structure hung on units and tickets

The structure follows the equation rather than a template. Four branches, and each one is a term of it. Written that way it is an issue tree built from the client's own arithmetic, and every branch is a quantity this case can put a number in.

• Units. Traffic, conversion, units per transaction, and the split between full price and markdown.
• Realized price. List price, markdown depth, markdown share, and returns.
• Unit cost. Landed cost, freight, and shrink.
• Fixed cost. Occupancy, store payroll, and corporate overhead.

Two firms publish opposite advice on how prepared that structure should be. Kearney's recruiting blog carries advice from Harsha, one of its consultants, dated 26 May 2022: "One of the most important things in a case interview is to lay a foundation. Come prepared with different kinds of structures for different kinds of cases", read at kearney.com. Bain's case page takes the other side, saying "We're not looking for a single right answer; we're looking for how you think, structure problems, and build on ideas as the conversation evolves".

Both are first-hand statements about what those two firms reward, and they do not agree. The way to satisfy either is to derive the branches from the client's own equation, which is what the four above are.

Computing the markdown mix, line by line

Season one. Markdown units are 30 percent of 10.0 million units, so 3.0 million units. Then 3.0 million units at 35 dollars is 105 million dollars. Full price units are 7.0 million, and 7.0 million at 50 dollars is 350 million dollars. Revenue is 455 million dollars.

Cost of goods is the same both seasons because units and unit cost are flat: 10.0 million units at 24 dollars is 240 million dollars. So gross profit is 455 minus 240, which is 215 million dollars. Gross margin is 215 over 455, which is 47.3 percent.

Season two. Markdown units are 40 percent of 10.0 million, so 4.0 million units, and 4.0 million at 35 dollars is 140 million dollars. Full price units are 6.0 million, and 6.0 million at 50 dollars is 300 million dollars. Revenue is 440 million dollars. Gross profit is 440 minus 240, which is 200 million dollars, and gross margin is 200 over 440, or 45.5 percent.

Operating profit. Season one is 215 minus 180, which is 35 million dollars. Season two is 200 minus 186, which is 14 million dollars. The fall is 21 million dollars.

Now split it. Gross profit fell 15 million and fixed costs rose 6 million, and 15 plus 6 is the whole 21. Price the mix directly as a check: the extra million units that moved from list to markdown each lost 50 minus 35, so 1.0 million units at 15 dollars is 15 million dollars. The two routes agree, which is what tells you the split is complete.

So mix carries 15 of the 21 million dollars, which is 71 percent of the fall. Cost growth carries the rest. Sales per square foot went from 455 million over 600,000 square feet, which is 758 dollars, to 440 million over 600,000, which is 733 dollars. The space did not get worse. The ticket did.

The stock number, against a public benchmark

Inventory at cost is 80 million dollars. Against a cost of goods of 240 million that is three turns a year. Against sales it is 80 million over 440 million, which is 2.18 months of cover at this season's rate.

That second figure has a public comparison, which is rare in a case and worth using. The Census Bureau's release CB26-154 of 16 September 2026 puts the inventories to sales ratio for clothing and clothing accessories stores at 2.11 for July 2026, with retail trade as a whole at 1.27 and department stores at 2.63. It is published at census.gov. Alder and Co sits slightly heavier than its own sector.

The caveat is what makes the comparison honest. The Census ratio divides sales measured at retail prices by inventories that its own glossary defines as "the value of stocks of goods held for sale" at cost, checked 29 September 2026. The numerator and denominator are on different bases, so the ratio is a benchmark against other retailers rather than a physical turn count. Say that when you quote it, and you are the only person in the room who has.

One channel fact belongs here too, because it decides where the next season's units should sit. Online sales were 17.1 percent of total United States retail sales in the second quarter of 2026, growing 12.2 percent against 6.7 percent for retail overall, in Census release CB26-133. A chain holding 2.18 months of cover in stores is financing a footprint the demand is slowly leaving.

The case interview exhibit, two seasons side by side

The last line is the one to say out loud. Realized price fell from 45.50 to 44.00 dollars, a drop of 1.50 dollars a unit, and 10.0 million units at 1.50 dollars is the 15 million dollars already found. A retail exhibit is usually asking you to find a per-unit number hiding inside two totals.

The lever that looks obvious and loses money

The instinct is to buy less. Fewer units into the season means less stock to discount, and the margin rate improves. Price it before recommending it.

