Skip to main content
37 min readmarkdown optimizationmarkdown optimization in retail

Markdown Optimization in Retail: Strategy, Triggers and Measurement

How to time, size and target markdowns so the season clears by its exit date: causes, five decision variables, trigger tables, a worked cadence and the KPIs.

Markdown optimization is the planning discipline that decides when to reduce price on unsold inventory, how deep each reduction goes, in which doors and channels, on which sizes and on what cadence, so that the season clears by its exit date for the least margin given up. It works at two points in the calendar. Before the season, it reduces how much inventory will ever need a markdown by getting buy depth, receipt timing, size curves and assortment breadth right. In the season, it chooses the cheapest way to clear whatever the plan still got wrong. This guide covers the discipline and its tactics; for the software category, what a markdown optimizer decides, how its price-response model and optimizer work and how to choose one, read What Is Markdown Optimization Software?.

What is markdown optimization?

Markdown optimization is the set of planning decisions and in-season actions that minimize the depth, frequency and volume of price reductions on unsold inventory while still clearing the season on time. In apparel, where seasonal product has a finite selling window, some markdown is unavoidable. The amount of inventory that needs marking down, though, is largely a planning outcome rather than a demand outcome, and that is the part a planning team controls.

Markdown optimization therefore has two halves that use the same numbers:

  • Preventive markdown planning sets the buy so that less inventory reaches markdown at all. Its levers are pre-season: the depth bought per option, the number of options, the size curve, the receipt flow and the chase reserve. It is measured in the open-to-buy and the margin plan.
  • In-season markdown optimization decides what to do with the inventory that is not selling at full price. Its levers are price, place and time: the depth and timing of each step, the doors and channels it applies in, and the sizes it is restricted to. It is measured in sell-through at exit and margin dollars kept.

Every markdown dollar comes directly out of the margin planned at the initial markup. The difference between a brand that plans markdown and one that reacts to it is rarely merchant talent; it is whether the triggers, depth steps and exit routes were agreed before the season or improvised in week nine.

Markdown strategy vs markdown optimization vs clearance vs promotion

Four terms get used interchangeably in trade meetings, and the confusion costs margin, because each one has a different owner, a different horizon and a different measure. A temporary promotion booked as a markdown hides a margin problem; a clearance taken as a promotion trains customers to wait.

TermWhat it isHorizonPrice changeMeasured by
Markdown strategyThe policy: triggers, depth steps, exit routes and owners, agreed by classificationSet pre-seasonRules, not pricesAdherence to triggers
Markdown optimizationChoosing the best action per style inside that policy, week by weekIn-season, weeklyPermanent reductionsMargin dollars kept, sell-through at exit
ClearanceThe final price stage before inventory leaves by another routeLast weeks of lifeDeepest, finalResidual units and recovery per unit
PromotionA temporary price event to drive traffic or a sales weekDays to weeksTemporary, revertsIncremental sales and margin per event

Three distinctions matter in practice:

  1. Permanent versus temporary. A markdown is a permanent ticket reduction; a promotion reverts. Promotions are usually booked to allowances rather than markdown dollars, which is why the markdown percent formula counts permanent reductions only. A sitewide code every other week is a markdown problem dressed as a promotion calendar.
  2. Policy versus decision. The strategy is written once per season; optimization happens inside it. When the trade meeting debates every style from scratch, there is no strategy.
  3. Clearance is a stage, not a strategy. If a large share of a season reaches clearance, the fix is in the buy, not a better clearance price.

Markdown cadence sits between the strategy and the optimization: it is the planned sequence of first touch, deepening steps and exit point that the strategy prescribes and the weekly optimization executes.

The real causes of markdowns (it's not weak demand)

Most merchants attribute markdowns to soft demand or a trend that turned. Our position is different: the primary drivers of markdown exposure are upstream planning decisions, and each of them is visible in pre-season data before the first unit ships. Demand misses happen, but they are smaller and rarer than the post-season story suggests, because a buy that was too deep, too late, in the wrong sizes or spread across too many options produces exactly the same symptom as weak demand — slow sell-through and a pile of residual stock.

