Merchandise Planning KPIs by Vertical

A merchandise planning KPI set is the short, fixed list of measures a planning team uses to judge whether a category is on plan and to decide what to change, each with a stated formula, grain, cadence and owner. It is not every number the reporting system can produce; it is the handful that a merchant, planner or allocator will act on this week, plus the few that the season hindsight uses to set next season's targets. The formulas are the same in every category; which KPI leads the conversation is not. This guide sets out the five KPI families, separates the leading measures from the lagging ones, works three of them through one category, and then follows ten verticals through the KPI that genuinely governs each.

It belongs to the by-vertical series beside open-to-buy by vertical, markup and pricing by vertical, markdown and exit strategy by vertical, assortment planning by vertical, allocation and replenishment by vertical, demand forecasting by vertical, store clustering by vertical, merchandise hierarchy by vertical and planning calendar by vertical. Retail math for merchandise planners carries the full derivation of each formula; this guide is about which of them to run the business on. Most formulas named here have their own page, with a calculator, in the retail formulas library.

What a merchandise planning KPI set is

Four properties separate a KPI set from a report. A KPI has a formula with its denominator named, because sell-through on receipts and sell-through on total units available are different numbers for the same season. It has a grain, because a category can be on plan in total while the sizes or shades the customer wants are gone. It has a cadence, because a measure reviewed less often than the decision it informs arrives too late to change it. And it has an owner — the merchant who can chase or cut, the planner who can move the receipt plan, the allocator who can move stock between doors — because a measure that no one has the authority to act on is commentary, not a KPI.

The set should be short. A weekly trade pack with forty columns does not give the meeting forty KPIs; it gives it no agreed view of which three numbers decide this week's actions. A workable shape is one primary KPI per category, two or three secondary KPIs that explain movement in the primary one, and a small number of guardrails — in-stock, aging, markdown percent — that stop the primary KPI being improved at the expense of something else.

The five KPI families and their formulas

Each formula below is stated in the form used across this site's formula pages, so a figure quoted from one of those pages and one computed here are on the same basis. Where a measure has two defensible conventions, the convention used is named.

Sales and productivity

Comp sales % = (this period's sales in the comp base − the same period's sales last year in the same base) ÷ last year's sales in that base. The comp base is the set of doors and channels open long enough, under the business's own comp rule, to compare like for like; the comp store sales formula page works it. Comp is the trading headline, but it is a lagging total: it says the business grew or shrank, not which category, style or size did it.

Sales per option = net sales ÷ number of options carried, where an option is one style-color. It is the breadth-and-depth check: a category that adds options faster than it adds sales is spreading the same demand across more inventory positions. Read it beside share of sales against share of options — illustratively, a class carrying 20% of the category's options and producing 12% of its sales is under-earning its space in the assortment. The SKU productivity entry defines the term and the SKU productivity calculator works net sales per option and sell-through by category.

Inventory

Sell-through % = units sold ÷ units received. This is the convention on the sell-through rate formula page: units received is the original buy for a finite-life style, or beginning-of-period stock for a replenishment read. Split it into full-price sell-through (units sold at the original ticket ÷ units received) and total sell-through, because the two diverge as soon as a markdown is taken.

Weeks of supply = units on hand ÷ average weekly unit sales, usually on a trailing four- or six-week window (weeks of supply formula). Forward weeks of supply runs the same stock against the forecast instead of the trailing rate (forward weeks of supply formula), and it is the one to trust ahead of a peak or a floor set.

Stock-to-sales ratio = beginning-of-month stock ÷ that month's sales, both at retail (stock-to-sales ratio formula). It is the monthly grain of the merchandise financial plan, and the ratio the receipt plan is built from.

Inventory turns = cost of goods sold ÷ average inventory at cost (inventory turns formula). Turns measures speed and is blind to margin.

Margin

IMU % = (initial retail − landed cost) ÷ initial retail (initial markup formula). Initial markup is the margin built into the first ticket. It is quoted on retail, not on cost, so a 60% IMU is a 150% markup on cost.

