Core and NOOS Replenishment Programs

A core or never-out-of-stock (NOOS) replenishment program is a standing commitment to keep a defined set of items continuously available, planned as an ongoing stock position governed by a replenishment rule rather than bought once per season and exited. It is a different planning discipline from seasonal buying, with a different unit of commitment, a different forecast, a different success measure and a different way of consuming money — and it is usually run by the same people, in the same meeting, on the same budget as the seasonal business, which is where the trouble starts.

This guide covers the program as a discipline: what qualifies an item and what disqualifies it, the review cadence that admits and removes items, how the trigger and safety stock are set and re-set, how core demand is forecast differently from fashion demand, how core and seasonal split one open-to-buy, how supplier lead time and minimum order quantity turn out to be the real constraint, and the four ways a program that looks healthy on a report is failing in the door. For the definition and the short form, see the never out of stock glossary entry. For how allocation and replenishment work in general, see allocation and replenishment for apparel and allocation and replenishment by vertical; this guide assumes those mechanics and concentrates on the program above them.

Core and seasonal are two different planning jobs

The seasonal buy answers a question with a deadline: what do we carry this season, how deep, at what price, and when does it leave. The core program answers a question with no deadline: does this item stay available, at what cover, until we decide otherwise. Almost every practical difference follows from that.

The measurement line matters more than it looks. Sell-through is the wrong metric for a core item and will always make a good program look bad, because a well-run core item is deliberately never sold through — holding cover is the point. Judging core on sell-through pushes planners to thin the position until the in-stock breaks, and the cost of that break does not appear in any report unless someone is estimating lost sales. Judging core on in-stock rate pushes the opposite way, toward cover that is never questioned, which is why in-stock has to be read next to the capital the cover ties up. The honest pair is in-stock rate and the inventory the program carries, reviewed together.

What qualifies an item for core status

Qualification is two sets of tests, on demand and on supply, and both must pass. The most expensive mistake in this whole discipline is promoting an item on the demand tests alone.

The demand tests

A rate you can set a rule from. Not a high number — a persistent one. The question is whether the item's weekly rate holds across a full year including both halves of the category's seasonal cycle, not whether it had a strong quarter. A style whose rate triples in one season and halves in the next has a seasonal shape, and a flat replenishment rule against a seasonal shape overstocks it for half the year and breaks in the other half. Where an item is genuinely core but seasonally amplified — a heavyweight core fleece, a core sunscreen — it needs a seasonal profile on the rule rather than a flat rate, which is a heavier thing to maintain and should be a deliberate choice.

No exit date. If the calendar already knows when the item leaves — a licensed window, a model year changeover, a fabric the mill has notified as discontinued — it is not core. It may be a long carryover, which is a different plan.

A stable curve. The item's internal split has to be stable, because the rule is set at that grain. In apparel that is the size curve; in footwear the size run within each width; in beauty the shade ladder. A style whose size curve moves every season is telling you its customer is changing, and a fixed per-size minimum will be wrong in a new way each quarter.

Demand that is not a promotion's shadow. An item that runs at a steady rate because it is permanently on offer, or because it is the anchor of a recurring bundle, has a rate that belongs to the promotion. When the promotion stops, the rate does not persist. Separate baseline from lift before reading the rate — the same discipline demand sensing applies in-season.

The supply tests

Re-orderable at all. Someone has to be willing to re-run the thing: the fabric or yarn, the trim and hardware, the last or mould, the finish, the shade base, the component. This is a question for sourcing, in writing, per item, and the answer changes without anyone telling merchandising.

A lead time safety stock can cover. Cover costs money, and the buffer scales with the lead time and its variability. An item on a sixteen-week refill can be core, but it will carry months of cover to be so, and that is a capital decision rather than a stock decision.

A minimum order quantity that buys a sensible number of weeks of cover on the item's slowest qualifying size, not just on its fastest. This is the test that fails most often and is checked least often, and it is covered in detail below.

A refill price that keeps the item in its band. Where the cost base moves — precious metal, a commodity fibre, a freight-exposed landed cost — each refill is priced at the day's cost, and an item can drift out of the price band it was built for. Planning margin on a moving cost base covers the mechanics; for the core program the test is simply whether the item survives a plausible cost move without a repricing decision.