1. Buy depth: over-buying the units

The most common cause is buying more units than the market can absorb at full price within the selling window. Over-buying happens when:

  • OTB budgets are set optimistically without checking the planned sell-through against what the category has actually delivered
  • The planning team adds "just in case" units instead of holding a chase reserve that is only released on evidence
  • Vendor minimums force buys above demand, and the minimum is accepted rather than consolidated across colors
  • Last season's best seller is bought to its peak rather than to its full-price sell-through

The depth error is the most expensive because it is multiplied by every unit: the style sells at its natural rate and leaves the excess for markdown. The fix sits in the open-to-buy: validate depth per option against sell-through capacity before the PO is placed, not after.

2. Receipt timing: late delivery and front-loaded flow

A style that arrives three weeks late into a twelve-week selling season has lost 25 percent of its selling window. The sell-through arithmetic no longer works at the planned depth: the brand either marks down early or carries the inventory past its window at full margin risk.

Timing errors come in two forms. Late receipts shorten the window for the units that arrive. Front-loaded receipts put the whole buy on the floor in week one, before the first sell-through read, so nothing can be held back when the read is weak. A receipt flow split into an initial and a follow-on drop gives the planner a decision point.

3. Size-curve error: the broken run

Size residuals are among the largest and most preventable sources of markdown inventory. If 15 percent of a buy is allocated to sizes that historically sell through at 40 percent, most of those units are destined for markdown before the season starts. The style can be selling well in its core sizes while the fringes accumulate.

Size-curve error is also the cause markdown is worst at fixing. A broken run — medium and large sold out, extra-small and extra-extra-large left — does not respond to price, because the customer who wanted a medium is not persuaded by a discount on an extra-small. See the size curve entry and the size and pack optimization guide for the upstream fix.

4. Assortment breadth: too many options competing

When two styles in the same category compete for the same customer occasion, they cannibalize each other and neither sells through cleanly. The result is markdown on both, where one well-bought style would have sold through at full price. Breadth also dilutes depth: a fixed budget spread across more options gives each one less inventory, which produces early stockouts on the winners and residuals on everything else.

This is the breadth versus depth tradeoff, decided in the line plan, not the markdown meeting.

5. Misaligned price architecture

When a brand's price points do not match the customer's willingness to pay, no amount of merchandising talent fixes it. A $180 dress in a market that wants $120 will not sell through at full price, and the markdown of a third that brings it to $120 is simply the market correcting the ticket. The tell is a style whose sell-through jumps sharply on a shallow first step: it was never a demand problem, it was a price problem.

How to tell which cause you are looking at

The same slow sell-through has different fixes depending on its cause, and the markdown decision should differ too. A quick diagnostic during the trade meeting:

Signal on the styleLikely causeMarkdown response
Slow across all sizes and doorsDepth or priceStep the whole style on trigger
Core sizes out, fringes leftSize-curve errorSize-restricted markdown; fix the curve next season
Strong in some doors, dead in othersAllocationTransfer before marking down
Arrived late, selling at plan rateReceipt timingRe-plan exit date; shallow step
Twin style selling, this one notBreadthStep the weaker twin only
Jumps on a shallow stepPrice architectureHold the new price; re-ticket next season

The point is to take the right markdown and attribute the residual to its real cause in the hindsight. Cutting depth to compensate for a size-curve error keeps the error and adds a stockout; the method for keeping causes apart is in decomposing a plan miss.

Prevention: the pre-season playbook

Preventive markdown planning is the larger half of markdown optimization, because it acts on the whole buy rather than on the residual.

Set realistic sell-through targets

Do not plan for 80 percent sell-through if a category has historically delivered 68 percent. Use your own rates by category and channel, adjusted modestly for changes you can name. A target that has never been achieved is a markdown budget nobody wrote down.

Build a markdown reserve into the margin plan

Plan for the markdown you expect to take. If a category has typically marked down a fifth of its units at an average reduction of about a third in your own history, build that into the maintained markup projection. Brands that plan to a zero-markdown scenario are planning to be surprised. The opposite failure is also real: a reserve set too generously normalizes the over-buy, so the plan builds in the failure and the failure keeps happening. The discipline is a reserve that matches history and a buy that is tightened when actuals drift.

Use a chase reserve for uncertain styles

Instead of buying full projected demand upfront, hold back part of the buy — 15 to 20 percent is an illustrative planning starting point, not a benchmark — as a chase reserve: committed capacity at the factory that is activated only when early sell-through confirms demand. This moves the risk from "hope it sells" to "prove it sells." The chase only works if its last responsible week is known; the in-season chase guide covers how to compute it.