MMU % = (net sales − cost of goods sold) ÷ net sales (maintained markup formula). Maintained markup is what the season keeps after markdowns, discounts and shrink. Gross margin % uses the same expression after cash discounts and workroom costs (gross margin percent formula); where neither applies, the two coincide.

Markdown % = markdown dollars ÷ net sales (markdown percent formula). The same markdown dollars over original retail is a smaller percentage; state which base a report uses.

GMROI = gross margin dollars ÷ average inventory at cost (GMROI formula). GMROI combines margin and speed into one return on the inventory investment, and it equals turns multiplied by gross margin dollars over cost of goods sold.

Flow

Receipts against plan % = (actual receipts − planned receipts) ÷ planned receipts, by week or month, in units and at cost. It is the plan-versus-actual variance applied to the receipt plan, and it explains a sell-through or weeks-of-supply movement before anyone blames demand.

PO coverage % = units on open purchase orders ÷ planned units (PO coverage formula). It is the buyer's status on whether the buy is complete, behind or over, and it is a leading measure for every later inventory KPI.

Fill rate % = units shipped ÷ units ordered (fill rate formula). On the wholesale side it measures what accounts actually received against their orders; order and line fill rate are its stricter variants.

Accuracy

Plan-versus-actual variance % = (actual − plan) ÷ plan, applied to sales, receipts, markdowns, margin and stock. Forecast accuracy is commonly stated as 1 − (sum of absolute forecast errors ÷ sum of actual units) at a named grain and lag; it is only comparable between periods when the grain and the lag are the same. Both are lagging by construction, and both are the inputs the season hindsight uses to decide whether a miss was the forecast or the execution.

Leading and lagging: which KPI belongs in which meeting

A KPI is leading when it moves before the season's result is fixed and while a decision can still change that result. It is lagging when it reports a result that has already happened. Both are necessary; they belong in different meetings. The weekly trade meeting needs the leading set, because its job is to decide this week's chase, markdown, transfer and receipt moves. The season hindsight needs the lagging set, because its job is to set next season's targets and buy rules. A trade pack built from lagging KPIs reports the miss without the means to prevent it; a hindsight built from leading KPIs re-litigates decisions instead of learning from the outcome.

Two rows deserve a note. IMU changes category depending on when it is read: during line review and the buy it is leading, because a style below target can still be re-costed, re-ticketed or dropped; after the goods land it is a lagging fact the season inherits. And fill rate is lagging for the order it measures but leading for the account relationship, because a short-shipped account sells through less of what it was promised and reorders less in the next window.

The weekly trade meeting guide sets out the agenda and decision rights that turn the leading set into actions, with the WSSI as the meeting's backbone. How to hindsight a season sets out how the lagging set is read before the next buy locks.

Worked example: sell-through, weeks of supply and GMROI for one class

The figures below are illustrative, chosen because they divide cleanly. They are not benchmarks and are not drawn from any brand.

An apparel brand's DTC channel carries a women's knit-tops class through a 13-week season. Every unit carries a $20 landed cost and a $50 initial ticket, an IMU of ($50 − $20) ÷ $50 = 60.0%. The class opens with no stock. It receives 2,400 units at the start of week 1 and a second drop of 1,600 units at the start of week 5, for 4,000 units received. A 30% markdown, to $35, is taken on part of the class in week 7. The read is taken at the end of week 8. Returns are left out to keep the arithmetic visible.

Weekly units sold and ending stock

Check: 250 + 270 + 290 + 310 = 1,120 units in weeks 1 to 4, and 350 + 330 + 310 + 290 = 1,280 in weeks 5 to 8, for 2,400 sold. Receipts of 4,000 less 2,400 sold leaves the 1,600 on hand at the end of week 8. Of the 2,400 units sold, 2,000 sold at the $50 ticket and 400 at the $35 markdown price.

Step 1 — sell-through. Units sold ÷ units received = 2,400 ÷ 4,000 = 60.0%. Full-price sell-through = 2,000 ÷ 4,000 = 50.0%. The ten-point gap is the markdown's contribution, and it is the gap a total sell-through figure hides.