What disqualifies

Fashion colors, limited drops and licensed product fail by construction: their exit dates are set in advance. Anything on a model-year cycle qualifies only until its changeover, which means the program needs the changeover date on the item record, not in someone's head. Single-source artisanal product usually fails the supply tests even when the demand is perfect. And a style that sells beautifully but whose vendor will only re-run it against a full season's commitment is a carryover bought once per season — planning it as core reserves budget for refills that will never arrive, which is the quietest way to lose open-to-buy.

The review cadence that admits and removes items

A core program is only as good as the calendar that maintains it. Five clocks run at once, and each owns different decisions. Collapsing them into one monthly meeting is how parameters go stale.

Entry: candidate, probation, promotion

Entry should be a three-step path rather than a switch. A candidate is nominated from the seasonal business — a style that has repeated across seasons, or a carryover whose rate has held. It goes onto probation: a provisional rule, a stated number of weeks, and a deliberately conservative maximum so that a failed probation costs a small residual rather than a year of cover. At the end of probation the item is promoted only if the rate held, the curve held, and sourcing has confirmed the re-run terms. Probation length follows the refill lead time, because a probation shorter than one full refill cycle has not tested the thing that matters.

Exit: the rule that most programs do not have

Exit needs a floor, a confirmation period and a run-down plan, all decided before the item is admitted. The floor is whatever the business will state: a weekly rate below which the cover is not worth its capital, a margin below which the item is not worth the space, or an in-stock cost that has become unjustifiable. The confirmation period stops a single soft quarter from demoting a good item — two consecutive review periods below the floor is a common shape. The run-down plan is the part teams forget: demotion means the trigger stops firing, not that the inventory disappears, and the remaining position has to be sold down through the existing channels or exited deliberately. Without a named run-down, demoted items sit at their old maximum for months.

The exit rule also has a date arithmetic that is easy to get wrong. The last replenishment that can land and still sell is placed at the planned exit date minus the refill lead time, so an item on a fourteen-week refill has to stop triggering more than three months before it leaves. Miss that, and the final order lands into a position the calendar is about to close.

Setting the trigger, the safety stock and the cover target

The arithmetic is not the hard part, and this site keeps it on its own pages: the safety stock formula, the replenishment trigger formula, weeks of supply and forward weeks of supply. The hard part is the grain the parameters are set at, and the two tiers most core programs actually run.

Two tiers, two lead times

A core item replenished through a brand's own distribution center has two rules stacked on each other, and they take different inputs. The door rule covers the DC-to-door lead time, which is measured in days, so its buffer is small and a miss is corrected on the next cycle. The DC rule covers the supplier lead time, which may be measured in months, so its buffer is sized mostly to lead-time variability rather than demand variability, and its order quantity is set by the vendor's minimum rather than by the doors' demand. Running one rule for both — a single fleet-level weeks-of-supply target — is how brands end up with plenty of total inventory and empty doors.

Illustrative worked example

The figures below are illustrative and were chosen because they divide cleanly. They are not benchmarks, not targets, and not drawn from any brand's data. All rates are demand net of returns; if a program is DTC-weighted, net the return rate before setting the rate, because a gross order rate sets a minimum against units that came back.

A core crew-neck tee in black runs at 600 units a week across the business, replenished from an overseas vendor into the brand's own DC, then from the DC to twenty doors and a DTC channel. DTC takes 20% of the demand; the doors split the rest evenly.

The size curve is stable, which is why the item qualified:

The DC rule. The supplier lead time is 10 weeks and the buffer is 21 days, because a slipped vessel is the dominant risk on a refill that long. Safety stock is weekly demand × 21 ÷ 7, which is three weeks of demand; the minimum is (10 × weekly) + (3 × weekly), or 13 × weekly. For size M: safety stock is 150 × 21 ÷ 7 = 450 units, and the minimum is (10 × 150) + 450 = 1,950 units. Across the run the minimum totals 13 × 600 = 7,800 units, and the per-size column above sums to the same 7,800.

The door rule. The DC-to-door lead time is one week and the buffer is seven days. For size M in one door: average weekly demand is 6.0 units, safety stock is 6 × 7 ÷ 7 = 6 units, and the minimum is (1 × 6) + 6 = 12 units. The DC ships that size in inners of 6, so the maximum is 12 + 6 = 18. At the minimum the door holds 12 ÷ 6 = 2 weeks of cover; at the maximum, 18 ÷ 6 = 3 weeks.

