Apparel costing in the connected stack

Target cost

The target cost is the cost the style is allowed to have. It is set in the line plan at the moment a slot in the range is given a ticket price and a target initial markup — and therefore a maximum landed cost — before there is a sketch, let alone a sample. It is the only figure in the ladder that runs backwards: it is derived from the price and the margin the plan needs rather than built up from fabric and labour. Because it is a landed figure, it has to be worked back to a target FOB through the plan’s assumptions about freight, duty and the other landing charges, and that working-back is the first place the figure acquires an assumption nobody writes down.

Its job is to discipline development. Every fabric choice, construction and trim is a decision against the target, and a costed sample that comes in above it is a margin decision whatever it is called in the room. The target belongs to merchandising and planning, because it is the margin plan expressed per unit; in practice it is typed into PLM as the cost objective development works to, which is where its ownership starts to blur. Why the margin is largely settled by the time the target has been met or missed is the argument of the margin is made before the buy.

Who sets it, and when

Merchandising and planning, at line-plan approval, before development begins.

System of record

The line plan and the plan of record; copied into PLM as the development cost objective.

The margin it implies

The planned initial markup — the margin the range was built to deliver at the ticket price, if the style lands at target.

Where it goes wrong

It is derived from a landed figure through unwritten freight, duty and currency assumptions, and the FOB target development works to inherits all of them.

Quoted or FOB cost

The quoted cost is the factory’s or agent’s price for the finished garment at an agreed handover point — FOB, free on board, with the goods loaded on the vessel at the port of export; ex-works or FCA; at the other end, DDP with freight and duty inside it. The quote is a first cost with a delivery term attached. It is the first number in the ladder that another company wrote, and it arrives in the factory’s currency, at the factory’s assumed quantity, against a sample that may not yet be the final construction. The delivery terms themselves, and what each one leaves to be added before the price is a landed cost, are set out stage by stage in the apparel purchase order lifecycle.

A quote is also a negotiating position. It moves between the first costed sample and the purchase order as fabric is substituted, consumption is tightened, a trim is sourced locally or a quantity commitment changes the price break, and each movement is a version. PLM is the natural home of the versions because the bill of material they are built on lives there. What the quote is not is a cost the margin can be read from: an initial markup computed on an FOB quote flatters the style by the whole landing stack, and the flattery is largest exactly where freight and duty are heaviest.

Who sets it, and when

The factory or sourcing agent, on the costed sample, revised through negotiation to purchase order.

System of record

PLM, as costing versions against the bill of material; the final quote becomes the purchase order unit price in the ERP.

The margin it implies

None on its own. It is a first cost with a delivery term, and a markup read on it overstates margin by the landing charges it excludes.

Where it goes wrong

It is read as the cost in the line review, in the factory’s currency, at a quantity the buy may not reach.

Bill-of-materials cost

The bill-of-materials cost is the cost of the garment built from its components: fabric by consumption and price per unit of length, trims and findings, labels and packaging, and beside them the cut, make and trim labour, the factory’s overhead and its margin. It is the anatomy of the quote. Where the quote is a single number the factory offers, the BOM cost is the brand’s own reconstruction of how that number is made, and the difference between the two is what the negotiation is about.

It lives in PLM because that is where the bill of material lives, and it changes whenever the bill of material does: a fabric moved from one mill to another, a consumption tightened by a marker change, a trim replaced. The BOM cost is the only figure in the ladder that explains itself — every other cost is a total, and this one is a sum — which makes it the figure to negotiate from and the figure development uses to hold the target. It is also the figure that varies most by colorway and by size: a printed colorway carries a different fabric price from a dyed one, and an extended size consumes more fabric than the base size the quote averaged across. Held at style, it says the style is on target while one of its colorways is not.

Who sets it, and when

Product development and sourcing, as the bill of material is built and revised through sampling.

System of record

PLM, at style-color, against the bill of material and its consumption.

The margin it implies

The margin the style would deliver if the factory charged exactly the built-up cost — components, labour, overhead and its own margin — a ceiling for the negotiation, not a figure to plan on.

Where it goes wrong

It is kept at style when fabric price and consumption differ by colorway and by size, so the average hides the colorway that is below target.

Standard cost

The standard cost is the single per-unit cost the ERP carries for a style-color — sometimes for the style — and values inventory, cost of goods sold and purchase price variances against. It is set at a point in time, usually from the purchase order cost plus a standard uplift for freight, duty and the other landing charges, and it stays fixed until someone re-sets it. It is a convention of cost accounting: it lets every receipt, transfer and sale be valued at a stable figure while the real costs arrive on their own timetable, and it collects the differences as variances.

That is its strength and its planning problem in one. A standard cost is a decision about what the cost will be assumed to be, not a measurement of what it was, and the variances it throws off — purchase price variance when the price paid differs from the purchase cost inside the standard, freight and duty variance when the landing charges differ from the uplift — are posted to finance accounts by period, not to the style the planner is reading. In a connected read the standard is a useful midpoint: it is the cost the business has agreed to value the goods at. In a disconnected read it is one more number called cost that neither the merchant nor the planner set. What an apparel-specific ERP holds the standard against, and at which grain, is covered in what is an apparel and footwear solution.

Who sets it, and when

Finance and the ERP team, at style set-up or purchase order, from the PO cost plus a standard landing uplift; re-set on a cycle.

System of record

The ERP item or style-color record; inventory valuation and cost of goods sold run on it.

The margin it implies

The margin the books carry between re-sets: retail less standard, with the real cost arriving later as a variance line.

Where it goes wrong

The variances that make it true are posted by period to finance accounts, so the style’s real cost is known to the ledger and not to the plan.

Landed cost

Landed cost is the full cost of one unit delivered to the brand’s own warehouse: the FOB or first cost, plus inbound freight, duty and tariffs, cargo insurance, customs brokerage, inland delivery to the distribution centre and, where a buying agent is used, its commission. It is the cost initial markup should be read on and the cost open-to-buy should be consumed at, because it is the cost the goods will actually have carried by the time they can be sold. The flagship’s landed cost glossary entry sets out the components and why planning decisions run on it rather than on first cost; the point here is what kind of number it is at each stage of the style’s life.

Definition — Landed cost in apparel

Landed cost in apparel is the full cost of one unit delivered to the brand’s own warehouse: the FOB or first cost paid to the factory, plus inbound freight, duty and tariffs, cargo insurance, customs brokerage, inland delivery and, where one is used, the buying agent’s commission. It is an estimate at costing and at purchase order, built from assumed freight, duty and currency, and a fact only after the factory, forwarder and customs invoices have been matched and allocated to the units that landed — which is why the plan’s margin and finance’s margin are read from two different landed costs for the same style.