Step 2 — weeks of supply. The trailing four-week average is 1,280 ÷ 4 = 320 units a week. Weeks of supply = 1,600 ÷ 320 = 5.0 weeks. With five weeks left in the season, the trailing figure says the class clears on time. The weekly column says otherwise: sales have fallen by 20 units in each of the last three weeks. If that continues, weeks 9 to 13 sell 270 + 250 + 230 + 210 + 190 = 1,150 units, leaving 1,600 − 1,150 = 450 units at the season exit. The trailing average is pulled up by week 5's peak; the trend is the leading signal.

Step 3 — margin. Net sales = 2,000 × $50 + 400 × $35 = $100,000 + $14,000 = $114,000. Cost of goods sold = 2,400 × $20 = $48,000. Gross margin dollars = $114,000 − $48,000 = $66,000, a gross margin of $66,000 ÷ $114,000 = 57.9%. Markdown dollars = 400 × ($50 − $35) = $6,000, which is $6,000 ÷ $114,000 = 5.3% of net sales.

Step 4 — GMROI. Average inventory is the mean of the eight weekly closing positions: (2,150 + 1,880 + 1,590 + 1,280 + 2,530 + 2,200 + 1,890 + 1,600) ÷ 8 = 15,120 ÷ 8 = 1,890 units, which at $20 is $37,800 at cost. GMROI = $66,000 ÷ $37,800 = 1.75 for the eight weeks. Cross-check through turns: eight-week turns = $48,000 ÷ $37,800 = 1.27, and gross margin dollars over cost of goods sold = $66,000 ÷ $48,000 = 1.375; 1.27 × 1.375 = 1.75.

What the three numbers say together. Sell-through alone says the class is close to a straight-line pace — 60.0% sold at week 8 of 13, against the 61.5% that eight weeks of thirteen would imply. Full-price sell-through says a sixth of that progress was bought with markdown. Weeks of supply says the stock lasts the season at the trailing rate, and the trend says it does not. GMROI says the inventory investment has earned well so far, at a margin that will fall as the residual 450 units are cleared. No single KPI tells the trade meeting what to do; the set does. Here the decision is about the residual: whether to move stock between channels, deepen the markdown on the slowest colors now, or hold full price and accept a larger end-of-season exit.

A note on GMROI periods. The 1.75 is an eight-week figure. It can be compared with another eight-week figure for the same class, but not with an annual GMROI, and multiplying it up to a year assumes the class trades the same way for fifty-two weeks, which a seasonal class does not.

The ten verticals compared

Two patterns run through the table. The primary KPI is almost never the total; it is a total read at the grain the customer buys at — a size, a shade, a configuration, a piece. And the trap is almost always a correct formula applied on the wrong base: the wrong denominator, the wrong period, the wrong cost.

Apparel: full-price sell-through against the weeks on floor

Seasonal apparel lives inside a price window. A style is bought once, lands on a floor set date, holds its full ticket for a planned number of weeks and then enters the markdown cadence. The primary KPI is full-price sell-through by style-color, read against the number of weeks the style has been on floor, because the same 40% sell-through is a chase signal in week 3 and a markdown signal in week 9. Total sell-through cannot carry that decision, because once a markdown is taken it rises regardless of whether the style was right.

The secondary KPIs explain movement in the primary one. Forward weeks of supply against the remaining full-price weeks says whether the stock clears at the ticket; markdown percent says how much of the season's progress was bought; size-level in-stock rate says whether a slowing style is losing demand or losing the sizes that sell. A style whose medium and large are gone will read as slow demand while the sizes left on the rail are simply the ones nobody fits, and a markdown on that style clears residue rather than responding to demand. Size and pack optimization covers the curve behind that read.