Where the rule stops working. Size XS in one door runs at 1.2 units a week. The same arithmetic gives safety stock of 1.2 and a minimum of 2.4, which is rounded up to 3 whole units — and 3 ÷ 1.2 is 2.5 weeks of cover, close enough to the target. But a rule computed on 1.2 units a week is reading noise: one customer buying two XS tees in a week puts that week two-thirds above the average rate. Where a position runs at a couple of units a week, the trigger stops being a demand signal and becomes a rounding artifact — where exactly that line falls is a judgement each business makes, but the failure is structural rather than a matter of tuning. The practical answers are to hold the fringe sizes at the DC and ship on demand, to carry a token single unit in the door and let DTC serve the tail, or to pool the fringe sizes across a door cluster and decide the curve there — the grain argument allocation and replenishment by vertical makes in full.

Re-setting: the parameters that go stale

Four inputs move, and each invalidates the rule silently.

• →The demand rate. A trigger computed from launch velocity keeps ordering long after the item has settled. The monthly re-set exists for this.
• →The lead time. A vendor that has quietly moved from ten weeks to fourteen consumes the whole three-week buffer and leaves the thirteen-week minimum a week short of the lead time itself, before any variability is allowed for.
• →The order multiple. A changed pack, carton, inner or container changes the maximum, and therefore the cover, without anyone touching the minimum.
• →The curve. A size curve or shade ladder that has shifted moves demand between positions whose minimums were set against the old split.

A re-set that nobody owns a date for is a re-set that does not happen, which is why the cadence table above names an owner per clock rather than describing the arithmetic again.

Forecasting core demand is a different job from forecasting fashion

Fashion forecasting is an analog problem: the style has no history, so the forecast is built from comparable styles, trend and judgement, and the useful output is a season's total with a phasing. Core forecasting is a rate problem: the item has its own history, and the useful output is a weekly rate with a stated tolerance. That difference changes the method in three places.

The base is estimated from the item's own history, not from analogs. A trailing average over a window long enough to be stable and short enough to be current is the working method, with the window length following the item's own volatility rather than a house default. What matters more than the averaging technique is what gets excluded from the base: promotional weeks, weeks where the position was broken, and weeks where a door was not stocked at all. Sales history from a stocked-out position is not low demand, it is censored demand, and a rate computed over it sets a minimum that guarantees the next break. In-stock rate at the sellable grain is the diagnostic that tells you which weeks to exclude.

Lift is planned separately and added back. Core items carry the promotional calendar for most categories — the basics that anchor a multi-buy, the core shades in a gift-with-purchase, the consumables in a season-opening offer. That lift is known in advance and belongs on the receipt plan as a one-off uplift against a specific week, not as a permanent change to the base rate. Folding a promotional week into the trailing average raises the minimum permanently, and the position never comes back down. The reverse error is just as common: dropping a promotion the item's base had come to depend on, and leaving the rate untouched.

Curve stability is itself a forecast output. For fashion, the size curve is an assumption inherited from a class. For core, the curve is measured from the item's own selling and is one of the things being watched — a drifting curve is an early signal that the item's customer is changing, usually before the total rate moves at all. Size curve allocation covers the mechanics of applying it.

One more difference is worth stating plainly, because it cuts against instinct: a core item's forecast error is less costly in units and more costly in duration. A fashion miss is bounded by the season — it ends in a markdown, and the number is knowable. A core miss repeats every week until someone re-sets the rule, in either direction: an over-set minimum ties up capital indefinitely, and an under-set one loses the same sale every week to the same empty position. That is why the monthly re-set is worth more on core than a better forecasting technique would be.

Core and seasonal compete for one open-to-buy

This is where the two disciplines actually collide, and where most of the damage happens.

Core receipts arrive month by month regardless of what the season decided. That makes them a prior claim on the receipt plan, not a discretionary spend inside it: the program's expected replenishment is reserved before seasonal breadth is set, and the seasonal buy is planned against what is left. Doing it the other way round — planning the season first and letting core replenish out of whatever the month has — means the trigger either fires and blows the receipt plan, or is held and breaks the in-stock commitment the program exists to keep. Both happen silently. What is open-to-buy planning covers the budget mechanics, and the open-to-buy formula the arithmetic.

Reserving it pre-season does not end the conflict, because rates move. The conflict is resolved in-season, weekly, or it resolves itself by starving newness.

