GMROI: the Inventory Number That Outranks Margin

Margin ranks products wrongly, and here is the proof

A shelf, a warehouse bay and a credit line are all finite. What you are really allocating is not space but cash, and the only honest way to compare two products is to ask what each returns on the cash it consumes. Margin does not answer that, because margin is measured per sale and says nothing about how many sales a given pile of stock generates in a year.

Set the cash equal and let margin and turns fight. Product A carries $10,000 of average inventory at cost, sells at a 60% gross margin and turns twice a year. Turns times average inventory at cost gives cost of goods sold: 2 × $10,000 = $20,000. At a 60% margin, sales are $20,000 ÷ 0.40 = $50,000, so annual gross margin is $30,000 and GMROI is $30,000 ÷ $10,000 = 3.00. Product B also carries $10,000 at cost, sells at 25% and turns twelve times: cost of goods sold $120,000, sales $160,000, gross margin $40,000, GMROI 4.00. The product with less than half the margin returns a third more gross margin on the same cash.

That inversion is not a curiosity built from convenient numbers; it is what the formula does whenever turns differ enough. Nothing in the comparison is unfair: same cash, same year, same measurement. The slow product simply spends most of the year sitting still while the fast one is sold, repurchased and sold again five more times.

The identity: markup on cost times turns

The derivation takes three lines. Gross margin in currency is sales × gross margin percentage. Cost of goods sold is sales × (1 − gross margin percentage). Inventory turns are cost of goods sold ÷ average inventory at cost, so average inventory at cost is sales × (1 − gross margin percentage) ÷ turns. Divide the first by the third and sales cancels: GMROI = gross margin percentage ÷ (1 − gross margin percentage) × turns.

The first factor deserves a second look, because gross margin percentage ÷ (1 − gross margin percentage) is exactly markup on cost — the same conversion between margin and markup that trips up so many pricing decisions. A 60% margin is a 150% markup; a 25% margin is a 33.3% markup; a 50% margin is a 100% markup. So GMROI is simply markup on cost multiplied by turns, and you can compute it in your head from a price ticket and a stock-turn figure. Check the two products: 1.5 × 2 = 3.00 and 0.3333 × 12 = 4.00.

The identity also shows you the trade-off explicitly. Hold GMROI at 3.00 and ask what turns each margin needs: a 20% margin needs 12 turns, 25% needs 9, 30% needs 7, 40% needs 4.5, 50% needs 3, 60% needs 2 and 70% needs 1.29. That curve is the whole of assortment strategy on one line. A discount grocer and a jeweller can earn the same return on inventory investment while looking nothing like each other, and the moment you know two of the three numbers you can solve for the third. It also tells you what a promotion really costs: cutting margin from 40% to 30% drops the markup factor from 0.667 to 0.429, so turns have to rise by 56% just to hold GMROI still.

Reading the number, and the line at 1.00

GMROI is a ratio of currency to currency, so it reads as a multiple: 3.00 means three dollars of gross margin a year for every dollar of cash held in stock at cost. The line worth marking is 1.00. Below it, the category returns less gross margin over a whole year than the cash permanently sunk in it — and gross margin is not profit. Rent, wages, shrinkage, markdowns and the cost of the capital itself are all still to be paid out of that same number. Product E in the table, at a 30% margin and 2 turns, comes to 0.86: it generates $8,571 of gross margin a year against $10,000 of standing inventory. It may still deserve shelf space for reasons the ratio cannot see — it completes a range, it brings people into the store — but the decision to keep it should be a conscious one.

The other useful reading is comparative rather than absolute. What a good GMROI looks like depends entirely on the format and the goods, and any number quoted as a universal threshold should be treated with suspicion unless it comes with a named source and a defined sector. Use the ratio to rank your own categories against each other and to track each one against its own history; that comparison is always valid, because the accounting basis and the denominator convention are the same on both sides.

Average inventory is where the number gets fudged

The numerator of GMROI is hard to argue with; the denominator is not. Average inventory at cost can be computed a dozen ways, and the spread between them is enormous for anything seasonal. Take a category whose stock at cost runs 5, 6, 8, 10, 12, 14, 18, 26, 34, 40, 20 and 5 thousand from January to December, opening the year at $5,000. The thirteen-point average — the opening figure plus the twelve month-ends, divided by thirteen — is $15,615. The two-point average that most spreadsheets use, opening plus closing over two, is $5,000. They differ by a factor of 3.12.