The grain moves through the season. In the first two or three weeks after a drop, the read is style-color full-price sell-through against plan, because it is the earliest evidence that the buy was right and the window for a chase is still open — the in-season chase covers how short that window is. Mid-season the read moves to size and door, where allocation and transfers can still help. At season end it becomes the hindsight's lagging set: maintained markup, the exit residual, and full-price sell-through by style compared with the buy's planned rate.

Apparel brands with a wholesale channel carry a second version of the same KPI. Sell-in is not sell-through: an account's order shipped is revenue for the brand, but the season's real result is the account's sell-through at full price, which decides the markdown money conversation and the next season's order.

Footwear: size-run and width completeness by model

Footwear is planned in pairs, by model and colorway, across a size run and, for many models, more than one width. A model sells only while the core of its run is on the shelf: the half sizes in the middle of the curve, in each width the model is planned in. The primary KPI is run completeness — the share of a model's core size and width positions that hold stock — read by door and by channel. A model can show a strong sell-through while its core sizes are gone, and from then on its remaining pairs are the edges of the run, which the headline figure counts as inventory and the customer does not.

Completeness is a positional measure, the footwear form of in-stock: in-stock core size-width positions ÷ planned core size-width positions, at a named grain. A model planned in eight core sizes in two widths at a door has sixteen core positions; when four of them are empty, the run is 75% complete at that door, and the model's sell-through from that point mostly measures what fits nobody. Set the core in advance from the curve, not after the fact from what sold, because a size that was never stocked never sells and drops out of the next curve.

Secondary KPIs follow the run. Sell-through and weeks of supply are held by size band rather than by model, because, illustratively, a model with 12 weeks of cover in total can have two weeks in the core and thirty at the edges. Carryover models — the year-round franchise colorways — run as replenishment, so in-stock and weeks of supply against the reorder lead time govern them, not seasonal sell-through. And the model year matters: when a franchise refreshes, the outgoing model's residual is read against the changeover date, because a pair carried past it competes with its own successor.

The trap is the reverse of apparel's. Apparel over-reads total sell-through; footwear over-reads model sell-through. A buyer who reorders a strong model to its original size curve, without checking which sizes drove the strength, rebuys the edges. Assortment planning for footwear brands covers the depth arithmetic behind the run.

Accessories & bags: two pools, two KPIs

Most accessories and bags assortments are two businesses under one category line. A carryover core — the signature bag in its hero colors, the black belt, the basic small leather goods — sells all year with little markdown exposure. A seasonal fashion layer — the trend shapes, the seasonal colorways, the gifting sets — behaves like apparel, with a full-price window and an exit. The two pools need two primary KPIs, and averaging them produces a number that describes neither.

On the core, the primary KPI is GMROI, with in-stock as its guardrail. Core accessories turn slowly against apparel and rarely mark down, so turns alone understates them, and GMROI — which credits the margin each turn carries — is the fair read of whether the inventory investment is earning. In-stock on the hero colors matters as much, because a core bag that is out of stock in black at a key door loses the sale entirely rather than substituting.

On the seasonal layer, the primary KPI is full-price sell-through against weeks on floor, read as it is in apparel. The secondary KPI on both pools is attach rate — accessory units or sales relative to the host category they sell beside — because an accessories plan often moves with the apparel or footwear it is merchandised with (attach rate). Attach rate is only meaningful in weeks when the host category was in stock; a belt's attach rate falls when the trousers it was shown with sell out, and that fall is a host-category problem, not a belt problem.

Material sets the cost base and therefore the margin reading. A leather goods program priced on a material ladder — coated canvas, smooth leather, exotic finishes — carries different IMUs at different rungs, and GMROI by rung shows whether the top of the ladder earns its inventory or merely anchors the middle. Gifting peaks add a timing question: a seasonal colorway that lands after the gifting window opens has a short full-price life, and its sell-through should be read against its own weeks on floor rather than the category's. Planning accessories lines covers the core-and-seasonal split.