Right-size the assortment

Fewer, deeper styles generally clear more cleanly at full price than broad, shallow assortments, because each option stays in stock in its core sizes and fewer twins compete for the same occasion. The test is whether each option has enough depth to cover its selling window without relying on markdown to clear the tail.

Pre-decide the exit for each classification

Core, fashion and test styles should each carry an agreed exit route before the season starts — carry forward, outlet, off-price or donate — so the end of the season is execution rather than negotiation. The end-of-season exit strategies guide sets out the five exits and their costs.

The five decision variables of markdown optimization

In-season markdown optimization is often described as "when and how much." In practice a planner is choosing five things at once, and getting four right does not save a style where the fifth is wrong.

1. When: the trigger and the week

Timing is the variable that matters most, because a markdown's effect depends on the traffic the style still has. A shallow reduction taken in week seven, while the style is still on the main floor and on the site's new-in pages, moves units that the same reduction in week eleven will not. The timing decision should be driven by a trigger — sell-through or weeks of supply against a threshold at a defined week of life — rather than by a calendar date applied to every style.

Week of life matters more than calendar week. A style received in week three of the season is in its fourth week of life in week six, and judging it against a style received in week one penalizes it for the receipt date rather than for its demand.

2. How deep: the size of the step

The depth of a step is the reduction needed to bring the selling rate up to the rate that clears the position by the exit date. That is a calculation, not a feel:

  • Required weekly rate: units on hand ÷ weeks remaining to exit
  • Current weekly rate: trailing average of recent weekly unit sales
  • Required lift: required rate ÷ current rate

The depth is then the smallest step on the brand's price ladder that the planner expects to deliver the required lift. That expectation is the whole decision, and it is where price elasticity enters: the more price-responsive the style, the shallower the step needed. A style that needs a lift of 1.5 times on a price-responsive fashion item may get there with a modest first step; the same lift on a basic that customers buy on need may not come at any depth worth taking. The price elasticity and promotion planning by vertical guide covers how that response differs across categories.

3. Where: doors and channels

A markdown does not have to apply everywhere. The style may be dead in a cluster of doors and selling in others, in which case a transfer is cheaper than a reduction. Channels carry their own cadence and exit routes, and running one blended markdown across full-price DTC, stores, outlet and wholesale usually lets one channel's clearance pressure set the price everywhere. The options, in roughly increasing cost:

  • Transfer units from slow doors to doors still selling at full price
  • Mark down in the doors or channel where the residual sits, hold elsewhere
  • Move the residual to outlet or off-price and hold the owned full-price channels
  • Mark down across every channel

Each has a cost: transfers carry freight and handling, and channel-specific prices need governance so the customer is not shown two prices for the same item.

4. Which sizes: whole run or the fringe

When the cause is a size-curve error, a whole-style markdown gives away margin on the sizes still selling at full price. A size-restricted markdown — only the fringe sizes, or only the sizes above a weeks-of-supply limit — targets the units that are actually stuck. It is operationally harder, especially in stores where a price is ticketed by style, and it is easier in DTC, where price can be set by SKU. The decision is whether the margin saved on the core sizes is worth the execution cost; on a style with a badly broken run it usually is.

5. Cadence: the sequence of steps

Cadence is the schedule of first touch, deepening steps and exit point. Steps taken too close together train customers to wait for the next one; steps too far apart leave stock ageing between decisions. A defined exit point stops the cadence deepening forever, spending margin on units price was never going to clear. Cadence is planned by classification and channel and then executed against triggers; the markdown cadence entry describes the three parts.

The rules-based approach: triggers by week of life

The rules-based approach is how most apparel planning teams run markdowns, and for good reason: it is transparent, it needs only data every team already has, and it can be governed in a single trade meeting.

How it works

The team agrees a trigger table per classification before the season. At each checkpoint in a style's life, the planner compares cumulative sell-through and forward weeks of supply to the table. A style that meets its threshold holds price. A style that misses takes the next step on the cadence. The weekly trade meeting confirms which triggers fired rather than debating each style.

Two tests carry the decision:

  1. Sell-through against the threshold for the week of life. Cumulative units sold ÷ units received, compared with the planned trajectory. This answers "is the style on its path?"
  2. Weeks of supply against weeks remaining. Units on hand ÷ trailing average weekly sales, compared with the weeks left to the exit date. This answers "will the style clear at its current price?" A style with more weeks of cover than weeks of selling left will not clear without a change.