Used by: Planning, sourcing, finance and merchandising teams

Related: FOB, first cost, Incoterms, duty, freight allocation, customs value, standard cost, purchase price variance, initial markup, open-to-buy

Landed cost is an estimate three times and a fact once. At line plan it is the target. At costed sample it is the quote plus the landing assumptions. At purchase order it is the PO cost plus assumed freight, assumed duty and an assumed exchange rate — a planning estimate with three assumptions in it, each of which will be settled by someone else, weeks or months later. At receipt it becomes the actual, once the factory invoice, the forwarder’s invoice and the customs entry have been matched and allocated to the units that landed. The estimates belong to planning, development and sourcing. The fact belongs to finance. The season’s margin was planned on the estimates and will be reported on the fact, and nothing in a disconnected stack carries the fact back to the estimate at the level of the decision that set it.

Who sets it, and when

Planning as a target, development and sourcing as the costed-sample and PO-time estimates, finance as the actual after receipt.

System of record

The plan of record for the estimate; the ERP costing record and accounts payable for the actual.

The margin it implies

The real initial markup, and the margin open-to-buy should be consumed at — planned on the estimate, delivered on the actual.

Where it goes wrong

The actual is calculated in finance after receipt and never travels back to the plan, so the season’s margin stays one estimate away from the truth.

Actual invoiced cost

The actual invoiced cost is what the brand was billed and paid: the factory’s invoice for the goods at the quantity shipped, the forwarder’s invoice for the freight, the broker’s invoice for duty and clearance, the inland carrier’s invoice for the last leg, the agent’s commission invoice where there is one. Each arrives separately, from a different company, weeks apart, in whatever currency the contract named, and each is matched — to the purchase order and the receipt for the goods, to the shipment for the freight, to the entry for the duty — in accounts payable.

It is the only figure in the ladder that is a fact about money rather than a figure about a product, and it is the one the margin is finally measured on. It is also the figure held at the wrong grain for planning: invoices are per shipment, per container and per entry, so the cost of one unit of one style-color-size is an allocation of several documents, done by finance, using rules the planner has never seen. Whether freight is spread by unit, by cube or by value, and whether duty is assigned by line or by entry, changes which style absorbs the cost — and the allocation is correct by its own rule and invisible to the plan either way.

Who sets it, and when

The factory, the forwarder, the customs broker, the carrier and the agent, as invoices; finance, as the matched and allocated result.

System of record

Accounts payable and the ERP costing record, by invoice, shipment and customs entry.

The margin it implies

The realised cost side of gross margin — the figure the season is reported on, by period, after the fact.

Where it goes wrong

It is held per shipment and per entry, so the per-unit cost of a style-color-size is an allocation finance makes by rules the plan never sees.

Six figures, who sets each, where it lives, and the margin it implies

The six figures in one view. The last column is the one to read twice: it names the margin each figure can honestly support, and two of the six — the quote and the bill-of-materials cost — support no margin a plan should be read on.

Three price figures that get called cost

Three more numbers sit beside the ladder and get pulled into the same conversation. The ticket price is the first retail the style is offered at. The wholesale price is what a retail partner pays for it, which for a wholesale brand is the revenue the cost is read against. The realised average unit retail is what the consumer actually paid once promotions and markdowns have done their work. None of them is a cost, but every margin figure is a relationship between one of them and one rung of the ladder, so a margin figure that does not name both ends is not a margin figure. Initial markup is ticket against landed cost, planned; the flagship’s initial markup glossary entry carries the formula. Maintained markup is realised average unit retail against the cost of the units sold; the maintained markup entry covers what erodes the first figure into the second. Wholesale margin is the wholesale price against landed cost, before chargebacks and allowances.

A brand that sells direct and wholesale reads two margins on the same landed cost, and the colorway that clears its wholesale margin may not clear its direct one once the direct channel’s markdown cadence is in. Why adding wholesale is a costing decision before it is a distribution decision is the argument of the second channel changes the company. The ladder says what a unit cost. The price figures say what it earned. The margin is the pairing, and the pairing has to be named.

Five systems, five versions of the same cost

Ask five teams what a style costs and each will give a figure it can defend from its own system. The product team reads PLM. Sourcing reads the purchase order. Planning reads the plan. Finance reads the ledger. The channel team reads the price file and does not think it is talking about cost at all. Each system is the honest system of record for one or two rungs of the ladder and a stale copy of the rest, and the trouble starts when a copy is read as a record. The split below follows the rule the system of record map applies to every data object: one writer per figure, everyone else reads, and every reader says which figure it read.

PLM — bill of material, target as objective, quoted cost

PLM is the system of record for the product, and so for every cost that is a property of the product’s construction: the bill of material with its consumptions and component prices, the cost objective development is working to — the target, copied in from the line plan — and the costing versions that run from the first costed sample to the quote the purchase order is cut at. It holds them at style, sometimes at style-color, almost never at size, and it holds them in the currency the factory quoted.

What PLM is not the record for is anything that happens after the quote: the price the order was actually placed at if it moved in negotiation after the costing was closed, the landing charges, the actual. A costing sheet in PLM is a development document, and it stops when development does. Where a brand has no PLM, the same sheets live in a costing workbook per season or per factory, with the same scope and without the versioning.

Sourcing and the ERP — purchase order cost, standard cost, landed-cost accruals

The ERP is the system of record for the transaction, and so for every cost that is a property of a commitment: the unit price on the purchase order line, in the order currency, at the delivery term the order names; the standard cost the style-color is valued at; and the accruals — freight, duty, brokerage, commission — the ERP or its costing module books against the receipt before the invoices arrive, usually at the standard uplift rates. Sourcing writes the first of those by cutting the order and agreeing amendments; finance and the ERP team set the second; the costing module or the accounting team books the third.

This is the layer the PO lifecycle runs through, and in a disconnected stack it is the layer where the cost is right and unread by the people who planned it. The order cost is right. The standard is a deliberate convention. The accruals are estimates by design, trued up when the invoices land. None of them is the planned margin, and the ERP has no object to hold one — which is why the plan and the ERP begin to disagree at the first cost-changing amendment, as the purchase order lifecycle follows in detail. The apparel-specific version of this layer, and the records it adds to a generic ERP, is the apparel and footwear solution.