How it is actually settled in the weekly trade meeting

The weekly trade meeting works the WSSI, and the core-versus-seasonal question arrives as a specific line: this month's projected receipts are over plan, and part of the overage is core replenishment running ahead of its reserve. There are four honest levers, and naming them is most of the job:

1. →Delay a seasonal receipt. Cheapest when the receipt is for product with a long selling window and no floor-set or marketing date attached. Expensive when it is not — a delayed receipt against a floor set or a campaign is a lost launch, not a deferred one.
2. →Reduce or cancel a seasonal order still open. Only available before the vendor commits material, which makes it a question about the order's status rather than the plan's preference. The in-season chase guide covers the same clock from the other direction.
3. →Lower a core maximum for one cycle. This is a deliberate, time-boxed reduction in cover — acceptable on a short-lead item where a miss is corrected next cycle, and dangerous on a long-lead one where the cover is the only thing standing between the item and a months-long break.
4. →Fund the overage explicitly. From planned reductions elsewhere, a markdown deferred, or an agreed increase. A real decision with a name on it.

What the meeting must not do is nothing, because the default is lever three applied by accident — the trigger fires, the receipt lands, the month runs over, and the money comes out of the next seasonal receipt nobody has placed yet. Core replenishment is the easiest spend in the business to approve and the hardest to see, because each order is small, routine and individually correct. The aggregate is what needs a decision, and the aggregate only exists if the program's receipts are a visible line in the plan rather than a stream of purchase orders.

Two structural habits make this tractable. First, the program's reserve is stated as a monthly number in the receipt plan and tracked against actual core receipts, so "core receipts are running ahead of their reserve this month" is a sentence the meeting can say, with a number behind it. Second, the program has a size — a share of receipts the business intends to commit to standing cover — that is set annually and defended, rather than emerging from the sum of everything that was ever promoted. Where the program has crowded out newness, the fix is an exit round, not a month of held triggers. What an option costs is the companion argument on the seasonal side.

Lead time and MOQ are the real constraint

Most core programs are not limited by forecasting quality. They are limited by the terms the vendor offers, and specifically by how the minimum order quantity is quoted.

An MOQ quoted per size-color and one quoted per color are different programs

Take the tee above, on the same illustrative figures: 600 units a week, the size curve in the table, a ten-week refill. Suppose the vendor quotes a minimum order quantity of 1,200 units per size, per color. For size M at 150 a week, 1,200 units is eight weeks of cover — workable. For size XS at 30 a week, the same 1,200 units is forty weeks of cover, and for XXL at 48 a week it is twenty-five weeks. The fringe sizes carry the MOQ, not the demand, and the program's inventory is decided by the vendor's setup cost rather than by anything the planner chose.

Now suppose the vendor quotes 6,000 units per color, split to the curve. The same order becomes XS 300, S 1,200, M 1,500, L 1,500, XL 1,020, XXL 480 — summing to 6,000, and ten weeks of cover evenly across every size. Fewer total units, the same vendor and the same demand — a completely different inventory profile and a completely different working-capital position.

The MOQ structure is therefore a planning decision disguised as a sourcing term, and it is worth more attention in the annual negotiation than an equivalent argument about unit cost. The same logic runs across categories: a case pack that cannot be split, a carton quantity, a fabric minimum that forces a color commitment ahead of the order, a container that has to ship full. Planning against fabric minimums and case packs and planned depth cover the two apparel versions in detail.

The commitments that sit behind the MOQ

Three arrangements make long-lead core items workable, and all three are negotiated once a year rather than per order.

• →Material held at the mill or vendor. The fabric, yarn or component is committed ahead, so the refill lead time is cut to the make-and-ship portion. This converts a capital commitment on finished goods into a smaller one on material, and it is the single most effective lever on a long-lead core program.
• →A blanket order with releases. A total quantity is committed for the year and drawn down in releases, which satisfies the vendor's planning need and the brand's cover need at once. The trap is that the blanket total is a commitment — if demand falls, it does not go away.
• →Reserved capacity. A production slot held per month, sized to the program rather than to a specific order. Useful where the constraint is the factory's calendar rather than its materials.

Each of these makes the annual review of the program's size a real financial decision rather than a formality, because each one commits money before the demand that justifies it has appeared.

What breaks

Four failure modes account for most of the damage, and all four pass a routine report.