Home & furniture: GMROI and forward cover across the lead time

Home and furniture is a capital business with long lead times. SKUs — a sofa in a frame, a fabric grade and a finish; a dining table in a size and a wood — are bought months ahead, often against container minimums that set the order quantity as much as demand does, and the stock sits on water for weeks before it is sellable. The primary KPIs are GMROI on the owned inventory and forward weeks of supply that includes on-water stock. Trailing weeks of supply, the default in apparel, is the wrong read here: it divides today's stock by last month's sales, while the decision that matters is the next booking, which has to cover demand through an arrival date months out.

GMROI earns its place because furniture turns slowly by construction and the question leadership asks is whether the capital is earning. Its base must match ownership: where the business takes title at the origin port, goods on water are owned inventory and belong in the average, and leaving them out flatters the return. Forward cover answers the operational question — will stock arrive before the current position runs out — and it must run against the forecast through the next container's landing date, not against a trailing average.

The secondary KPIs are specific to how furniture is sold. Floor-set productivity — sales per floor sample or per floor-set slot — measures whether the range shown in stores earns the space, because a configuration that is not on the floor rarely sells from stock. The special-order share splits sales into stocked and made-to-order; special orders carry no inventory and inflate GMROI if they are counted in the numerator without a matching stock base, so read GMROI on stocked sales and track special orders as a demand signal for the next stock decision. For brands selling through dealers, dealer prebook coverage against plan is the flow KPI that tells the buyer how much of the next container is already spoken for.

The trap is reading apparel's KPIs on a furniture program. A trailing weeks-of-supply read chronically under-orders against the lead time; a sell-through read by season misreads a product that sells for several years. Container fill is a guardrail: open dollars against a half-full container do not buy the goods. OTB planning for home goods and merchandise planning for home and furniture brands carry the detail.

Outdoor: dealer sell-through against the model-year changeover

Outdoor brands plan on the model year, and most of the year's quantity is fixed at the dealer prebook close, before the selling season opens. Sell-in to specialty dealers is booked revenue for the brand; the season's real result is how much of that sell-in the dealers sell to consumers before the next model year arrives. The primary KPI is dealer sell-through, read against the model-year changeover date, because a dealer carrying last year's jacket into the new model year will discount it beside its own successor and order less of the new one.

Weather sets when the season opens, and the season opens at different weeks in different climate zones. Sell-through measured on a calendar week compares a dealer whose season has started with one whose season has not; measured from each zone's opening week, the comparison holds. The store clustering guide covers grouping dealers by the week the season opens.

The secondary KPIs are flow measures. Prebook coverage — prebook orders against the model-year plan — is the outdoor form of PO coverage and is the leading KPI before the year starts. At-once fill rate — in-season reorders shipped against reorders placed — measures whether the brand held enough uncommitted stock to support dealers who are selling through, and a low fill rate on reorders is lost demand that no sell-through figure records. Where dealers sell under a minimum advertised price policy, advertised reductions are constrained by the policy, and the closeout plan at changeover becomes the main reduction line.

The trap is counting sell-in as success. A model-year plan that hits its prebook target and leaves dealers with a heavy residual at changeover produced the year's revenue and the next year's resistance. Merchandise planning for outdoor brands and planning a model year changeover carry the detail.

Sporting goods: in-stock and sell-through inside each sport's window

Sporting goods is several seasonal businesses under one roof, each on its own calendar: the sport's season, the school term, league registration, team cycles. Baseball equipment, soccer cleats and team uniforms each have a selling window that opens and closes on dates the brand does not set. The primary KPIs are in-stock and sell-through inside each sport's window, because demand outside the window is structurally close to zero, and a category-level read blends a sport in season with one that is not.

In-stock leads because the window is short. A door out of the core sizes of a cleat in the first two weeks of the season loses the sale to a competitor, and the customer does not return when the reorder lands. Sell-through inside the window, measured from its opening week, says whether the buy was right for that sport; measured across the year, it says nothing. Zero sales in a closed window are not weak demand, and a forecast or a sell-through comparison that treats them as observations will under-buy the next window.

