The In-Season Chase: Which Levers Are Still Open, and the Last Week Each One Works

What an in-season chase is

An in-season chase is any action taken after the buy is committed that changes what a brand will actually have to sell, and each such action has its own lead time, which is what decides how long it stays available. The word "chase" is usually used more narrowly than that, to mean a reorder placed against a style that is selling ahead of plan. That definition is too small to plan with, because it excludes most of what a merchandising team actually does between commitment and season end. A chase is any change to the future inventory position, in either direction — pulling more units toward a style that is working is one shape of it, and taking units away from one that is not is the other. Both are governed by the same constraint, which is time.

This guide covers the period after commitment and before exit. Three neighbouring pieces cover what sits either side of it. How often to reforecast derives the cadence at which you should be asking the question, from the same deadline logic used here, and deliberately does not enumerate the decisions themselves — this page is that enumeration. The in-season reforecast is the argument for why the exercise matters at all; read it as the case, and this as the lever list. End-of-season exit strategies picks up after every lever here has closed, when the only remaining question is which route the residual stock leaves by.

There is one more boundary worth naming, because it is the one teams collapse most often. Change control before commitment and lever management after it are different disciplines, with different owners and a different cost curve. What sign-off on a range means, and how a locked line gets amended before anything is bought, is covered in locking the line on line-board.com. Everything on this page assumes the line is locked, the purchase orders are cut, and the money is spent or committed.

The last responsible week

The organising idea of the whole page is a date, computed per lever.

A lever's last responsible week is the week the season stops rewarding the change, minus the lead time that lever needs. Both halves have to be derived; neither is a given.

The second half is the easy one. Lead time here means the lever's own end-to-end duration — for a reorder, from the moment the order is released to the moment the units are sellable in the channel, including receiving and processing at the far end, which is routinely left out and is routinely the part that slips.

The first half is where the arithmetic lives, and where most teams substitute the season end date by default. The season stops rewarding a change earlier than it ends, because units that land in the final weeks cannot sell through at full price before the exit begins. What matters is not the last day of the season but the last week the product still trades at full price — the week before the planned markdown cadence starts, or before the floor set changes to the next delivery, whichever comes first. Call that the full-price horizon.

Then the quantity matters, because a bigger chase needs more weeks to clear:

Weeks needed to clear = chase quantity ÷ current weekly full-price sell rate

Reward-stop week = full-price horizon − weeks needed to clear

Last responsible week = reward-stop week − lever lead time

Three consequences follow from that chain, and each of them is a real behaviour change.

First, the last responsible week is a property of the lever, not of the season. Two levers considered in the same meeting on the same style can have expiry dates many weeks apart. A conversation that concludes "it's too late to do anything" has almost always tested one lever and generalised.

Second, the last responsible week moves earlier as the chase quantity grows. Doubling the order doubles the weeks needed to clear it, which pulls the deadline in. A quantity that made sense in week 4 may be the wrong quantity in week 8 even if the lever is still technically open — the lever survives, the size does not.

Third, the whole chain runs off the current sell rate rather than the planned one, which is the number that triggered the conversation in the first place. If the style is running at, say, half again over its planned rate, the plan rate would overstate the weeks needed and produce a deadline that is comfortably, wrongly, later.

The lever table

This is the artifact to keep. Seven levers, four columns held constant: what it changes, the lead time it consumes, what it costs, and the last week it can be pulled. The last column is expressed as a formula rather than a number, because a fixed week would be a fabricated benchmark — every brand's lead times, and therefore every brand's deadlines, are its own.

Two things about that table are worth stating explicitly, because they are the reason it is useful rather than decorative.

The order of the rows is the order in which the levers expire, roughly, for most brands. Supply levers close first because they carry the longest lead times; distribution levers stay open much longer because they only move units that already exist; price never closes. That ordering means a team working down the list from the top is testing the levers in the order they disappear, which is the correct search order.

The cost column is not all denominated in money. A colorway substitution is nearly free in dollars and expensive in range integrity. A transfer costs freight and costs the sale it removed from the origin door. Reading the table as though only the cash column counts is how a cheap-looking lever gets pulled repeatedly until the assortment stops making sense.

The cancellation lever, which nobody writes down

One row deserves expanding because it is the one most often assumed rather than checked. A cancellation does not save the order value; it saves the unspent balance of the order.

What the vendor has already committed on your behalf usually stays with you. Piece goods bought against your order, trims ordered to your specification, and anything already cut are costs that exist whether the finished units ship or not. What survives cancellation is the make cost, the remaining material not yet purchased, and the freight — which is a real saving and a smaller one than the number on the purchase order.