Core that quietly became fashion

An item is promoted on a steady rate, and then the market moves under it. The color becomes dated, a fit block changes, a competitor's version resets the category, a licensed equivalent takes the shelf. The rate decays gradually rather than dropping, so no exception fires, and the rule keeps ordering into a declining position. The tell is a trailing rate that has moved one direction for several consecutive review periods, which is a different pattern from noise and is worth flagging explicitly rather than waiting for the floor. This is the case the exit rule exists for, and the case where a missing exit rule is most expensive.

Safety stock set once and never revisited

The buffer is computed at launch, correctly, and then the lead time moves. A vendor's real lead time drifts from ten weeks to fourteen through a capacity change nobody flagged; freight moves; a port issue adds a fortnight. The buffer still holds 21 days of cover and is now covering a fraction of the lead-time exposure it was built for. A safety stock is a statement about a lead time, and it expires when the lead time changes — which means the vendor's actual measured lead time, not the quoted one, has to be an input the monthly re-set reads. Measure it from purchase order to receipt and keep the variance, not just the mean, because on a long-lead item the variance is what the buffer is actually for.

A program with no exit rule

The most common failure, and the most structural. Every season adds items — each promotion individually justified by a good sell-through — and no season removes any. The program's reserved receipts grow year on year. Nobody experiences this as a decision, because the growth is distributed across dozens of small, correct promotions, and the cost lands somewhere else entirely: as seasonal newness that could not be funded, breadth that got trimmed, a chase that was not affordable. The symptom to watch is the program's share of the receipt plan over time. If it has never gone down, there is no exit rule in practice regardless of what the policy document says.

Size-level stockouts hidden inside a healthy style-level position

The style holds four weeks of cover and the report is green. Underneath, the two highest-demand sizes are empty and the fringe sizes are carrying the average. Every customer for those sizes leaves; the style records low sales on the sizes that remain; and the next rate computation reads that as demand softening and lowers the minimum. A style-level cover reading averages across positions that customers do not substitute between, which makes it structurally incapable of seeing the break that matters. The commitment and the measurement both have to sit at the grain a stockout happens — size in apparel, size within width in footwear, shade in beauty, certified configuration in juvenile hard goods — and how modern apparel brands approach sizing and replenishment covers the diagnostic in depth.

A fifth failure sits underneath all of these and is worth naming: the rule counting a position that cannot be sold. The furniture floor sample, the beauty tester, the jewelry piece out on memo, the damaged carton, the unit already allocated to an order — each inflates on-hand and delays the trigger. Available to sell is the quantity the rule should read, and for a wholesale program the trigger is only as good as the fill rate behind it, because an order the DC ships short restores the run in the plan and not in the door.

What counts as core, by vertical

The program's shape is the same everywhere; what changes is which items can qualify, the grain the commitment is made at, and what ends it. Vertical mechanics are covered in allocation and replenishment by vertical; this table is the qualification view.

Three hard ceilings appear in this table that the apparel version does not have, and in each case the ceiling overrides the demand arithmetic entirely. Beauty's maximum is bounded by dating regardless of demand — a batch justifies only the cover that sells before its ceiling, which dating rules and weeks of supply computes. In toys the ceiling is the licence itself: a licensed window is an exit date set in advance, so licensed product never qualifies for the program at all, and the evergreen pool a toys core program can draw on is correspondingly narrow — planning a licensed product window covers how a licensed line is planned instead. And footwear, outdoor and sporting goods carry the model-year boundary, where the program ends on a date that is set by the line review rather than by demand — planning a model year changeover covers the handover.

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. Because OTB planning, assortment planning and allocation sit on that shared model, receipts committed to core items are visible in the open-to-buy at the point where seasonal breadth is set, rather than arriving as a stream of purchase orders nobody has aggregated. OTB planning covers the budget side and allocation planning the distribution side; allocation and replenishment for apparel covers how the trigger sits on top of it.

Related resources

• →Never Out of Stock (NOOS) — the definition and the short form
• →Replenishment — the underlying mechanic
• →Replenishment Trigger Formula — the trigger arithmetic
• →Safety Stock Formula — the buffer component
• →Weeks of Supply and Forward Weeks of Supply — the cover measures
• →How to Set a Stock-to-Sales Ratio — the monthly cover target core sits inside
• →The Weekly Trade Meeting — where the OTB conflict is settled
• →How to Hindsight a Season — separating a demand miss from a placement miss