Planning — planned landed cost and planned margin

The plan is the system of record for the planned landed cost and for the margin it implies against the planned retail — the only two figures in this section that are decisions about the future rather than records of a transaction. The planned landed cost starts as the target, is replaced by the quote plus the landing assumptions when the costed sample is read, is replaced again by the PO cost plus the same assumptions when the order is committed, and should be replaced a last time by the actual when it is known. The planned margin moves with it, and so does the open-to-buy the order consumes, if the brand consumes open-to-buy at landed cost.

In a disconnected stack the plan holds whichever of those replacements someone remembered to make. A planned cost that was never re-keyed after the costed sample is not a planned cost; it is a target with a date on it, and the margin read from it is the margin the range was meant to have rather than the margin it is going to have. Which versions a plan of record must hold, and who holds the pen on each, is set out in the plan of record.

Finance — the actual landed cost after freight, duty and commission

Finance is the system of record for what the goods actually cost: the matched factory invoice, the forwarder’s invoice allocated to the shipment’s lines, the customs entry’s duty assigned to the goods it was assessed on, the broker’s and carrier’s charges, the agent’s commission, each at the exchange rate on the day it was settled or the hedged rate the treasury applied. The actual landed cost per unit is the sum of those allocations, and it is a fact — the only rung on the ladder that is.

Finance holds it at the grain it was incurred at — by invoice, shipment and entry — and reports it by period, through cost of goods sold and the variance accounts that reconcile actual to standard. The actual is complete, correct and in a shape no planning decision can read: it is a shipment’s freight and a period’s variance, not a style-color’s cost beside the estimate the buy was approved on. Carrying it back into that shape is the single most valuable unmade handoff in costing, and nobody in a disconnected stack owns it.

Retail and ecommerce — ticket price and realised average unit retail

The channel systems — the price file behind the stores and the storefront, the promotions engine, the point of sale and the ecommerce order record — are the system of record for what the unit was offered at and what it was sold for: the ticket price by channel and market, the promotional and markdown prices by date, and the realised average unit retail that results. They hold no cost at all, or a stale standard pulled across for a margin report. They are in this section because the margin is a pairing, and they own one end of it.

The planning consequence is that the realised margin is computed in a system that owns the price and borrows the cost, and the cost it borrows is whichever one was pushed to it — a standard, a planned landed cost, occasionally an FOB. A merchandising report that shows gross margin by style is pairing a true realised price with a cost of unknown rung, and it looks exactly as authoritative as one that is not. Why the full-price figure the same report shows has no system of record of its own is the subject of the full-price number nobody owns.

Seven reasons the same style costs a different amount in each system

None of the disagreements below is an error, and that is what makes them durable. Each system applies its own correct rule to its own correct input and arrives at a different figure for the same garment. Reconciling harder does not close them; naming the mechanism does, because each one is a decision the brand can make once instead of an argument it has every season.

Timing — quote, purchase order, invoice

The quote is written at the costed sample, months before the order. The purchase order is cut later, after negotiation, sometimes at a different price. The factory invoice is raised at shipment, at the order price unless an amendment moved it. The forwarder and the broker invoice later still. Each system holds the figure that existed at the moment it received one, and a system that is not told about a later event keeps the earlier figure as its truth. PLM holds the quote; the ERP holds the order; finance holds the invoice; and the plan holds whichever of the three was current when it was last touched. The three can be different for a style whose cost moved exactly as it was supposed to.

Currency

A factory in one country quotes in one currency, often a trading currency rather than its own; the brand plans in its home currency; the retail partner buys in a third. The quote is converted at the rate on the costing date, the plan is built at a planning rate finance set for the season, the ERP posts the order at the rate on the order date and the receipt at the rate on the receipt date, and treasury settles the invoice at the spot or hedged rate on payment. Five rates for one price. The movement between them is a real cost or a real saving, and a stack that folds it silently into landed cost cannot tell a currency effect from a freight increase, so the lesson for next season’s costing is lost in the same line as the lesson for next season’s logistics.

Duty, and the value it is assessed on

Duty depends on the classification of the garment — fibre content, construction, gender and use all move it — on the country of origin, and on any preferential agreement between that country and the market of import. The rate the plan assumed at costing was for the fabric and the origin on the costed sample, and a fabric substitution or a move to a second factory in another country changes it after the plan has been approved. The value the duty is assessed on also differs by market: under the transaction-value rules of United States customs valuation it is generally the price paid for the goods excluding international freight and insurance, while under the European Union’s Union Customs Code it includes the cost of transport and insurance to the point of entry into the customs territory. The same FOB price, at the same rate, attracts a different duty amount depending on where the goods clear. Finance then assigns the entry’s duty to lines by its own rule, and a planned duty uplift and an actual duty allocation can both be right and still differ by the whole reclassification.

Freight allocation

Freight is charged per container, per shipment or per kilogram, never per garment, and a part-air shipment to hold a delivery date is charged on a different basis again. Turning a forwarder’s invoice into a cost per unit is an allocation, and the brand chooses the rule: by unit, which charges a camisole the same freight as a parka; by cube or weight, which follows what the forwarder actually charged for; or by value, which loads freight onto the expensive styles. Each rule is defensible. Each produces a different landed cost for every style in the container. The planning uplift assumed one rule — usually a percentage of FOB, which is a value rule — and finance applied another, so the plan and the actual disagree on freight for every style even in a season where total freight came in exactly on budget.

Minimum-order surcharges and price breaks

A quote is a price at a quantity. Below the factory’s minimum for a colorway, a fabric or a size, the price rises — as a stated surcharge, as a fabric minimum the brand has to take anyway, or as a quietly higher quote next season — and above a volume break it falls. The quantity the quote assumed was the line plan’s intended depth; the quantity the order carries is the buy plan’s, after the assortment review trimmed a colorway or the open-to-buy cut the depth. So the purchase order price for the trimmed colorway is above the quote PLM holds, and the style-level average in the ERP sits between them. The surcharge is a cost event caused by a planning decision, which is why it should be visible to the planning decision that caused it rather than discovered on the order confirmation.

Sample and development cost

Proto samples, fit samples, salesman samples, pattern and grading work, lab dips and strike-offs, testing and certification all cost money, and brands treat that money three ways: expensed in the period as a development cost, carried against the style as a development cost that never enters the unit cost, or amortised into the unit cost across the planned quantity. The factory may do the same on its side, folding sample charges into the quote or billing them separately. A style costed with development amortised in carries a higher unit cost than the same style costed without, and two teams comparing their cost figures can differ by exactly the development spend without either having made a mistake. When the planned quantity falls, an amortised development cost per unit rises, which is a second cost event caused by a planning decision.