The second test is the more important, because it accounts for time. A style at 35 percent sell-through with ten weeks left may be fine; the same sell-through with four weeks left is not.

An illustrative trigger table

The table below is an illustrative planning starting point for a fashion classification with a fourteen-week life — not a benchmark. Thresholds and depths should be set from your own history by classification and channel.

Week of lifeSell-through triggerCover testAction if missed
425%—Flag; check sizes, doors, transfers
640%WOS above weeks leftFirst step, 30% off
1070%WOS above weeks leftSecond step, 50% off
1385%—Final step or route to exit

Week four is a review, not a markdown: it is too early for a reliable price decision, but early enough to fix an allocation or a size problem with a transfer. The first price step sits where the read is reliable and the style still has traffic. Basics carry a gentler table, or none, because their exit is a carry-forward rather than a clearance.

Data requirements

Weekly unit sales and on-hand inventory by style-color (ideally by size and door), receipt dates so week of life can be computed, the exit date per classification and channel, and the agreed price ladder. Every team that runs a weekly sales report has it.

Strengths and limits

The rules-based approach is easy to explain to merchants and finance, easy to audit after the season and hard to game, because the trigger is set before anyone knows which styles will miss. Its limit is that it treats every style in a classification the same. A highly price-responsive style and a style customers buy on need both get 30 percent at week six, so one is over-discounted and the other is under-discounted. Teams usually tighten this by splitting classifications more finely — fashion tops versus fashion knitwear, for example — before reaching for a model.

The model-based approach: elasticity under constraints

The model-based approach replaces the fixed step with a calculated one. Instead of "missed the week-six trigger, take 30 percent," it asks: given how this style responds to price, what sequence of prices from now to the exit date keeps the most margin dollars while clearing the position?

How it works

A model-based markdown process has three parts:

  1. Estimate the price response. From historical sales at different prices, estimate how unit sales change when price changes for this style or, more often, for a group of similar styles, since a single style rarely has enough price history of its own. The estimate has to separate the price effect from everything else moving at the same time: the natural decay of a seasonal style's selling rate, the seasonality of the category, promotions, and stockouts that capped sales in some weeks. Price elasticity is the summary of that response.
  2. Forecast forward at each candidate price path. For each sequence of prices the brand is willing to show, project weekly units to the exit date, including the decay in the style's base rate.
  3. Choose the path under constraints. Pick the path that maximizes margin dollars — full-price sales, markdown sales and the value of any residual at its exit route, against a unit cost that is already sunk — subject to the rules the business sets.

The constraints are what make it a planning tool rather than a pricing experiment:

  • Exit date: the position must clear, or reach a planned residual, by a fixed week
  • Margin floor: no step below a minimum margin per unit or a minimum season margin rate
  • Price ladder: only the price endings and step sizes the brand shows
  • Step limits: a maximum number of markdowns per style and a minimum interval between them
  • Channel rules: price parity between stores and DTC, or a defined gap to outlet
  • Wholesale commitments: no owned-channel reduction below what accounts were promised, and any markdown money owed to accounts counted as part of the cost

Data requirements

The model-based approach needs everything the rules-based one needs, plus:

  • Clean price history: every price a style actually sold at, by week and by channel, with permanent markdowns separated from temporary promotions
  • Promotion and event calendar: so promotional weeks are not read as the response to a permanent markdown
  • Stockout and broken-run flags: so weeks where sales were capped by availability are not read as low demand
  • Product attributes and several seasons of history: the basis for pooling similar styles, since response to price is noisy and a single style has little history

Where it fails

A model-based markdown is only as good as its elasticity estimate, and the estimate fails in predictable ways. Promotions overlapping markdowns make the response look larger than it is. Broken size runs make it look smaller. New styles with no history inherit the response of the group they were pooled into, which may be wrong for a style that is genuinely new. And a model optimizing margin dollars for the season will happily discount early and deep if its estimate says that pays, which may conflict with the brand's pricing position. The constraints are not optional extras; they are where the brand's judgment lives in the model.

Rules-based versus model-based

Rules-basedModel-based
Decision unitClassificationStyle or style group
DepthFixed steps per triggerCalculated per style
Data neededWeekly sales, stock, receiptsPlus clean price and promo history
TransparencyHigh; anyone can audit itLower; needs explanation
Main riskSame step for unlike stylesWrong elasticity estimate
GovernanceTrigger table in trade meetingConstraints and override rules

Many teams run a hybrid: the trigger table decides when a style is reviewed and sets the default step, and an estimate of price response adjusts the depth for styles with enough history to trust it. Whichever approach is used, the triggers, constraints and override rules should be agreed before the season, and every override recorded so the hindsight can tell whether the overrides helped.