The consequence is that the cancellation window is keyed to the vendor's procurement dates, not to your ship date. Those dates sit upstream of anything printed on the order, they differ by vendor and often by fabric, and they are very rarely written into the supply agreement in a form anyone can look up mid-season. The only reliable version of this information is the version you asked for and recorded before the season started — a per-vendor note of the last date at which an order can be cancelled outright, the last date it can be reduced, and what the liability is on each side of both dates. Teams that have this can act in a week. Teams that do not spend that week finding out, and the week is usually the whole margin of the decision.

The same applies to a push, which is the softer form of the same lever and is frequently the better one: the units still arrive, later, against a re-phased plan. A push costs storage and pressure on the following season's receipt flow, and it preserves optionality that a cancellation destroys.

The chase break-even

A chase pays when the incremental full-price margin it lands covers the expedite premium on the whole order plus the margin outcome of the units that arrive too late to sell at full price.

Written as a threshold on the share of the order that lands in time:

Break-even full-price share = (expedite premium per unit − margin per late unit) ÷ (full-price margin per unit − margin per late unit)

Three things move in that expression as the weeks pass, and they all move the same way.

The actual full-price share falls, because the arrival week slides later while the full-price horizon does not move. The required share rises, because the units landing late land into a deeper part of the markdown cadence, so the margin per late unit falls. And the expedite premium per unit rises or holds, because the later you commit the more of the transit has to be bought as air rather than ocean.

The break-even is therefore moving toward the actual number from both directions at once. That is the mechanism behind the single most expensive habit in in-season management, which is watching a style for another two weeks to be sure. Waiting is not a neutral act that preserves the decision. It is a decision, it has a price, and the price is computable in advance.

Worked example: one style, two dates

Illustrative figures throughout, chosen because they divide cleanly. Not benchmarks, and not drawn from any brand. Lead times below are inputs to this example, not norms — substitute your own.

The setup. A 20-week season. One style-color, bought at 2,000 units, planned at 100 units per week. The brand's planned markdown cadence starts in week 17, so the full-price horizon is week 16. Unit economics: retail 80, landed cost 36, so full-price margin is 44 per unit. First markdown is 50 per cent off, giving a margin of 4 per unit. Anything left at season end goes to an off-price route recovering 22 against a cost of 36, so a loss of 14 per unit. The expedite premium on a chase order is 9 per unit.

Week 6, reading the numbers. Weeks 1–6 sold 900 units — 150 per week, half again over plan. On-hand is 1,100. At 150 per week that runs out during week 13. From week 14 to the full-price horizon in week 16 there are three full-price weeks with no stock: 3 × 150 = 450 units of full-price demand the current buy cannot serve, plus whatever the markdown weeks would have absorbed.

Computing the last responsible week for each lever. The reward-stop week is the week the chase units have to be sellable, which is the full-price horizon less the weeks needed to clear them. The vendor's minimum for a chase run is 1,000 units, so the quantity is not free: 1,000 ÷ 150 is 6.7, call it 7 weeks needed — but only 3 full-price weeks exist. That mismatch is itself the finding, and we come back to it. Taking the arrival deadline as week 13, when the existing stock runs out:

Two supply levers are already closed at the first moment anyone had enough data to want them. That is not a process failure; it is the normal geometry of a season, and it is the reason reserved capacity and held material are pre-season decisions rather than in-season ones. The lever that is open is open only because someone held fabric before the season started.

The break-even at week 6. Ordering 1,000 units in week 6 against a 6-week lead lands them in week 12. Existing stock covers through week 13, so the chase units sell from week 14. Full-price weeks 14, 15 and 16 absorb 3 × 150 = 450 units. The remaining 550 are on hand when the markdown starts in week 17; at a markdown rate of 175 per week over weeks 17–20 the markdown period can absorb 700, so all 550 clear at markdown and nothing goes to off-price.

• →Full price: 450 × 44 = 19,800
• →Markdown: 550 × 4 = 2,200
• →Off-price: 0
• →Expedite: 1,000 × 9 = −9,000
• →Net: +13,000

Checking it against the threshold: margin per late unit is 4, so the break-even full-price share is (9 − 4) ÷ (44 − 4) = 12.5 per cent. The actual share is 450 ÷ 1,000 = 45 per cent. The chase clears the bar by a wide margin.

The same decision at week 10. Nothing about the order has changed — same vendor, same 1,000 units, same 6-week lead, same prices. It now lands in week 16. Weeks 14 and 15 are stocked out and stay stocked out. Only week 16 sells at full price: 150 units. The remaining 850 are on hand at the markdown; the markdown absorbs 700, and 150 go to off-price.