Now attach a result. Suppose that category sold $155,000 at a 40% gross margin: cost of goods sold $93,000 and gross margin $62,000. On the honest thirteen-point denominator, GMROI is $62,000 ÷ $15,615 = 3.97, and inventory turns are $93,000 ÷ $15,615 = 5.96 — and the identity checks, since 0.6667 × 5.96 = 3.97. On the two-point denominator, GMROI is $62,000 ÷ $5,000 = 12.40. The same business, the same year, the same sales, and a number three times better because of one choice about how to average.

Three rules keep the denominator honest. Use at least thirteen points for anything seasonal, and weekly points if your season is short. Value the stock at cost, never at retail, or the ratio stops meaning a return on cash. And apply the same convention to every category you intend to compare, because a ranking built from two different averaging methods is worse than no ranking at all.

Sell-through and weeks of cover: the same idea in daily clothes

Nobody on a shop floor recomputes GMROI daily. What they watch instead are two operational cousins of the same ratio. Sell-through is units sold divided by units available in a period — receive 800 units on top of 200 already in stock, sell 750, and sell-through is 750 ÷ 1,000 = 75% with 250 units left. Weeks of cover is current stock divided by average weekly sales: it says how long the pile in front of you will last.

Weeks of cover connects to GMROI directly, because turns and cover are reciprocals of each other on a 52-week year: turns ≈ 52 ÷ weeks of cover. Product B at twelve turns is carrying 4.3 weeks of cover; product A at two turns is carrying 26 weeks. Substituting into the identity gives GMROI = markup on cost × 52 ÷ weeks of cover, which is the version a buyer can use in front of a supplier: if you want me to hold three months of your product instead of one, the markup has to triple for the shelf to earn the same.

Sell-through does something GMROI cannot: it flags the problem early enough to act. GMROI is an annual, backward-looking ratio; sell-through after four weeks tells you whether a buy is going to end in a markdown while there is still time to reorder, transfer or stop. Use sell-through and cover to run the week, and GMROI to decide what to buy next season. And remember the third lever the ratio hides — payment terms. GMROI measures the return on inventory at cost; if the supplier is financing 60 days of it, the cash you actually risk is smaller than the denominator suggests, which is the honest argument for taking a slower product on longer terms.

Frequently asked questions

Is GMROI the same as inventory turnover?

No. Turnover counts how many times the stock is replaced in a year and knows nothing about margin; GMROI multiplies turnover by the markup on cost, so it measures money returned rather than movement. Two categories can turn identically and produce completely different GMROI, and a category can raise its turnover by discounting while its GMROI falls. Watch both: turnover tells you about the flow, GMROI about the return on the cash inside it.

Should average inventory be valued at cost or at retail?

At cost. The whole point of the ratio is the return on the cash you have committed, and the cash you committed is what you paid the supplier, not what you hope to charge the customer. Valuing at retail inflates the denominator and deflates the ratio, and it does so unevenly across categories, so the ranking itself becomes wrong. Retailers using the retail method of accounting still convert to cost before computing GMROI. Whichever you use, apply it to every category or do not compare them.

What GMROI should I be aiming for?

There is no defensible universal target, and any figure offered as one should come with a named publisher, a defined sector and a stated denominator convention or be discarded. Build your own reference instead: compute GMROI for every category on the same basis, rank them, and look at what the top and bottom quartiles have in common. The number that matters is whether a category clears the return you need on working capital, and that threshold is yours, not the industry's.

Can GMROI be negative?

Only if gross margin itself is negative, which happens when markdowns, shrinkage or freight push the realised cost of goods sold above the money collected. It is rare in a full year and common in a clearance month, which is why an annual figure can hide a category that has been selling below cost since the season broke. If you see a GMROI near zero, look at the monthly gross margin before concluding anything about turns.

How does GMROI relate to the cash tied up in the business?

GMROI looks at inventory alone, which is only one of the three components of working capital; receivables and payables sit beside it. A category with a superb GMROI financed on 90-day supplier terms consumes far less of your own cash than the ratio implies, and one with the same GMROI paid cash on delivery consumes far more. Read GMROI alongside the cash conversion cycle rather than instead of it — the companion article in this series works that calculation through in full.