Roll-ups at style, colorway and size

Cost varies below the style. A printed colorway carries a different fabric price from a solid; a colorway below the dye minimum carries a surcharge; an extended size consumes more fabric and may carry a size surcharge, while the quote averaged consumption across the size run. PLM holds cost at style or style-color. The purchase order holds price at the line, which is style-color-size or a prepack. Finance holds the actual at the shipment. Each level is the honest grain for its own purpose, and a figure rolled up to the style is an average that hides every colorway and size that is off target — in both directions, so a style on target can contain a colorway below target and a size above it. Why opposite errors cancel at every level above the one they occur at is the subject of the planning grain; the structure that lets a cost be held at the right level and rolled up rather than averaged is the subject of the apparel data model.

The target is set

The first cost event is a margin decision disguised as a cost. When the line plan gives a slot a ticket price and a target initial markup, it has set a maximum landed cost and, through the landing assumptions, a target FOB. The number is written by merchandising and planning in the line plan and handed to development as a cost objective — typed into PLM, or into the costing sheet, by someone in product development.

Two things are lost at the handoff. The landed basis: the figure arrives in PLM as an FOB, and the freight, duty and currency assumptions that turned a landed target into an FOB objective are not on it. And the margin: development sees a cost to hit, not the markup the cost protects, so a trade-off that moves the cost by a little and the duty classification by a lot is judged on the cost alone. The target is the most important number in the ladder and the one with the least context attached to it when it reaches the people who work to it.

When it happens

At line-plan approval, before any sample exists.

Who writes the number

Merchandising and planning, as a landed figure derived from the ticket price and the target initial markup.

Where it lands

The line plan and the plan of record; PLM or the costing sheet as the development cost objective.

Where it is re-keyed

Into PLM as an FOB objective, by hand, with the landing assumptions and the margin stripped off it.

The factory prices the sample

The costed sample is the first time the garment has a cost that someone outside the brand wrote. The factory or agent prices the sample — a quote at a delivery term, in its currency, at an assumed quantity — and development records it in PLM against the bill of material, usually as one of several costing versions as the sample is revised and the quote moves.

The planning consequence is immediate, and in a disconnected stack unread. The planned landed cost should move from the target to the quote plus the landing assumptions, and the planned margin should move with it. In a disconnected stack it does not: the quote is compared with the FOB objective on the costing sheet, the gap is negotiated or accepted, and the plan keeps the target. A style accepted two points under target at the costed sample enters the buy as a style on target, because the only document that changed was a development document, and the buy is not read from development documents.

When it happens

When the factory or agent prices the costed sample, often across several revisions.

Who writes the number

The factory or agent, recorded in PLM by product development against the bill of material.

Where it lands

PLM costing versions; the costing sheet where there is no PLM.

Where it is re-keyed

Nowhere, unless a planner re-keys the quote and rebuilds the landed estimate by hand; otherwise the plan keeps the target.

The purchase order cost is locked, and then amended

The purchase order fixes the price the brand will pay: the negotiated FOB, in the order currency, at the quantity by size the buy plan settled on, under the delivery term the order names. Sourcing writes it by cutting the order in the ERP, and from this moment the ERP is the record for the price. Two further things happen in the same event. The standard cost is set or confirmed, usually as the order price plus the standard uplift. And open-to-buy is consumed — at the landed estimate, at first cost or at retail, depending on a choice the brand should have made once and written down.

The handoff that fails here runs in the other direction from the first two. The plan should read the order price and re-cost itself; instead the buyer or the planner types the order quantities into the plan and leaves the cost as it was. The order is complete in the ERP and the plan is still costed at the sample, or at the target, and the two disagree by exactly the negotiation.

The event then recurs. Every cost-changing amendment — a price change, a ship-mode change to air to hold a date, a quantity change that crosses a minimum and triggers a surcharge, a currency clause invoked — is written in the ERP, or agreed in email and then written in the ERP, and reaches the plan only if someone carries it. Amendments are the subject of their own section in the purchase order lifecycle, because they are where the plan and the ERP diverge for quantities and dates as well as for cost.

When it happens

At purchase order issue, and again at every cost-changing amendment.

Who writes the number

Sourcing or buying, on the purchase order in the ERP; the factory, on the confirmation and each amendment.

Where it lands

The ERP purchase order line; the standard cost; the open-to-buy consumed in the plan.

Where it is re-keyed

Into the plan as quantities by size, with the price only if the planner carries it; amendments into the plan only when someone remembers.

The actual landed cost is known

The fourth event is the only one that produces a fact. After receipt, the factory invoice is matched to the order and the receipt, the forwarder’s invoice is allocated across the shipment’s lines, the customs entry’s duty is assigned to the goods it was assessed on, and the broker’s, carrier’s and agent’s invoices follow. Finance writes the actual landed cost per unit and posts the difference to standard as purchase price, freight and duty variances. Inventory stays on the books at the standard; the real cost of the goods is the standard and the variances read together, which is the shape the next paragraph describes.

It is written weeks after receipt, by a team that does not own the plan, at a grain the plan cannot read. The actual is the one figure every earlier rung was an estimate of, and it is the one figure that never travels. In the worked example below it lands 1.75 a unit above the planned landed cost. The plan is still carrying 46.25 and reporting 63.0%; the colorway is being sold at 61.6%; and the person who could act on the difference — by moving the next delivery’s retail, by changing the ship mode on the open orders, by re-costing the carryover — has not been shown it.

When it happens

Weeks after receipt, as each invoice is matched; complete only when the last one is.

Who writes the number

Finance, from the factory, forwarder, broker, carrier and agent invoices.

Where it lands

Accounts payable and the ERP costing record; variances to standard on finance accounts, by period.

Where it is re-keyed

Nowhere. It reaches the plan only as a period-level freight or duty variance in a finance review, if at all.

The season’s cost is closed

The fifth event closes the season’s cost. Invoices that arrived after period close, a duty reclassification or refund, a freight rebate or surcharge settled with the forwarder, a currency gain or loss on payment, a sample charge the factory billed late, a chargeback from a retail partner that the brand books against the goods — each adjusts the cost of goods already sold or still in stock. Finance writes it at the level it was incurred at and books it in the period it was settled.