Worked example: one style-color through a markdown cadence

The example follows one fashion style-color through a fourteen-week life against the illustrative trigger table above. All figures are invented for illustration, and the weekly sales after each markdown are assumptions about price response — the point is the arithmetic, not the numbers.

The style is a women's linen shirt in one color.

  • Units received: 1,200, all in week one
  • Full price: $80; unit cost: $32, so initial markup is ($80 − $32) ÷ $80 = 60%
  • Exit date: end of week 14; any units left then go to an off-price buyer at an assumed $16 per unit
  • Total cost of the buy: 1,200 × $32 = $38,400

Weeks 1 to 6 at full price

WeekUnits soldCumulativeSell-through
170705.8%
27514512.1%
37021517.9%
46528023.3%
56034028.3%
66040033.3%

Week 4 check

Sell-through is 23.3 percent against a 25 percent trigger. The style is flagged. The planner checks sizes and doors and finds the shortfall is spread evenly across both, so it is not a size or allocation problem and there is nothing to transfer.

Week 6 check

Sell-through is 33.3 percent against a 40 percent trigger, so the first test fails. For the cover test: 800 units are on hand (1,200 − 400). The trailing four-week average is (70 + 65 + 60 + 60) ÷ 4 = 63.75 units a week, so weeks of supply are 800 ÷ 63.75 ≈ 12.5 weeks. Weeks left to the exit date are 8. Cover exceeds time, so the second test fails too. Both triggers have fired.

Sizing the first step

To clear 800 units in 8 weeks at a single price, the style would need 800 ÷ 8 = 100 units a week, against a recent rate of about 60 to 64. The required lift is roughly 1.6 to 1.7 times. The trigger table's first step is 30 percent off, taking the price to $56. The planner's assumption is that the rate rises to 110 units in week seven and then decays as the season ages.

Three ways to play the remaining eight weeks

Scenario A: hold full price to the exit date

The style keeps selling at its decaying rate: 55, 50, 45, 40, 35, 30, 25 and 20 units in weeks 7 to 14, a total of 300.

Scenario B: take the first step on trigger

30 percent off from week 7: 110, 100, 90 and 80 units in weeks 7 to 10, a total of 380. At the week-10 check, cumulative sales are 400 + 380 = 780, a sell-through of 65.0 percent against a 70 percent trigger; 420 units are on hand against a trailing four-week rate of 95, so cover is about 4.4 weeks against 4 weeks left. Both tests fail again, so the second step takes the price to 50 percent off, $40: 90, 80, 70 and 60 units in weeks 11 to 14, a total of 300. At the week-13 check, cumulative sales are 1,020, exactly the 85 percent trigger, so no further step is taken.

Scenario C: hold, then go deep late

Full price through week 10 (55, 50, 45 and 40 units, a total of 190, taking cumulative sales to 590 or 49.2 percent), then straight to 50 percent off for weeks 11 to 14: 95, 85, 75 and 65 units, a total of 320.

The result

A: HoldB: Step on triggerC: Late and deep
Units sold at $80700400590
Units sold at $5603800
Units sold at $400300320
Residual to off-price500120290
Owned-channel sales$56,000$65,280$60,000
Off-price recovery$8,000$1,920$4,640
Total revenue$64,000$67,200$64,640
Gross margin dollars$25,600$28,800$26,240
Gross margin %40.0%42.9%40.6%

The components, recomputed:

  • A: 700 × $80 = $56,000; 500 × $16 = $8,000; total $64,000; less $38,400 cost = $25,600
  • B: 400 × $80 = $32,000; 380 × $56 = $21,280; 300 × $40 = $12,000; owned total $65,280; 120 × $16 = $1,920; total $67,200; less $38,400 = $28,800
  • C: 590 × $80 = $47,200; 320 × $40 = $12,800; owned total $60,000; 290 × $16 = $4,640; total $64,640; less $38,400 = $26,240

In every scenario the units add to 1,200: A is 700 + 500, B is 400 + 380 + 300 + 120, C is 590 + 320 + 290.

