Monday's report changes a replenishment call
A merchant sees a new jacket drop sell 270 units from a receipt of 600 in its first month: 45% sell-through. That can support a reorder, price hold, or review of weak sizes only when sales are tied to the right stock intake. If some of the 270 sales came from an older delivery, 45% describes trading activity, not the new drop's performance.
Sell-through rate measures how much received inventory sold within a defined period or receipt cohort: demand against intake. Unlike a revenue report, it focuses on units, availability, and capital that could remain tied up in stock requiring later markdowns.
The basic formula is:
Sell-through rate = (Units sold / Units received) × 100
Use units from the same scope: SKU, colour-size variant, collection, supplier delivery, channel, or store group. Calculation is simple; selecting a defensible denominator is harder.
Receipts define the denominator
Count units when physically received, quality-checked, and available for sale. A purchase order is not a receipt, nor is stock in customs, failing inspection, or not yet at a fulfilment location. Including unavailable units makes a product appear slower than customers had a fair chance to buy.
For a seasonal collection, a receipt cohort is usually the cleanest denominator: divide sales from that delivery by units received in it. For replenishable basics, teams may use a rolling receipt cohort or period-flow measure, dividing weekly sales by weekly receipts. The latter monitors intake pace but misleads when delivery timing is uneven.
A week with no receipt and steady sales produces an undefined or inflated flow figure, not infinite demand: numerator and denominator describe different inventory flows. Keep a cohort measure beside any weekly operating report.
Inventory systems need shared definitions for receipt, transfer, cancellation, shipment, return, and write-off. A receipt, sale, and return event taxonomy stops analysts rebuilding the metric differently in every dashboard.
Calculate one cohort before aggregating
Assume a retailer receives 600 jackets on 1 September and customers buy 270 during September. Gross sell-through is:
(270 / 600) × 100 = 45%
If 18 are returned and restocked in sellable condition during the agreed return window, net units sold are 252:
270 gross units sold - 18 returned units = 252 net units sold
Return-adjusted sell-through is:
(252 / 600) × 100 = 42%
Both measures matter. Gross sell-through shows checkout demand; net sell-through shows realised demand after returns. Do not subtract returns from a current cohort when they belong to a prior delivery; match them to the original sale or state that the figure is provisional.
Returns that cannot be resold should not re-enter available inventory. They still reduce net sales, while their disposition belongs in a separate damage or write-off record. Blending the cases obscures weak demand, poor fit, or product quality.
Three clocks answer different questions
Weekly, monthly, and season-to-date sell-through are not interchangeable; each supports a different decision.
Weekly sell-through catches early demand changes: a campaign lift in one colour, a breaking size run, or stockouts suppressing sales. It is volatile: a creator mention, public holiday, late delivery, or single wholesale order can distort seven days.
Monthly sell-through is steadier for trading reviews and category planning, giving paid media, email, store traffic, and replenishment time to affect demand. It can hide a sharp final-week decline, so pair it with weekly trend lines rather than replace them.
Season-to-date sell-through asks whether a collection is moving fast enough to clear before its selling window ends. It suits spring assortments, holiday gifting ranges, and limited collaborations. A high season-to-date rate can coexist with a weak current week after an early launch spike.
Label every rate with its period: “week 4 from launch,” “September receipts,” or “season-to-date through 30 November.” Comparing a four-week result with a season-to-date result creates a false narrative even when both calculations are correct.
Turnover and stock cover are not substitutes
Sell-through, inventory turnover, and weeks of supply answer separate questions. Treating them as synonyms produces poor purchase orders.
| Metric | Core calculation | Decision it supports | Main limitation |
|---|---|---|---|
| Sell-through | Units sold / units received | Is this delivery or collection converting into sales? | Depends on matching sales and receipt scope |
| Inventory turnover | Cost of goods sold / average inventory at cost | How quickly is inventory capital cycling? | Can hide weak individual styles inside a broad category |
| Weeks of supply | Usable on-hand units / average weekly unit sales | How long will stock last at the current sales pace? | Breaks when recent demand is abnormal or stockouts limited sales |
A fast-selling SKU can have high sell-through yet excessive stock cover after a large reorder. A mature evergreen product can have modest sell-through on a recent delivery while posting strong turnover across the year. Turnover is a financial, portfolio-level lens; sell-through is closer to an item-level intake decision.