The secondary KPI is prebook coverage. Much of sporting goods quantity is committed through prebooks placed months ahead, with an at-once layer bought for in-season reorders, so the buyer's leading measure is how much of each sport's plan is covered by prebook and how much at-once capacity remains. Team business adds its own flow measure: fill rate against team orders, which have fixed delivery dates tied to the start of the season and little value if they arrive late. Closeouts follow the model year, as in outdoor.

The trap is annual turns. A sporting goods category turning at its annual rate can hold a sport's entire residual from a closed window, a figure that only appears when the KPI is read by sport and window. Merchandise planning for sporting goods brands covers the calendar.

Health & beauty: in-stock by shade and cover against dating

In health and beauty the grain is the shade, not the SKU family. A foundation, concealer or lip range is a ladder of shades, and demand across it is uneven: the mid-depth shades carry most of the volume, and the ends of the ladder serve customers who have few alternatives. The primary KPI is in-stock by shade, read across the full ladder at each door and online, because a range averaged across its shades can look balanced while its mid-depth shades are out and its tails are aging.

The second primary measure is cover against dating. Beauty stock carries a shelf life and, once opened, a period after opening (PAO); retailers often set acceptance and sell-by rules against the manufacture date. Weeks of supply therefore has a ceiling that apparel's does not: cover that exceeds the remaining sellable life is a write-off scheduled in advance, whatever the sell-through says. Dating rules and weeks of supply covers holding cover against an expiry rather than a season.

The secondary KPIs are productivity measures tied to how beauty is merchandised. Shade-range productivity — sales per shade across the ladder — shows which shades earn their position on the planogram and which are held for range completeness, a legitimate decision as long as it is a decision. Sales per gondola reset measures whether the space allocated at each reset earns its share, and the reset calendar is the date when shade additions and deletions take effect. Testers are a real reduction line: units consumed at counter and units written off at a dating cut-off belong in the shrink and markdown figures, not in an unexplained margin gap.

Never-out-of-stock lines — the core shades and the hero products — run on replenishment, so in-stock and weeks of supply against the reorder lead time govern them, as on any never-out-of-stock program. The trap is SKU-level reporting: a rebuy calculated on the family's average rate rebuys the tails and leaves the shades that sell out. Merchandise planning for health and beauty brands covers the shade ladder.

Toys & games: sell-through into the gifting peak and the post-peak residual

Toys and games concentrate the year into a single gifting peak in the fourth quarter. The same unit of stock is a shortage in the weeks before the peak and dead inventory in the weeks after it, so annual measures are close to meaningless as in-season signals. The primary KPIs are sell-through into the peak, read by item and account by week, and the post-peak residual: the units left in the channel when the peak ends, against the plan for how they exit.

Sell-through into the peak is the leading measure. By the time the peak weeks arrive, replenishment from overseas production is usually closed, so the read that matters is whether early-season sell-through and forward weeks of supply against the peak phasing say an item will run out early, sell through as planned or be left over — while domestic stock can still be moved between accounts. Forward cover against the peak curve, not trailing weeks of supply, is the working number; a trailing figure in October divides stock by pre-peak sales and reads as heavy when it is about to be thin.

The post-peak residual is the lagging measure that decides the next year. A heavy residual becomes the markdown allowances in each account's terms and the returns and closeout negotiation that follows. Licensed product adds a fixed end date: a licensed item has to be sold off by the contract's sell-off date, so its residual is read against that date rather than the season, and stock left past it cannot be sold under the license at all. Planning a licensed product window covers the window arithmetic.

Toy brands selling mostly to retailers also face the sell-in versus sell-through gap. Shipments are booked revenue; the retailer's sell-through to consumers decides the post-peak conversation and the next year's open orders. Where retailer point-of-sale data is shared, it is the primary read; where it is not, reorders and fill rate are proxies and should be labelled as such. Planning the gifting calendar and merchandise planning for toy and game brands cover the peak.