What the example shows

Stepping on trigger kept $3,200 more margin than holding, and $2,560 more than the late, deep markdown. The reason is visible in the residual: B moved 380 more units through the owned channels than A, each one sold for well above the $16 it would have fetched at exit. C waited until the style had little traffic left, so even a 50 percent reduction moved fewer units than B's two steps.

The markdown decision has a simple break-even. Every unit sold at the markdown price that would have sold at full price anyway gives up the step (here $80 − $56 = $24). Every extra unit the markdown moves that would otherwise have gone to off-price gains the gap to the exit price (here $56 − $16 = $40). The step pays when:

incremental units ÷ units that would have sold anyway > (full price − markdown price) ÷ (markdown price − exit recovery per unit)

For the first step that ratio is $24 ÷ $40 = 0.6: the markdown needs to move at least 0.6 extra units for every full-price unit it discounts. In this example it did comfortably. On a style customers buy on need, it might not, and holding would be right. That is the honest limit of the example: the answer depends entirely on the assumed response to price, which is why the rules-based team watches the week after the step closely and the model-based team invests in the elasticity estimate.

The example also shows why markdown percent alone misleads, which the measurement section takes up below.

In-season: timing and depth principles

The early markdown advantage

Conventional wisdom says hold the price as long as possible. For a style that has missed its triggers, the example says otherwise. The first touch does most of the work, because it lands while the smallest reduction still moves units. Depth cannot buy back time: once traffic has moved on, a deeper step moves fewer units per point of discount. And holding is a decision too — a style held past its triggers has chosen a larger residual at the exit price.

When holding price is right

Early markdown is not universal. Hold price when the style is meeting its cover test even though sell-through is behind — a late receipt with a shifted exit date is the common case — when the residual has a good carry-forward route because the style is a core item, or when the brand's pricing position means a visible reduction costs more than the margin it recovers. In those cases the right action is often a transfer, a size-restricted step or a planned move to outlet rather than a ticket change. Dynamic pricing for fashion covers why frequent price movement is risky for apparel brands specifically.

Measuring markdown performance

Markdown performance is measured on five numbers, each with a formula, read together. Any one of them alone can be made to look good while the season loses margin.

1. Markdown percent of sales

Markdown % = markdown dollars ÷ net sales × 100

Markdown dollars are the permanent reductions taken: (original price − selling price) × units sold at the reduced price. In the example, using owned-channel sales and counting the off-price residual separately as an exit rather than a markdown:

  • A: $0 ÷ $56,000 = 0.0%
  • B: ($24 × 380) + ($40 × 300) = $9,120 + $12,000 = $21,120; $21,120 ÷ $65,280 = 32.4%
  • C: $40 × 320 = $12,800; $12,800 ÷ $60,000 = 21.3%

Scenario A has the best markdown percent and the worst margin. That is the single most important point about this metric: it measures the price reduction taken, not the margin kept, and a team managed on markdown percent alone will hold price and push the cost into the residual. The markdown percent formula page and the markdown rate entry cover the definition and its conventions. State your convention for off-price sales and wholesale markdown money explicitly, because it changes the number.

2. Markdown dollars against plan

Markdown variance = actual markdown dollars − planned markdown dollars

The margin plan should carry a markdown reserve by category and month. Actual against that reserve tells finance whether the season is spending more or less than it planned, and when. A season that is under its markdown plan in the early months and far over it at the end is the late-and-deep pattern of scenario C, visible in the monthly numbers before the hindsight.

3. Margin dollars recovered

Margin dollars recovered = gross margin dollars of the action taken − gross margin dollars of the hold-price baseline

This is the measure markdown optimization should be judged on, because it compares what was done with what would have happened without it. In the example, B recovered $28,800 − $25,600 = $3,200 and C recovered $26,240 − $25,600 = $640. The baseline is a projection, so it should be calculated the same way every time — the style's pre-markdown rate, decayed forward — and kept with the decision so the hindsight can test it.

4. Sell-through at exit

Sell-through at exit = units sold through owned channels by the exit date ÷ units received × 100

In the example: A 700 ÷ 1,200 = 58.3%, B 1,080 ÷ 1,200 = 90.0%, C 910 ÷ 1,200 = 75.8%. Split it into full-price sell-through (units sold at full price ÷ units received: A 58.3%, B 33.3%, C 49.2%) and markdown sell-through. Full-price sell-through is the measure of the buy; total sell-through at exit is the measure of the markdown execution. The sell-through rate formula has the conventions.