Two things should happen. The books are corrected. In a disconnected stack the costing for next season is not: the true-up tells the brand what the landing stack really cost for this origin, this lane and this classification, and that is exactly the assumption the next line plan’s target will be built on — built instead from last year’s uplift rates, because nobody carried the actual back into them. The true-up is the cost event that should set next season’s first cost event, and in a disconnected stack nothing connects the two but a planner’s memory.

When it happens

At or after period close, as late invoices, refunds, rebates and currency effects settle.

Who writes the number

Finance, against the goods and the period they belong to.

Where it lands

Cost of goods sold and inventory valuation; the finance view of the season’s margin.

Where it is re-keyed

Into next season’s landing assumptions, if anyone carries it; otherwise the next target is built on last year’s uplift rates.

Cost event, who writes it, where it lands, and where it is re-keyed

The five events in one view. The last column is the diagnostic: it names the handoff a disconnected stack leaves to chance, and the second row’s entry — the plan, only by hand — is the one that costs the most, because it is the earliest.

One style, one ticket price, every rung of the ladder

The numbers below are illustrative, chosen so the arithmetic is visible rather than to describe any brand, factory or market; the freight, duty and other rates are planning assumptions for the example, not benchmarks. Initial markup throughout is on retail — retail less cost, over retail — which is the figure the flagship’s initial markup entry defines; markup on cost is a different figure, and the two should never be read as one.

Event one: the target

A brand plans a woven shirt into its spring range at a ticket price of 125.00 in its own currency, with a target initial markup of 64.0%. The target landed cost is therefore 125.00 × (1 − 0.64) = 45.00. The plan’s landing assumptions for this origin and ship mode are freight at 5% of FOB, duty at 17% of FOB and a further 3% of FOB for insurance, brokerage and inland delivery — 25% in all — so the FOB objective development is given is 45.00 ÷ 1.25 = 36.00. That is the first cost event: a landed figure derived from a margin, carried into PLM as an FOB objective with three assumptions attached and none of them written on it. Read on FOB alone, the objective shows an initial markup of (125.00 − 36.00) ÷ 125.00 = 71.2%, which is the first of several flattering numbers the style will carry.

Event two: the costed sample

The costed sample comes back from the factory at an FOB quote of 38.00 — 2.00 over the objective. Read on the quote alone, the style shows an initial markup of (125.00 − 38.00) ÷ 125.00 = 69.6%, and that is the figure the line review sees if the costing sheet is what it reads. Read on a landed basis, the quote becomes 38.00 + 1.90 freight + 6.46 duty + 1.14 other = 47.50, and the initial markup is (125.00 − 47.50) ÷ 125.00 = 62.0%: two points below target, where the FOB read showed 5.6 points above it. Whether the planned margin moves to 62.0% now depends on whether anyone rebuilds the landed estimate from the quote.

Event three: the purchase order

Sourcing negotiates. A trim is swapped and the marker is tightened, and the purchase order is cut at 37.00 FOB. The planned landed cost, at the same assumptions, is 37.00 + 1.85 + 6.29 + 1.11 = 46.25, and the planned initial markup is (125.00 − 46.25) ÷ 125.00 = 63.0%. On FOB alone it reads (125.00 − 37.00) ÷ 125.00 = 70.4%. The ERP’s standard cost for the colorway is set at 46.25 — the order price plus the standard uplift — and the 1,200 units of the colorway consume 1,200 × 46.25 = 55,500 of open-to-buy at the planned landed cost. The plan now carries 63.0% if the order price reached it, 62.0% if the quote did and the negotiation did not, and 64.0% if nothing after the line plan did.

Event four: the actual

The goods ship. To hold the delivery date after a fabric delay, part of the order travels by air, and freight rates on the lane have risen since costing. When the forwarder’s invoice is allocated, freight comes to 3.60 a unit against the 1.85 estimated. Duty is assessed at the assumed classification, 6.29; the other charges come in at the 1.11 estimated. The actual landed cost is 37.00 + 3.60 + 6.29 + 1.11 = 48.00, and the initial markup actually achieved at ticket is (125.00 − 48.00) ÷ 125.00 = 61.6%. Against the standard of 46.25, finance posts a freight variance of 1.75 a unit — 1,200 × 1.75 = 2,100 on the colorway — in the period the invoice arrived. The margin plan does not change, because nothing in the stack tells it to.

Set the reads side by side. The line review saw 69.6% on the quote, or 70.4% once the order price was known, if it read FOB. The plan carried 64.0% if nobody re-keyed the target after the costed sample, 62.0% if someone did and stopped there, 63.0% if the order price reached it. Finance will report the colorway at 61.6%. Every figure is correct for its basis, and the spread between the most flattering read of the price actually paid and the true one is 70.4 − 61.6 = 8.8 points of margin on a style whose retail never moved. Against the 64.0% the range was built to deliver, the colorway is 2.4 points short, and nobody who approved it at 63.0% has yet been told. The last column is the quiet lesson: the FOB read never changes after the order, however much the landed cost does, which is why a brand that reads FOB believes its costing is stable.

The colorway below the minimum

The same style has a second colorway. The assortment review trimmed it to 400 units, which is below the mill’s illustrative dye minimum of 600 units for that shade, and the factory applies a surcharge of 0.80 a unit: 37.80 FOB. At the same landing assumptions its planned landed cost is 37.80 × 1.25 = 47.25 and its planned initial markup is (125.00 − 47.25) ÷ 125.00 = 62.2%. The style-level order cost the ERP shows is the weighted average, (1,200 × 37.00 + 400 × 37.80) ÷ 1,600 = 59,520 ÷ 1,600 = 37.20, which lands at 46.50 and reads 62.8%.

At style level the buy reads 62.8%, 1.2 points under target, and nothing in that figure says that colorway B is 1.8 points under while A is 1.0 under. The decision the colorway read supports — drop B, raise its retail, accept the surcharge knowingly, or add the 200 units that would clear the minimum if the open-to-buy allows — is unavailable to a plan that holds cost at style. And the surcharge was caused by the assortment review that trimmed B to 400 units: a planning decision that changed the cost without being shown the cost it changed. This is the roll-up mechanism from the previous section in a single style, and it is the reason the grain a cost is held at is a planning decision rather than a data-entry convenience.

The arithmetic here is deliberately plain, and the free tools on retail-plan.com run it on real figures: the landed cost calculator takes an FOB cost with freight, duty and other charges to a landed cost per unit and the gross margin it yields at a target retail, and the initial markup calculator reads initial markup on retail and markup on cost from the result. What neither can do, and what this page is about, is make the figure they produce the one every team reads.