• →Full price: 150 × 44 = 6,600
• →Markdown: 700 × 4 = 2,800
• →Off-price: 150 × −14 = −2,100
• →Expedite: 1,000 × 9 = −9,000
• →Net: −1,700

The threshold has moved too. Blended margin per late unit is now (2,800 − 2,100) ÷ 850 = 0.82, so the break-even share is (9 − 0.82) ÷ (44 − 0.82) = 18.9 per cent, against an actual share of 15 per cent. The required share went up while the achieved share went down, which is why the result does not just shrink — it inverts.

What the four weeks cost. The same order swings from +13,000 to −1,700, a difference of 14,700 on 1,000 units, or roughly 3,675 for each week of delay. That number is the price of "let's watch it another two weeks", and it was computable in week 6. The decision that was actually made in week 10 was made in week 6, by not making it.

One more reading of the same example, which is the part that survives the arithmetic. Even the week-6 chase serves only 450 units of full-price demand while committing to 1,000, because the vendor minimum is larger than the window can absorb at full price. The chase is worth doing and it is still 55 per cent a markdown buy. A team that understands that goes into the decision knowing it is buying 450 good units and 550 planned markdown units — which is a defensible thing to do, and a very different thing from believing it bought 1,000 good units.

What to do when nothing upstream is open

This is the common case, and it deserves an honest section rather than a paragraph of encouragement.

When every supply lever has closed, four things remain. The first three move units or margin. The fourth is the one people skip.

Re-allocate what is not yet distributed. Warehouse holdback is the cheapest lever on the table and the one most often left unpulled, because it does not feel like a decision — the stock is already bought, so moving it registers as logistics rather than merchandising. It is merchandising. Holdback sent to the doors and channels now proving demand converts inventory that was doing nothing into full-price sales, at handling cost, in days. This is the lever the inventory allocation view exists to make visible.

Transfer between locations, and shift between channels. Units sitting in doors where the style is not working are inventory in the wrong place, and a transfer is a chase whose lead time is the pick, the truck and the receiving rather than production. The constraint is size curve integrity: pulling the sizes that are not selling out of a deep door and consolidating them where the run is broken restores complete runs in fewer places, which sells better than the same units spread thin everywhere. Channel shifts work the same way and are usually faster, since a channel reallocation may need no physical movement at all.

Use price deliberately in both directions. On a style running hot with no chase available, the correct price action is often to do nothing for longer — hold full price, let it sell out cleanly, and keep it out of any broad promotion, because every unit sold at a discount on a style that is going to stock out anyway is margin given away for no volume. The mirror lever is on the styles that are long: an earlier, shallower markdown taken while the style still has velocity protects more average unit retail than the deep clearance that a late decision forces. Both of those are covered in more depth in end-of-season exit strategies.

Take the stock-out, and price it. When no lever lands units in time, the answer is that the style will be out of stock from a known week, and the cost of that is computable: lost full-price weeks × the current sell rate × full-price margin. Writing that number down does three things. It stops the team spending real money on a rescue that will not work. It gives the pre-season conversation a fact rather than a feeling — this is what not reserving capacity cost, in this style, this season. And it makes the open-to-buy reserve argument concretely, which is the only version of that argument anyone ever wins.

That last point is the honest summary of the whole page. Most of what determines which levers are open in season is decided before the season opens — capacity reserved, material held, cancellation windows negotiated and written down, and open-to-buy deliberately withheld rather than discovered as an underspend. The in-season work is executing against that pre-season position quickly. It cannot create a position that was never bought. The mechanics of buying that responsiveness are set out in speed to market in apparel.

How the levers differ by vertical

The framework is the same everywhere: reward-stop week, minus lead time, per lever. What changes is which levers exist at all and how early they close.

Apparel

Apparel is the base case used above, and its distinguishing feature is that the enabling lever is almost always material already in stock. Fabric held against a hero style is what converts a theoretical reorder into an achievable one, because it removes the longest single block from the lead time. The complication is the size curve: a chase order lands a full run, but a stock-out is rarely a full-run event — it is the middle of the curve going first. Chasing a complete run to fix a shortage concentrated in three sizes buys units in the tails that will exit at markdown, which is a cost that belongs in the break-even and is usually left out of it.