5. Residual inventory

Residual units = units received − units sold by the exit date, and residual recovery = residual revenue ÷ residual units

In the example the residual was 500, 120 and 290 units, each recovering $16. Residual units are what the exit channel has to absorb, and the recovery per unit is what the clearance actually earned. Track both by cause — depth, timing, size, breadth — so the residual feeds next season's buy rather than just this season's write-down.

Two supporting measures explain why the five moved: average markdown depth = markdown dollars ÷ (original price × units sold at a reduced price), which in B is $21,120 ÷ ($80 × 680) = $21,120 ÷ $54,400 = 38.8%; and size residual share = residual units in fringe sizes ÷ total residual units, which shows whether the size curve is driving markdowns.

MetricFormulaTells you
Markdown % of salesMarkdown $ ÷ net salesPrice reduction taken
Markdown $ vs planActual − planned markdown $Reserve spend and timing
Margin $ recoveredAction GM $ − hold GM $Whether the markdown paid
Sell-through at exitUnits sold ÷ units receivedExecution of the clearance
Residual unitsReceived − sold by exitWhat the exit must absorb

Channel notes

Markdown optimization looks different in each channel, because each has its own price rules, its own exit route and its own customer.

Stores. Price is ticketed, so every step has a labor cost and a lag, and size-restricted markdowns are harder to execute. Transfers are the cheapest lever before a markdown.

DTC. Price can change by SKU, which makes size-restricted markdowns easy and discipline harder. The common failure is sitewide codes stacked on markdowns, so the effective price falls below any planned step and the customer learns to wait for the code. Promotions and markdowns should share one calendar.

Outlet. An outlet that is a planned channel, with its own buy and its own size curve, is a lever. An outlet that is a destination for residuals is a second markdown problem in a different building. See planning an outlet channel for the difference.

Wholesale and off-price. Wholesale accounts often expect support on slow sellers, through markdown money or allowances, which is a markdown cost the brand pays without changing its own ticket. Off-price buyers set their price against the season's timing and against every other brand clearing the same weeks, so the recovery per unit falls the later the residual is offered. Both belong in the markdown plan, not in a separate negotiation after the season.

Vertical notes

The trigger-and-cadence method above is written for seasonal apparel, where the exit is forced by a sell-through threshold and a season change. Other verticals are forced out by different clocks, and the markdown approach has to follow the clock. The markdown and exit strategy by vertical guide covers ten verticals in depth; the short version for four of them:

  • Footwear. The exit is driven by model-year changeover and by broken size runs. A model is closed out when the new model year lands, often across every colorway at once, so the clearance is scheduled backwards from the changeover date. Width fits and half sizes break runs faster than apparel, and a run with the core sizes gone does not respond to price.
  • Home and furniture. Long lead times and container minimums mean the overhang is often a container quantity rather than a demand miss. Floor sets and floor samples have their own clearance at each reset, and dealer prebooks constrain what the brand can do on price while dealers still hold the item.
  • Beauty. Seasonal gift sets and limited shades exit on a fixed date, and dating — shelf life to expiry, and the minimum remaining life a retailer will accept on receipt — sets a ceiling no markdown can extend. Gondola resets force an exit on the retailer's calendar, and markdown is often the last lever rather than the first.
  • Toys. The gifting peak dominates, so post-holiday exit is planned before the peak, and licensed product exits on the license window regardless of sell-through.

Exit strategies for end-of-season inventory

Markdown is one exit among several. Off-price channels preserve pricing on owned channels; bundling pairs slow movers with best-sellers without a visible ticket change; carry-forward suits core styles that are not trend-dependent; donation can recover partial value through tax deductions. Destruction is closing as a route: the EU's Ecodesign for Sustainable Products Regulation (ESPR, Regulation (EU) 2024/1781) bans the destruction of unsold apparel and footwear at large companies from July 2026, with medium-sized companies to follow. The end-of-season exit strategies guide sets out the margin math for each exit.

Common markdown optimization failures

These are the patterns that turn a planned markdown budget into an unplanned clearance bill.