Baby & juvenile: in-stock on registry core over the life of a configuration

Baby and juvenile hard goods — car seats, strollers, cribs, high chairs — are long-lived products, certified in specific configurations and sold across several model years. A configuration that sells steadily for years will show a low sell-through in any single season, which is a normal reading rather than a problem. The primary KPI is in-stock on the registry core: the configurations most often added to registries, which a parent expects to be available when the registry is fulfilled. Registry demand is intent recorded ahead of purchase, and an item out of stock at fulfilment loses the sale to whatever is in stock.

Productivity is read over the life of the configuration rather than by season. A car seat shell certified in a set of configurations carries testing cost that is recovered over the units sold on it, so the useful question is whether each configuration is earning across its life and whether a short-run configuration justifies its place. Life-of-configuration sell-through — units sold since launch against units received since launch — is the secondary KPI, and it is the one that informs the changeover decision when a new chassis or shell replaces the old.

Juvenile soft goods and apparel are read differently: they are sized by age band, and the age-band mix of demand at a door decides which sizes sell. Sell-through by age band, like size-level in-stock in apparel, stops a total from hiding a gap in the bands that move. Fill rate is a guardrail on both sides: registry retailers expect orders to ship complete, and a short shipment on a registry item fails the customer at a fixed date.

The trap is applying a seasonal apparel KPI to a product that sells for years. A configuration flagged as slow on a one-season sell-through and cut from the range can be the one that registries carry. Planning with registry demand and merchandise planning for baby and juvenile brands cover the detail.

Jewelry & watches: GMROI and inventory value at replacement cost

Fine jewelry is a piece-level business at low velocity, built on a precious metal cost base that moves with the market. A ring's cost is its metal content priced by weight and purity, its stones, and its bench labor and findings, and only the first of those moves daily. Turns are low by construction, so speed alone says little. The primary KPIs are GMROI and inventory value, both held at replacement cost: what the pieces in the case would cost to make today, at today's metal price.

Replacement cost matters because the case holds pieces made at different metal prices. A piece made when metal was cheaper shows a healthy margin on its historical cost; its replenishment will be costed at the current price. GMROI computed on historical cost therefore describes a business that no longer exists. Tracking metal exposure beside it — the metal content of the inventory in grams by purity, valued at the current price — tells leadership how much of the inventory value is a position in the metal market rather than a merchandise decision.

The secondary KPIs follow how the category sells. Sales per piece and sales per design family show which designs earn their place in the case; with depth of one or two pieces per door, sell-through is a piece-level yes or no rather than a rate, and the useful read is time to sell by price band. Memo stock — pieces placed with a retailer on consignment and invoiced only when sold — stays in the brand's inventory and depresses turns without producing a sale, so memo and owned stock should be reported separately. Watches add the reference and the brand's annual novelty cycle: an authorized dealer's planning variable is which references to hold and how deep, and a reference the brand discontinues is the category's closeout.

Markdown percent is the wrong guardrail here. Fine jewelry exits through remount, melt, consolidation between doors and private sale rather than public price reductions, so a low markdown figure says nothing about whether the inventory is healthy; aging by piece and the value of pieces past a set age does. Merchandise planning for jewelry and watch brands covers the metal and memo detail.

How KPI sets go wrong

A KPI set fails in four predictable ways. Each passes every arithmetic check, because the formulas balance on a wrong base as readily as on a right one.

Measuring sell-through without receipt timing

Sell-through divides sales by receipts, and receipts arrive in drops. Read the worked example's class at the end of week 4: it has sold 1,120 units against the 2,400 received, a 46.7% sell-through. Measured against the full 4,000-unit buy — including the 1,600 units that have not yet landed — the same week reads 28.0%, and a trade meeting comparing that figure with plan would mark the class down for stock it did not yet have. The reverse error happens when a drop lands late: sell-through on the units received looks strong while the class is short of the stock the plan assumed, and the lost sales never appear in any rate. Read sell-through on the units actually received to date, and read receipts against plan beside it, so a sell-through movement caused by flow is never mistaken for demand.