Suppose the buy comes down by 700,000 units and the units removed are the ones that would have been marked down. Full price units stay at 6.0 million for 300 million dollars. Markdown units fall to 3.3 million, and 3.3 million at 35 dollars is 115.5 million dollars, so revenue is 415.5 million. Cost of goods is 9.3 million units at 24 dollars, which is 223.2 million dollars. Gross profit is 415.5 minus 223.2, which is 192.3 million dollars.

That is 7.7 million dollars worse than the 200 million the chain just earned, while the margin rate rises to 46.3 percent. The reason is one number: a markdown unit sells for 35 dollars against a cost of 24, so it still contributes 11 dollars of gross profit. Removing a unit that earns 11 dollars to protect a percentage is a rate improvement bought with cash.

This is the arithmetic the published guidance skips. Bain's own apparel case lists raising prices among the ways to drive revenue, with the honest counterargument that "volume would fall if we increased the price", checked 29 September 2026 at bain.com. The counterargument is right and unquantified, and quantifying it is the whole job.

The call, and the sell-through it needs

The recommendation, stated as a partner would want it.

Hold the buy flat and cut markdown depth from 30 percent to 20 percent, taking the markdown price from 35 to 40 dollars. Full price revenue is unchanged at 300 million dollars. Markdown revenue becomes 4.0 million units at 40 dollars, which is 160 million dollars. Revenue is 460 million and gross profit is 460 minus 240, which is 220 million dollars. That is 20 million dollars more than this season, and it recovers 95 percent of the 21 million dollar fall.

Now the condition, because a shallower cut will move fewer units. Cost of goods is paid on everything bought, so write gross profit as 300 plus 160 times the sell-through, minus 240. Set that equal to today's 200 million dollars and the sell-through solves to 140 over 160, which is 87.5 percent.

So the test is specific: at least 87.5 percent of the units currently clearing at 35 dollars must still clear at 40 dollars. Below that threshold the deeper discount was the better policy, and the chain should find out on a category before it finds out on a season.

The risk sits in the unsold remainder. Every unit that fails to clear stays in inventory, pushes stock cover above the 2.18 months computed earlier, and comes back next season at a deeper cut. That is the mechanism by which a margin repair becomes next year's markdown problem, and it is the sentence to end the case on.

What the scoring punishes in a retail case

The weights here are ours, taken from the rubric code in src/lib/grading, and no consulting firm publishes a rubric. Exhibit answers are scored on So what 30, Observation 25, Implication 25 and Next test 20. Numeric answers are scored on Answer 40, Working 25, Units 15 and Interpretation 20. Each line is reported on its own with the first correction named, and the reasoning behind every weight is on how an answer is assessed.

1. Reading the margin rate and stopping. Saying margin fell 1.8 points is an observation with no implication. So what 30 is the largest line on the exhibits rubric because that is the most common place to stop.
2. Quoting a benchmark without its basis. Comparing a cost-based inventory figure to a retail-based ratio without saying so is a Units failure worth 15 points, and it is the error a public benchmark invites.
3. A lever with no threshold. Recommending shallower markdowns without the 87.5 percent condition gives the client something to disagree with rather than something to test. On the synthesis rubric that is the Risk line, worth 20.

The fix for all three is one habit: after every number, write the sentence that says what changes because of it.

Practicing the retail case mix arithmetic

Rerun this case with one input changed and mark yourself against the method above. Set markdown depth to 40 percent off, so the markdown price is 30 dollars: markdown revenue becomes 4.0 million units at 30 dollars, which is 120 million, revenue is 420 million, and gross profit is 180 million. Compare that to the 200 million the chain earned and you have priced depth as its own lever, separate from mix.

Then change the mix instead. At a 50 percent markdown share, 5.0 million units at 35 dollars is 175 million and 5.0 million at 50 dollars is 250 million, so revenue is 425 million and gross profit is 185 million. Two inputs, two different sizes of damage, and now you know which one to ask about first.

Our chart drill runs exhibit reads of this shape and returns the four exhibit lines separately, and the math drill times the unit arithmetic. Both are free and need no account. Neither teaches retail, and neither will tell you whether a merchant can hold price, which is the judgment this case rests on. Our assessment simulators cover firm built formats, and that page states that none of the six is live yet.

For the same equation applied without a mix question, see the profitability framework. For the version where the client is deciding whether to open the format at all, see market entry.