  1. Calendar triggers instead of performance triggers. Marking everything down in the same week over-discounts the styles that were selling and under-discounts the ones that were not.
  2. Judging by calendar week, not week of life. Late receipts are marked down before they have had a fair read.
  3. Managing to markdown percent. Holding price to protect the rate pushes the cost into the residual, as scenario A shows.
  4. Skipping the cover test. Sell-through without weeks remaining is half a signal.
  5. Whole-style markdowns on a broken run. Discounting core sizes to clear fringe sizes that will not clear at any price.
  6. No exit point. A cadence without a defined end keeps deepening on units that needed a different route.
  7. Re-debating every style. Without agreed triggers, each markdown is negotiated, and negotiation is systematically late.
  8. No record of overrides. If the trigger said step and the merchant held, the hindsight needs to know, or the trigger table never improves.

Where a planning system fits

Spreadsheet-based planning can surface markdown risk, but usually late. By the time a merchant notices slow sell-through in a pivot table, the shallow steps have often expired. The work a connected planning environment makes easier is the work this guide describes: holding the buy, receipts, size curves and the markdown reserve in one plan, so a planner can see where planned depth exceeds the style's history, check sell-through and weeks of supply against the trigger table each week, and record which lever each residual used and what it recovered.

The same record closes the loop into next season. The full-price split, the trigger history and the residual by cause are the inputs to the next buy, and a team that holds them in one place starts the next season from evidence rather than from memory. Replacing spreadsheet planning and margin planning cover that side.

See how RetailNorthstar keeps the open-to-buy, the receipt plan and the markdown reserve in one connected plan, so the markdown budget is planned with the buy instead of discovered at the end of the season.

Start Free Trial →

No credit card. No commitments.

Related resources

Common questions

What is markdown optimization in retail?

Markdown optimization is the planning discipline that decides when to reduce price on unsold inventory, how deep each reduction goes, in which doors and channels, on which sizes, and on what cadence, so that the season clears by its exit date for the least margin given up. It has two halves: preventive planning that reduces how much inventory reaches markdown at all, and in-season decisions that clear what does reach it at the lowest total cost.

What is the difference between markdown strategy and markdown optimization?

A markdown strategy is the policy: the triggers, depth steps, exit routes and owners agreed before the season, usually by product classification. Markdown optimization is the work of choosing the best action for each style inside that policy, week by week, from its sell-through, weeks of supply, remaining selling weeks and expected price response. The strategy sets the rules of the game; optimization plays each style within them.

Is a lower markdown percent always better?

No. Markdown percent measures how much price reduction was taken against sales, not how much margin the season kept. In this guide's worked example, holding full price to the exit date produces a markdown percent of zero and the lowest gross margin dollars, because 500 units leave through off-price at $16. Taking a first markdown in week 7 and a second in week 11 produces a markdown percent of 32.4% and $3,200 more gross margin. Read markdown percent next to gross margin dollars, sell-through at exit and residual units.

When should the first markdown be taken?

When a style misses its sell-through or weeks-of-supply trigger at a defined checkpoint in its life and the remaining selling weeks are no longer enough to clear the position at full price. The calendar week alone is the wrong trigger. The test is cover against time: if weeks of supply on hand exceed the weeks left to the exit date, the style will not clear at its current price and the first step should land while the style still has traffic.

What is the difference between rules-based and model-based markdown optimization?

A rules-based approach fires pre-agreed markdown steps when a style misses sell-through or weeks-of-supply thresholds at set weeks of life. It needs only weekly sales and inventory by style, is transparent, and is easy to govern, but it treats every style in a classification the same. A model-based approach estimates each style's price response from history and chooses the price path that maximizes margin dollars subject to the exit date, a margin floor and the brand's price rules. It can be more precise, but it needs clean price and promotion history, and it is only as good as its elasticity estimates.

Which metrics measure markdown performance?

Five measures cover it: markdown percent of sales (markdown dollars divided by net sales), markdown dollars against plan, margin dollars recovered against a hold-price baseline, sell-through at the exit date, and residual inventory left for an exit channel. Two supporting measures, average markdown depth and size residual share, explain why those five moved; sell-through at exit is also split into full-price and markdown sell-through.

RetailNorthstar Editorial Team
RetailNorthstar ·

Share this guide with your team

Copy a link or a pre-written message for Slack, Teams, or email.

// Know where your operation stands

Apply this to your planning operation.

The free Apparel Planning Maturity Assessment benchmarks your operation and tells you exactly which gaps to fix first.

Take the assessment →

Apply these insights with RetailNorthstar.

See how modern apparel brands use RetailNorthstar to put this planning framework into practice.

No credit card. No commitments.

Connected merchandise planning — live in weeks, not quarters.