Footwear

Footwear closes its supply levers earlier than apparel does, and for a reason that has no apparel analogue: the chase depends on tooling, not on material. A model-color reorder needs the last and the outsole tooling to be free, and tooling that has been reassigned to the next model year is not a lead-time problem that money can compress — it is a hard gate. Component lead times for outsoles and uppers also run longer than fabric, so even where tooling is available the arithmetic starts from a longer number.

Run integrity is the second difference. A footwear chase that lands only part of a size run has not fixed the position, because a broken run is functionally dead inventory long before the unit count says so. The useful measure of what a footwear chase has to deliver is complete runs, not pairs, and where widths are carried the grid multiplies again. The practical effect of both differences together: the last responsible week for a footwear supply chase sits materially earlier in the season than the same computation in apparel, and the redistribution levers therefore carry more of the load.

Outdoor

Outdoor changes the first half of the equation rather than the second. The full-price horizon in outdoor is set by weather, not by the calendar, and weather does not respect the markdown cadence — an insulation or shelter category can stop rewarding a chase weeks before the planned exit simply because the conditions that drive the purchase have passed. That makes the reward-stop week both earlier than the calendar suggests and less predictable, which argues for computing it against observed conditions rather than against the plan.

Two structural constraints compound it. Commitment sits in the dealer prebook, which means the supply position was largely fixed long before the sell window opened and there is often no open capacity to chase into. And where minimum advertised price policy applies, price is not fully available as an exit lever, so units that land late do not simply clear at a worse margin — they sit, and they roll into the next model year. A late chase in outdoor is therefore more expensive than the same late chase in apparel, because the residual has fewer routes out.

Sporting goods

In sporting goods the honest answer is usually that most levers were never open. Prebook commitment means the season's supply is fixed before it starts, and the vendor's capacity is allocated against a book that closed months earlier. The in-season chase is therefore not a supply question at all; it is an allocation and channel question — moving units between dealers, between the dealer network and direct channels, and into or out of team and institutional orders, which behave as their own channel with their own timing.

That has a practical implication for where the effort goes. A sporting-goods team working the supply row of the lever table is working the row least likely to yield anything, while the re-allocation and transfer rows — the cheap ones, whose lead time is handling rather than production — are where the recoverable margin actually is. The pre-season equivalent of a chase capability here is not reserved capacity but a deliberately unallocated share of the prebook, held back to be placed once demand is observed.

Accessories and bags

Accessories and bags has the most reliable chase enabler of the five, which is component stock. Hardware, zips, linings and hides held in inventory shorten the chase to make-and-ship, and because these components are shared across styles within a family, holding them is a hedge across several possible outcomes rather than a bet on one style — which makes the holding decision much easier to justify pre-season than a fabric commitment to a single style.

The absence of a size axis cuts both ways. A chase quantity is one number rather than a curve, so the decision is faster and there is no tail risk from buying sizes that will not sell. But depth concentrates: the entire chase lands on one style-colorway, so an oversized chase has nowhere to hide and the whole error shows up in one line. The real constraint is upstream, in minimum order quantities and material yield, which set the smallest sensible chase and are frequently larger than the remaining window can absorb at full price — the same mismatch the worked example ran into, but sharper, because there is no size run across which to spread it.

Where this goes wrong

Five failure modes, in rough order of how often they appear.

1. →Testing one lever and generalising. "We can't do anything" almost always means the reorder was checked. The distribution levers are still open for weeks afterward and cost almost nothing.
2. →Using the season end date as the reward-stop week. This produces deadlines that are comfortably wrong by the length of the markdown cadence, and it is the single most common arithmetic error in the chain.
3. →Computing weeks of cover off the planned sell rate. The style is in the conversation precisely because it is not selling at plan. Using the plan rate overstates remaining cover and hides the deadline.
4. →Treating a delay as neutral. Every week of waiting moves the break-even and the achieved share in opposite directions. The cost of another two weeks is a number, and it should be on the slide.
5. →Discovering the cancellation terms during the season. The vendor's material-commitment dates are not on the purchase order. Collected pre-season, they make the cancellation lever usable in a day; collected in season, they cost the week that was the whole decision.

Related resources

• →How Often to Reforecast a Merchandise Plan — the cadence at which these questions should be asked
• →The In-Season Reforecast — why the exercise matters, and why monthly is too slow
• →End-of-Season Exit Strategies — what happens after every lever here has closed
• →Speed to Market in Apparel — buying the responsiveness that makes these levers exist
• →Dating Rules and Weeks of Supply — the same backwards-from-the-deadline arithmetic, applied to dated product
• →For Apparel Buyers — the buyer's view of the open-to-buy and the chase reserve
• →For Inventory Allocation — the re-allocation and transfer levers in practice