Weeks of supply looks forward; sell-through looks back at a receipt cohort. Use stock cover to decide whether to replenish, then cohort sell-through to judge whether the prior buy was sized correctly.
Healthy bands follow product life cycles
No healthy sell-through percentage fits every category. A 30% rate can be alarming for a limited fashion drop two weeks before planned markdown, yet acceptable for furniture, luggage, or equipment with a longer consideration cycle and slower replenishment plan.
Teams often begin with planning bands, not universal benchmarks. For fast-fashion or short-drop products, a typical first-four-week band is around 40-70%, with a higher target by the planned end of the selling window. For durable goods, a first-four-week band around 15-35% can fit a healthy longer cycle if traffic, margin, and stock cover support it. These are heuristics, not evidence that one percentage fits every retailer.
Set internal targets by product life, lead time, price point, return behaviour, and channel. A premium coat needs a different pace than a low-cost T-shirt, and a SKU with a twelve-week supplier lead time needs a different reorder trigger than one replenished in days.
Variant detail matters. A collection can look healthy because one colour sells out while two stagnate. Aggregate sell-through can delay a colour-specific markdown and leave only broken size runs for full-price customers.
Use the signal to price and buy
Sell-through matters when linked to a decision rule, not left in a weekly deck. Start with the planned selling window, gross margin, available and incoming stock, lead time, and current stock cover. Then classify items by demand and inventory risk.
High sell-through with low weeks of supply is a reorder candidate when margin allows, the supplier can deliver before demand fades, and returns are not eroding realised sales. Reorder selling variants, not the assortment average. If size M is depleted while other sizes remain, buying another full size pack can magnify the problem.
High sell-through with high returns requires diagnosis before expanding the buy. Strong imagery or promotion may attract demand while fit, quality, or expectations fail after delivery. Check return reason codes, reviews, and whether returned units are resellable.
Low sell-through with normal traffic can call for a controlled markdown test, revised product page, bundle, or placement change. Low sell-through with low traffic is not yet a pricing verdict: the product may be buried in navigation, excluded from campaigns, unavailable in requested sizes, or shown after an out-of-stock period.
Discount only when expected contribution is better than holding stock. Track sell-through with selling price, product cost, fulfilment cost, return cost, and paid acquisition cost. Unit economics as a systems audit examines this link between inventory pace and profit, treating a sale with weak contribution as a system issue rather than a win.
Promotions need a comparison period or holdout where feasible. A higher rate during a campaign may reflect seasonality, a new channel, or stock arrival rather than the discount. Apply the logic behind testing promotional incrementality before crediting every sold unit to the offer.
Bad data creates false winners
The commonest error divides sales by on-hand inventory. On-hand is a snapshot after sales, transfers, and returns; receipts are historical intake against which sell-through is measured. Substituting one can make an almost-sold-out SKU appear to have sold more than 100% of its stock.
State return treatment. Use gross sell-through for an early demand read, then return-adjusted sell-through after enough time for the return window to mature. Do not mix gross results for one category with net results for another.
Stockouts are another trap. An item at 80% can look stronger than one at 55%, though the first was unavailable for half the period. Record in-stock days and lost availability before declaring a winner. Transfers, marketplace cancellations, bundles, and units reserved for wholesale create similar distortions.
Compare like with like. A monthly rate for a new receipt cannot be judged against a season-to-date rate for an older cohort. Every report should include date range, channel, unit status, return rule, and cohort definition.
Choose the cadence that matches risk
Use weekly cohort sell-through when a fashion drop, limited stock position, or short lead time requires fast intervention. Use monthly readings when weekly noise would trigger needless price changes, and season-to-date sell-through for exit plans and seasonal open-to-buy decisions.
Reorder when sell-through is ahead of plan, stock cover is tightening, return-adjusted demand is sound, and replenishment can arrive before the selling window closes. Otherwise protect cash: diagnose availability and variant issues, then test markdown depth or reduce the next buy instead of treating every low rate as demand failure.