A retail case asks you to diagnose or improve the performance of a shop, chain or retail brand — why sales have fallen, whether a store network is productive, how an assortment or a channel should change. It runs on a small set of metrics, and most retail cases are decided by knowing which of them to ask for. The sector appears often in consulting interviews because it is familiar enough to need no explanation and quantitative enough to test real analysis.
This page covers those metrics and works three practice prompts through to a recommendation. Figures are fictional and constructed for practice.
The decomposition that opens most retail cases
Retail revenue breaks down cleanly, and the break is where the case starts:
Revenue = Traffic × Conversion rate × Average basket size × Frequency
Each term has a different cause and a different fix, which is why locating the change before analysing it saves most of the case:
| Term | What it measures | Typical causes of a decline |
|---|---|---|
| Traffic | People entering | Location, competition, catchment change, marketing |
| Conversion | Share who buy | Availability, price, staffing, store experience |
| Basket size | Spend per transaction | Assortment, merchandising, promotion mix |
| Frequency | Visits per customer per period | Loyalty, category, competitor proximity |
A candidate who asks “has traffic fallen, or is traffic flat and conversion down?” has narrowed the case in one question. A candidate who builds a full profitability tree has not.
On the cost side, the metrics worth knowing:
- Sales per square foot — the productivity of the space, and the number that reveals whether a store is underperforming its footprint.
- Inventory turns — how many times stock cycles a year. Low turns mean capital tied up and markdown risk.
- Gross margin after markdown — the headline margin is rarely the realised one.
- Shrinkage — theft, damage and error. Small as a percentage, large against a thin net margin.
- Like-for-like sales — growth excluding openings and closures.
That last one matters more than candidates expect. A chain opening 30 stores a year can report growing total sales while every existing store declines. Interviewers ask for like-for-like precisely to see whether the candidate knows the difference.
The structural feature underneath all of this is that retail carries a high gross margin and a thin net margin, because the gross margin has to cover a large fixed operating base. A grocery chain operating on low single-digit net margins loses a disproportionate share of profit to a one-point rise in shrinkage or markdown. Understanding that ratio is what makes a retail recommendation credible.
Case 1: Sales are down, and it is not the market
Prompt. A 60-store fashion chain has seen revenue fall 8 per cent over two years while the overall market grew slightly. Store count is unchanged. Management believes online competition is responsible. Is it?
Solution
Decompose before accepting the explanation. Online competition would most likely show up as reduced traffic. Ask for the split:
Suppose the data shows traffic down 3 per cent, conversion down from 24 to 21 per cent, and average basket flat.
Conversion is doing most of the work: a fall from 24 to 21 per cent is a 12.5 per cent relative decline, against 3 per cent on traffic. Combined: 0.97 × 0.875 = 0.849, roughly the 8 per cent revenue decline once basket is held flat, allowing for rounding.
That reframes the case. Online competition explains the traffic decline. It does not explain conversion, because customers who are already in the store have already chosen not to buy online today. Something is happening in the store.
Investigate conversion causes:
- Availability. Is the size and colour the customer wants in stock? Ask for the rate of stockouts on core sizes. In fashion this is the most common conversion killer and the most fixable.
- Staffing. Have store hours been cut? Conversion in assisted categories is sensitive to staff presence, and labour reductions made to protect margin frequently cost more revenue than they save.
- Assortment. Has the range shifted away from what the catchment buys?
- Price perception. Not price level — perception. A customer who believes the item is cheaper online will not buy it in front of you.
Recommendation. The declared cause is wrong in its most important part. Traffic loss to online is real but small; the conversion decline is the larger problem and is internal. Prioritise a stockout analysis on core sizes and a review of whether store labour hours were reduced during the period, since both are measurable within weeks and both are reversible. Reserve judgement on assortment until the availability question is settled, because a range that looks wrong when half the sizes are missing may not be.
What this case tests. Whether you test the client’s stated cause rather than analysing it, and whether you decompose before hypothesising.
Case 2: Omnichannel as a cost-to-serve problem
Prompt. A home goods retailer with 120 stores has grown online sales to 25 per cent of revenue, but total profit has fallen over the same period. Online gross margin is the same as in-store. Why is profit down, and what should they do?