What the planning side needs from costing

The ladder is not an accounting curiosity. Four planning decisions depend on reading the right rung at the right time, and each one is made wrong in a specific way when it reads the wrong one. Naming the decision names the figure it needs, which is a more useful way to specify a costing workflow than listing the fields a cost record should carry.

The line review: target against quote, on a landed basis

The line review decides which styles stay in the range at which retail. It needs the target landed cost and the costed sample’s quote brought to the same landed basis, side by side, by colorway — not the FOB objective against the FOB quote, which is how the costing sheet shows them. In the worked example the FOB comparison shows 2.00 over objective and the landed comparison shows 2.0 points of margin under target; they are the same fact, and only the second is a decision the review can take. A line review that reads FOB approves styles on target that are not, and the range it approves is under its planned margin before a single order exists.

Buy approval and open-to-buy: one figure, consumed once

Buy approval commits units by size against open-to-buy. It needs the planned landed cost at the purchase order price — the rung the order will actually be cut at — and it needs open-to-buy consumed at that same figure, so that the margin the buy was approved on and the budget it spent are one number. A buy approved at the target’s margin and consumed at the order’s cost has spent more than it reported; a buy approved at the order’s margin and consumed at first cost has reported a budget it did not spend. The choice of which figure consumes open-to-buy is a planning rule, made once; what open-to-buy is and how receipts draw on it is covered in the flagship’s open-to-buy planning guide.

The in-season read: actual against planned, by colorway

The weekly and monthly reads decide markdowns, chases, reallocations and the second delivery’s retail. They need the actual landed cost of the units that landed, by colorway, beside the planned landed cost those decisions were built on — in the period the actual is known, not at period close. A markdown decision that reads planned cost is deciding how much margin to give away from a figure that is already wrong; in the worked example it would be giving margin away from 63.0% that is really 61.6%, and the markdown that looked affordable at the first figure is not at the second. The read belongs on the clock set out in the operating cadence, and the actual has to arrive on that clock to be useful.

Next season’s target: the true-up as the first input

The next line plan sets its targets through the same landing assumptions this season’s targets were built on, and the only honest source for those assumptions is this season’s actuals: freight by lane and ship mode as it was actually allocated, duty by classification as it was actually assessed, the currency effect as it was actually settled, the surcharges the minimums actually triggered. The true-up is where those are known. A planning function that sets next season’s targets from last year’s uplift rates is planning on a costing it already knows to be wrong, and the gap between the rates it used and the rates it paid is the first number the next season will miss by.

What a connected stack does instead

A stack can hold every cost figure correctly in its own system and still be disconnected, because the failure is not in any figure. It is in the fact that the plan’s margin is read from a copy, and the copy was made on a date. Integrated means each system holds its rung of the ladder accurately; connected means the plan reads the rung rather than a typed copy of it, and changes when the rung does. The first is what PLM, the ERP and the finance system each already do. The second is a workflow decision about where the margin is read from and what happens when a cost moves, and it rests on three arrangements that are described below as work, because that is what they are.

Definition — Connected costing

Costing is connected when every function reads the same costed record for a style — one that holds each cost figure with its basis, its currency, its date and its owner, rather than one field called cost — and when a cost event written by any one of them re-flows into the planned margin of the plan of record as a new version. In practice that means a cost change anywhere re-prices the plan, the margin is read at the colorway and size the cost actually varies at, and a landed cost that lands above target by more than an agreed threshold opens a review while the decision it affects can still be taken.

Used by: Merchandising, planning, sourcing, finance and technology leaders deciding where cost should live

Related: Costed record, plan of record, cost event, landed cost, planned margin, cost basis, review threshold, system of record, merchandising operating layer

One costed record the plan reads from

The first arrangement is that there is one costed record per style-color — one place that holds every rung of the ladder side by side, each with its basis, its currency, its date and its owner — and that the plan of record reads its planned landed cost from it rather than holding a number someone typed in at the line review. PLM still writes the bill-of-materials cost and the quote. The ERP still writes the order price and the standard. Finance still writes the actual. What changes is that the plan stops being a fourth writer of cost and becomes the reader of the other three, and that each figure it reads says which rung it is, so a margin shown anywhere can be traced to the cost it was read on.

This is the costing case of the general rule in the system of record map: one writer per object, everyone else reads. The record is held at the grain the cost varies at — colorway where fabric price and minimums differ by color, size where consumption or a surcharge does — and the style figure is a roll-up of those, computed rather than typed. A single field called cost on the style master is the opposite of this arrangement, however accurately it is maintained.

Cost events that re-flow into planned margin

The second arrangement is that each of the five cost events produces a new version of the planned margin in the plan of record, stamped with the event that caused it, without anyone re-keying anything. A quote revised in PLM re-prices the planned landed cost of the style-color it belongs to and moves its planned margin, and the line review opens on the revised figure beside the target. An order cut or amended in the ERP does the same at commitment. An actual landed cost posted in finance replaces the estimate for the units that landed and moves the in-season margin read for the colorway. The true-up updates the landing assumptions the next target will be built on.

The re-flow is a workflow, not a report. The writer writes once, in the system that is the record for that rung, and the plan’s version increments; the reviewer sees the margin delta by style-color and in aggregate, with the cause attached, instead of a rebuilt spreadsheet whose numbers moved for reasons it cannot say. What a versioned plan has to hold for this to work — the prior value, the reason, the owner and the date on every change — is the subject of the plan of record, and which clock is allowed to act on a given cost change is set out in the operating cadence: a quote moves on the seasonal clock, an amendment on the monthly one, an actual on the weekly read, a true-up at close.

A review when landed cost exceeds target by a threshold

The third arrangement is a rule. For each category the brand agrees a threshold — a number of margin points, or a cost per unit — above which a planned or actual landed cost that exceeds the target opens a review, and it names who convenes the review and who decides. The point of the threshold is timing: the earlier the event, the more decisions are still available. At the costed sample the review can change the construction, the fabric, the factory or the retail. At purchase order it can change the price, the quantity against a minimum, the ship mode or the retail, and it can accept the shortfall knowingly and write that down. After receipt it can only move the next delivery’s retail, re-cost the carryover and correct next season’s assumptions.

