Retail KPI Cheat Sheet: 21 Formulas, Worked From One Store's Numbers

How the number is calculated

Twenty-one formulas sounds like twenty-one things to learn. It is four. Net sales, cost of goods sold, average inventory at cost, and a unit count. Every metric here is one of those divided by another of those, or by a physical count of something in the store: square feet, transactions, door entries, staff, order lines. Once you can produce those four cleanly, the rest is division. Producing them cleanly is the part that goes wrong. Net sales means after returns, after discounts, after allowances, and without sales tax. A POS export labelled "sales" is often gross, and a gross sales figure inflates every ratio here that carries sales in the denominator, quietly and in the flattering direction. Cost of goods sold means the landed cost of what actually sold in the period. It is not the cost of what you purchased in the period, and in any month where receipts and sales are out of step those two figures differ enough to move turnover by half a turn. Why margin and markup are not the same percentage is the single most expensive confusion in retail math, and it is worth being slow about. Take one line. It cost $100 and it sells for $161.29. The spread is $61.29. That spread expressed against cost is a 61.3 percent markup. The same spread expressed against the selling price is a 38.0 percent margin. Both statements are true, both describe the identical transaction, and there is no arithmetic error anywhere. The conversions are Markup % = Margin % ÷ (1 − Margin %) and Margin % = Markup % ÷ (1 + Markup %). The damage happens when someone buys to a target expressed in the wrong convention. A buyer told to protect a 38 percent margin who buys at 38 percent markup on cost lands at 0.38 ÷ 1.38, which is a 27.5 percent margin. That is 10.5 margin points short of the target, on every unit, discovered a season later. The gross margin vs markup guide works through the conversion table in both directions, and the gross margin and markup calculators will do it on a single line. Some categories quote a third convention, keystone, meaning a 100 percent markup on cost, which is a 50 percent margin. When a vendor says keystone and a planner hears 50 points, they happen to agree. When a vendor says 40 percent and a planner hears 40 points, they do not. Average inventory appears in three formulas here and almost every published cheat sheet leaves it undefined. The two-point average of beginning and ending inventory is what almost everyone uses, and it is what this page uses: (268,000 + 242,000) ÷ 2 = $255,000. It is defensible for a full year in a business without violent seasonality. It is poor for a quarter that contains a peak, and close to meaningless for a single month, because a two-point average of a curve with a spike in the middle is not the average of the curve. The alternative is a thirteen-point average, the twelve month-end balances plus the opening balance, divided by thirteen. If your system can produce it, use it, and say on the report which one you used. Substituting ending inventory for average inventory is the common shortcut, and it flatters. For this store, turnover on ending inventory reads 4.22 instead of 4.01, and GMROI reads $2.59 instead of $2.46. Both overstate performance by 5.4 percent, and the overstatement is not random: it grows every time inventory is drawn down late in the period, which is exactly what a store does going into a slow season. So the shortcut is most flattering in the periods you were most likely to have underperformed. Then there is the basis question. Inventory turnover has two constructions. At cost it is COGS over average inventory at cost. At retail it is net sales over average inventory at retail. Published formula lists print one or the other and rarely say which. For this store they agree exactly: at cost, 1,021,760 ÷ 255,000 = 4.01, and at retail, 1,648,000 ÷ 411,290 = 4.01. That agreement is not a validation. It happens because a single uniform cost complement of 62.0 percent was applied to convert one side to the other, so the same ratio appears on both sides of the conversion by construction. They separate the moment realized margin diverges from the margin assumed in the conversion, which is precisely what markdowns and mix shift do all season. A store carrying a fashion classification at 55 percent planned margin that actually realizes 41 percent after clearance has an inventory-at-retail figure derived from an assumption its own P&L already contradicted. The retail inventory method guide covers how that conversion is built and where it drifts. The practical rule: pick the cost construction when you can get clean COGS, keep it, and never compare a cost-basis turn to a retail-basis turn across stores or years. GMROI puts a retail-basis numerator over a cost-basis denominator on purpose, which offends people the first time they see it. Gross margin dollars are earned at retail. Average inventory is the money actually tied up, which is a cost figure. The ratio answers a cash question: for every dollar of capital sitting in stock across the year, how many dollars of gross margin came back. It decomposes exactly into margin and turnover, since 0.38 ÷ 0.62 = 0.6129 and 0.6129 × 4.01 = $2.46. That identity is the reason GMROI is the primary output here. Two stores can post the same GMROI from opposite operating models, and the decomposition tells you which one you are running before you decide what to change. There is more on reading the two factors against each other in the GMROI guide, and the GMROI calculator computes the ratio directly. Four metrics here depend on a physical count rather than a ledger figure, and each has a definitional trap of its own. Selling area means the selling floor. Including the stockroom, the receiving bay and the manager's office in the denominator will drag sales per square foot down by a quarter or more and make a productive store look weak against its own history the year someone remeasured. Transactions means completed sales, not line items, which is what separates average transaction value from average line value. Traffic means door entries, and a door counter that counts staff, deliveries and the same customer twice will hold conversion rate down permanently. Average FTE means full-time equivalents, not headcount: two people at twenty hours are one FTE, and a store that flexes heavily on part-time labor will read very differently on the two bases. Sales per employee is only comparable across stores when every store computes FTE the same way. OTIF and fill rate measure different failures. OTIF is order level and unforgiving by design. An order arriving on time with one line short fails, and an order arriving complete a day late fails. Fill rate is line level, so the same delivery scores most of its lines as successes. For this store, OTIF reads 89.0 percent against a fill rate of 95.5 percent, and the gap between the two numbers is the useful part: it says the failures are scattered across many orders rather than concentrated in a few bad ones. Two suppliers with identical fill rates can have very different OTIF, and the one with the lower OTIF is the one disrupting more of your receiving days. The OTIF calculator computes both conditions separately so you can see which one is failing. The full argument about which denominator belongs under which line, and which convention should have authority when a target passes between buying and finance, is in the denominator problem.