Solution
Equal gross margin and falling profit means the cost sits below the gross margin line — in fulfilment, returns and the cost of serving the order.
Build cost to serve by fulfilment method. This is the analysis that makes the case:
| Method | Pick cost | Delivery cost | Returns cost | Total per order |
|---|---|---|---|---|
| In-store purchase | — | — | $1.50 | $1.50 |
| Ship from distribution centre | $3.00 | $8.00 | $6.00 | $17.00 |
| Ship from store | $6.00 | $9.00 | $6.00 | $21.00 |
| Click and collect | $3.00 | $1.00 | $2.00 | $6.00 |
Two findings fall out. Shipping from store is the most expensive method, because store labour is more expensive per pick than warehouse labour and store shipments are small and uneconomic. And returns are a large, recurring cost on delivered orders — in home goods, return rates are high enough that the returns column can exceed the pick cost.
Then check the mix. If the growth in online has come disproportionately through ship-from-store — which is common, because it is the easiest capability to switch on — the average cost to serve has risen sharply even while gross margin held.
Work it: if 25 per cent of revenue is online, and of that 60 per cent ships from store at $21 versus 40 per cent click-and-collect at $6, the blended online cost to serve is $15 per order. Against an average order of, say, $95 at a 40 per cent gross margin — $38 of gross profit — that is 39 per cent of gross profit consumed by fulfilment.
Recommendation. The problem is fulfilment mix, not channel. Shift volume from ship-from-store to click-and-collect and to distribution-centre fulfilment by changing the defaults the customer sees and the incentives at checkout — free collection against paid delivery is the standard mechanism and it works. Attack the returns rate on the highest-return categories through better product information and sizing, since returns are the single largest line. Use ship-from-store only where it genuinely wins a sale that would otherwise be lost to a stockout.
What this case tests. Whether you go below gross margin when gross margin is flat, and whether you know that fulfilment method rather than channel drives retail economics.
Case 3: Assortment and the tail
Prompt. A speciality grocery chain carries 14,000 SKUs. The merchandising director wants to cut 20 per cent of them to simplify operations. The commercial team objects that range is the chain’s point of difference. Who is right?
Solution
Both claims are testable, so test them. The merchandising claim is that the tail costs more than it earns. The commercial claim is that the tail drives traffic. These are not contradictory, and the answer depends on which SKUs.
Ask for the sales distribution. In most grocery assortments, a small share of SKUs generates the majority of units. Suppose the bottom 20 per cent of SKUs by unit volume — 2,800 items — generates 3 per cent of units.
Cost of the tail. Each SKU carries shelf space, replenishment labour, inventory, and markdown or waste risk. If the fully loaded cost of carrying a slow SKU is roughly $400 a year across 60 stores, 2,800 SKUs cost around $1.1 million a year.
Value of the tail. This is the harder number and the one that decides the case. The commercial team’s argument is not that the tail sells; it is that some customers shop the chain because of it, and those customers buy the rest of their basket there too.
Test it with basket data: what share of baskets contain at least one tail SKU, and what is the average value of those baskets? If 12 per cent of baskets contain a tail item and those baskets are 40 per cent larger than average, the tail is attached to a materially valuable customer group and cutting it risks the whole basket rather than 3 per cent of units.
Recommendation. Neither position as stated. Cut on basket attachment rather than on unit volume — remove the slow SKUs that appear in no distinctive baskets, and protect the slow SKUs that anchor large ones, even where their own sales are negligible. That yields most of the operational saving while leaving the differentiation intact. If the data to distinguish them does not exist, building it is the first recommendation, because a 20 per cent cut made blind is the one decision here with real downside.
What this case tests. Whether you recognise that a low-volume SKU can carry high-value baskets, and whether you can find the version of the answer that is neither of the two positions presented.
Practising retail cases
Retail rewards preparation because the metrics recur. Learn the revenue decomposition until asking for the split is automatic, know roughly what net margins look like in the sub-sectors — grocery thin, fashion higher and markdown-exposed, speciality higher still — and get comfortable moving between per-store, per-square-foot and per-basket views of the same problem.
The consumer packaged goods primer covers the supplier side of the same value chain, which is where a retail case often turns when the question moves to supplier terms. More worked cases in other formats are in the case study library, the arithmetic is in case maths techniques, and the case interview hub sets out a full preparation plan.