In the worked example, a threshold of one and a half margin points opens the review at the costed sample, where the quote landed 2.0 points under target; does not open it at purchase order, where negotiation had narrowed the gap to 1.0; and opens it again at receipt, where the actual landed 2.4 points under — the first time with four decisions still available on the style itself, and the last with none that changes the cost of the units already landed. Without the rule, the same style passes every review it is shown to, because each review reads a different rung and none of them reads the gap. Who should hold the decision at each stage is a question for the decision rights map; the threshold itself is a planning decision, agreed once per category and adjusted after a season.

Five tests

Five tests tell a stack with accurate costing from a connected one. They are deliberately operational: each can be checked against last season’s figures without a systems review, and a stack that passes the first and fails the rest is integrated, not connected.

1. 01
   Every margin figure on every surface names its cost basis.

   A line review, a buy approval, a weekly trade read and a period close each show which rung of the ladder they were read on — target, PO estimate, standard or actual landed — and two figures with different bases are never shown as one number.

2. 02
   A cost change written anywhere re-prices the plan without being re-keyed.

   A quote revised in PLM, a PO cost or ship-mode amendment agreed with the factory, a re-set standard, an actual landed cost posted in finance: each arrives in the plan of record as a new version of the planned margin, stamped with the event that caused it.

3. 03
   The plan reads cost at the grain the cost actually varies at.

   Colorway where fabric price and minimums differ by color; size where consumption or a surcharge differs by size; shipment where freight and duty were actually allocated. The style-level average is a roll-up of those, never the only figure held.

4. 04
   Actual landed cost returns to the style and the order that incurred it, within the period it is known.

   Not as a freight variance on a finance account, but as a cost against the style-color, read beside the estimate the buy was approved on, while the season it belongs to is still selling.

5. 05
   A landed cost above target by more than an agreed threshold opens a review while a decision is still available.

   At costed sample the review can change the construction; at purchase order it can change the price, the quantity or the retail; after receipt it can only move the next delivery’s retail, re-cost the carryover and correct next season’s assumptions. The threshold is agreed once, per category, and the review has a named owner.

Agreeing one cost definition across design, sourcing, planning and finance, step by step

None of the above can be bought. The three arrangements rest on definitions four functions have to agree, and the agreement is the work — a connected stack without it holds four private definitions of cost in one record and reconciles them faster. The sequence below is ordered so that each step makes the next one decidable: the ladder before the owners, the owners before the assumptions, the assumptions before the grain, and the events and the threshold last, because they are what the earlier steps make it possible to route. It is a working session or two per step, with the people who write each figure in the room, and it should produce a written document that every surface showing a margin can be checked against.

1. 01
   Write the ladder down

   Name every cost figure the brand uses — target, quoted or FOB, bill-of-materials, standard, landed estimate, actual landed — in the brand’s own words, with the basis each one includes and excludes. A term two teams use for different figures gets two names.

2. 02
   Give each figure one owner and one system of record

   PLM for the bill-of-materials, target-as-objective and quoted figures; the ERP for the purchase order cost, the standard and the landed-cost accruals; the plan of record for the planned landed cost and planned margin; finance for the actual. Every other system reads, and says which figure it read.

3. 03
   Fix the landing assumptions as data, not habit

   Freight as a rate by ship mode and origin, duty by classification and country of origin, agent commission by agent, insurance and brokerage and inland delivery by lane, and the exchange rate the plan is built at — each with an owner and a date, each visible to anyone who reads a landed estimate built from them.

4. 04
   Decide which figure consumes open-to-buy and which margin the line review reads

   Usually the planned landed cost for both, so that the margin approved in the line review and the budget the order consumes are the same number. Whatever is chosen, it is chosen once, and the retail-math surfaces are labelled with it.

5. 05
   Agree the grain

   Cost is held at colorway wherever fabric price or minimums differ by color, and at size wherever consumption or a surcharge differs by size; the style figure is a roll-up. The margin is read at the same grain, so a colorway below target cannot hide inside a style average.

6. 06
   Agree how currency is handled

   One planning rate per season, set by finance, applied to every quote and purchase order in the plan; the hedged or actual rate reconciled against it at receipt; the difference posted as a currency effect against the style, not folded silently into landed cost.

7. 07
   Agree the sample and development cost treatment

   Whether proto and salesman sample charges, pattern and grading work, testing and lab dips are expensed in the period, carried as development cost against the style, or amortised into its unit cost — one rule per cost type, applied the same way in PLM, the ERP and the plan.

8. 08
   Define the cost events and the return path for each

   Line-plan target, costed sample, PO cost locked, cost-changing amendment, actual landed cost, post-season true-up: for each, who writes the number, where it lands, and how it reaches the plan of record as a new version of the planned margin without being re-keyed.

9. 09
   Set the review threshold and its owner

   The variance from target, in margin points or in cost per unit, above which a costed sample, a purchase order or a receipt opens a review; who convenes it; and which decisions are still available at each stage. Then run one season on it and adjust.

This is a seam in the sense the connected planning implementation playbook uses — one handoff, one owner per phase, the data the seam needs, the failure modes — and the costing seam is a good early one because its inputs already exist in systems that hold them well. Where a brand stands on it, among the other seams, is one of the questions the operating model assessment asks.

Where the cost base behaves differently

The ladder holds in every merchandise vertical: there is always a target, a quote, a built-up cost, a book cost, a landed estimate and an actual. What changes is which component dominates the cost base and which cost event matters most, following the logic of the connected stack by vertical: each vertical has a cost component the apparel costing sheet was not designed around, and the question is whether the stack carries it as a figure the plan can read or as a note on a quote. Four verticals change the shape of the problem rather than its size.

The per-vertical treatment of the full stack, beyond costing, is on RetailNorthstar’s industry pages.

One style family, one season, six figures side by side

The tempting first project is a costing module. The useful one is a table. Pick one style family from last season — enough colorways to show the roll-up, enough imported volume for freight and duty to matter — and for each style-color collect the six figures from the systems that hold them: the target from the line plan, the quote and the bill-of-materials cost from PLM, the order price and the standard from the ERP, the planned landed cost from the plan, the actual from finance. Put them in one table with the basis, the currency, the date and the owner beside each, and read the initial markup on every rung. Then write down every decision the gap between the rungs would have changed: a colorway dropped or re-priced at the costed sample, a ship mode changed before the order, a retail moved on the second delivery. That list is the business case for connecting costing, and it is specific to the brand in a way no outside figure could be.