Averaging percentages

Percentages roll up by summing their numerators and denominators, never by averaging the rates. Add a second class to the example: it receives 400 units and sells 360, a 90.0% sell-through, beside the knit tops at 60.0%. The simple average of the two rates is 75.0%. The category's actual sell-through is (2,400 + 360) ÷ (4,000 + 400) = 2,760 ÷ 4,400 = 62.7%. The averaged figure gives a class one-tenth the size equal weight and overstates the category by more than twelve points. The same rule applies to in-stock, fill rate, markdown percent, gross margin and GMROI: roll up the dollars and units, then divide.

GMROI on the wrong inventory base

GMROI's denominator is average inventory at cost, and every shortcut changes the answer. On the worked example's $66,000 of gross margin:

All four divisions are arithmetically correct; only the first answers the question GMROI asks. The two-point average more than doubles the return because its opening point is taken before the first drop lands, so it never sees the stock the class actually carried through the eight weeks. Ending inventory flatters any period that sold down; a retail base mixes a retail denominator with a margin numerator and understates the return. The other base errors are vertical: on-water stock left out in furniture, memo stock left in or out without saying so in jewelry, historical cost on metal that has moved. State the base in the report header, and compare GMROI only between periods of the same length on the same base.

A KPI nobody can act on

The quietest failure is a measure that is accurate, reported every week and owned by no one. Comp sales reviewed in a trade meeting that cannot move stock between doors; sales per square foot reviewed by a planning team that does not decide the space — a boundary space planning vs merchandise planning sets out; forecast accuracy shown weekly at a total where no forecast decision is made. A KPI earns its place in a meeting only if someone in the room can change it this week. Move the others to the meeting where their owner sits, or drop them from the set.

Building the set for a category

The sequence is short, and each step is cheaper to change than the one after it.

1. →Name the binding constraint — the scarce resource whose shortfall decides the season, with a lever someone owns: time at full price, the size run, the lead time and capital, the shade ladder and dating, the peak, the registry date, the metal price. The KPI that measures it is the primary one.
2. →Write the formula with its denominator, grain and period. Receipts or total available for sell-through; trailing or forward for cover; average at cost, over how many points, for GMROI.
3. →Add two or three secondary KPIs that explain the primary one, and the guardrails that stop it being gamed — in-stock against a sell-through push, markdown percent against a clearance, aging against a cover target.
4. →Assign the cadence and the owner. Leading KPIs to the weekly trade meeting with the role that can act; lagging KPIs to the monthly review and the hindsight.
5. →Set the targets from the plan, not from last year's actuals. A target copied from history carries last year's assortment, receipt timing and markdown plan into a season that may not share them. Where a target range is used as a planning starting point rather than a computed figure, label it as illustrative.

Where RetailNorthstar fits

RetailNorthstar is a merchandise planning platform — OTB, assortment and line planning, buy planning, allocation, sizing, PO and WIP tracking and analytics on a shared data model. As retail math for merchandise planners sets out, it holds the merchandise financial plan, assortment planning, buy planning, open-to-buy and allocation on one shared data model, so margin, inventory productivity, sell-down and flow measures compute from the same sales, cost, stock and on-order lines rather than from separate exports. As the market and performance intelligence page puts it, velocity, weeks of supply and sell-through rate are shown at the point of decision, not in a separate BI report.

Choosing the primary KPI for each category, setting its target and deciding what the trade meeting does with it remain judgements that belong to the merchant and the planning team. Merchandising planning holds the financial plan the lagging KPIs are measured against, and OTB planning holds the receipt budget the flow KPIs track.

Related resources

• →Retail Formulas — most formulas in this guide on their own page, with a calculator
• →Retail Math for Merchandise Planners — the derivations behind each formula
• →The Weekly Trade Meeting — where the leading KPIs become decisions
• →How to Hindsight a Season — where the lagging KPIs become next season's buy rules
• →SKU Productivity Calculator — net sales per option and sell-through by category
• →Markup and Pricing by Vertical and Markdown and Exit Strategy by Vertical — the margin KPIs in practice
• →In-Stock Rate, Sell-Through Rate, Weeks of Supply and GMROI — the terms