Then fix the landing assumptions as data — freight by lane and ship mode, duty by classification and origin, the planning rate — with an owner each, because every landed estimate in the table was built from them and nobody could say which version. Define the return path for the actual next: the one cost event that today reaches nowhere the plan reads, carried to the style-color and the order it belongs to, within the period it is known. Set the threshold last, once a season of the table has shown what size of gap is normal and what size is a decision. The sequencing logic, one seam at a time with one owner per phase, is in the connected planning implementation playbook.

The cost of the whole pattern across the business, rather than one number, is the subject of the cost of disconnected apparel workflows; the executive version of the margin gap this page works through in detail is why ERP and Excel still leave apparel margins exposed.

What one cost definition does not fix

One costed record makes the cost a style actually carries visible to the people deciding its margin, at the time they decide. It does not make the cost lower. Freight rates move with the market, duty follows the classification and the origin, a factory’s minimum is the factory’s, and metal is priced by an exchange. A forwarder will still consolidate the container after the order, and the actual will still arrive weeks after receipt, because invoices do. A brand that connects its costing has the same costs it had before; it has them earlier and in one shape.

Nor does it make the trade-offs. Whether to accept a surcharge to keep a colorway, to air-freight a late delivery and give up two points of margin to hold a floor date, to raise a retail or to take the shortfall — those are commercial decisions about margin, service and the range, and no record of the cost makes them for the brand. And the record is only as honest as the allocation rules behind it: a freight allocation by value is a decision too, and a connected stack that carries it carries its bias faithfully. What one agreed costing does is put every such choice in front of the people who own the margin while the choice is still open, with the number they will later be measured on already attached. That is enough to be worth doing.

Frequently asked questions

What is apparel costing?

Apparel costing is the set of cost figures a style carries through its life — target cost, quoted or FOB cost, bill-of-materials cost, standard cost, landed cost and actual invoiced cost — and the margin each of them implies against the ticket price. The target is set backwards from the price and the planned initial markup before the product exists; the quoted or FOB cost is the factory’s price at a delivery term; the bill-of-materials cost is that price rebuilt from fabric, trims, labour and factory overhead; the standard cost is the single figure the ERP values inventory at; landed cost adds freight, duty, insurance, brokerage, inland delivery and any agent commission to the first cost; and the actual invoiced cost is what was finally billed and paid. Each is set by a different team in a different system, and a margin read from any of them is only a margin figure if it names which one it used.

What is the difference between FOB cost, first cost and landed cost?

First cost is the factory’s price for the goods themselves. FOB — free on board — is a delivery term under the Incoterms rules: the price includes getting the goods on board the vessel at the port of export, and everything from that point is the brand’s to arrange and pay for, so an FOB price is a first cost with a defined handover point. Landed cost adds everything between that handover and the brand’s own warehouse: inbound freight, duty and tariffs, cargo insurance, customs brokerage, inland delivery and, where a buying agent is used, its commission. A quote sheet shows FOB; a margin plan needs landed. In the illustrative example on this page, a purchase order cost of 37.00 becomes a planned landed cost of 46.25 with freight at 5%, duty at 17% and other landing charges at 3% of FOB, and an actual landed cost of 48.00 once the forwarder’s invoice allocates freight at 3.60 a unit instead of the 1.85 estimated.

Which system should own product cost?

Cost has to be split by cost figure, because the figures are different objects sharing a name. The line plan and the plan of record own the target, set backwards from the ticket price and the planned initial markup; PLM carries it as the objective development works to and owns the bill-of-materials cost and the quoted cost. Sourcing and the ERP own the purchase order cost, the standard cost and the landed-cost accruals — sourcing writes the order price, finance and the ERP team set the standard, and the costing module books the accruals. The plan of record owns the planned landed cost and the planned margin it implies. Finance owns the actual landed cost, after freight, duty and any agent commission have been invoiced, matched and allocated. Retail and ecommerce own the ticket price and the realised average unit retail the margin is finally earned against. The mistake is treating them as one field: a merchant quotes initial markup from the figure PLM shows while finance reports margin from the figure the goods actually cost, and both are defensible even though they disagree.

Why does the same style have a different cost in PLM, the ERP and the plan?

Seven mechanisms, none of them an error. Timing: the quote, the purchase order and the invoice are written months apart and each system holds the one it received. Currency: a factory quotes in one currency, the plan is built at a planning rate, the ERP posts at the rate on the day, and finance settles at the hedged or spot rate. Duty allocation: duty is assessed on a customs value that differs by market and by entry, and assigned to lines by a rule the plan never sees. Freight allocation: freight is charged per shipment by cube or weight and spread to units by unit, by cube or by value. Minimum-order surcharges and price breaks: the quote assumed a quantity the buy did not reach. Sample and development cost: proto and salesman samples, patterns, testing and lab dips are expensed, carried or amortised into the unit cost depending on the team. Roll-ups: PLM holds cost at style or style-color, the purchase order at style-color-size, finance at the shipment, and an average at one level hides the variation at the level below.

How much does landed cost change initial markup?

By the whole landing stack, and unevenly across the range, because freight follows cube and weight rather than value while duty follows the classification. In the illustrative example on this page, a style ticketed at 125.00 with a purchase order cost of 37.00 reads 70.4% initial markup on the FOB cost alone, 63.0% on the planned landed cost of 46.25 and 61.6% on the actual landed cost of 48.00 — 8.8 points between the first read and the last, on one style, with no change to the retail price. Against the 64.0% the line plan targeted, the actual is 2.4 points short, and on the 1,200 units of the colorway the 1.75 a unit between the planned and actual landed cost is 2,100 of margin the plan never saw. The retail-plan.com landed cost and initial markup calculators run the same arithmetic on your own figures.

What is a standard cost in an apparel ERP, and why does it differ from the actual?

A standard cost is the single per-unit cost the ERP carries for a style-color — sometimes for the style — and values inventory, cost of goods sold and purchase price variances against. It is set at a point in time, usually as the purchase order cost plus a standard uplift for freight, duty and the other landing charges, and it stays fixed until it is re-set. The real costs arrive later and separately — the factory invoice, the forwarder’s invoice, the customs entry — and the differences between them and the standard are posted as variances to finance accounts by period rather than to the style. In the illustrative example on this page the standard is set at the planned landed cost of 46.25 and the actual landed cost is 48.00, so the 1.75 a unit posts as a freight variance in the period the invoices arrive, visible to the ledger and not to the plan.

What does connected mean for costing?

Costing is connected when every function reads the same costed record for a style — one that holds each cost figure with its basis, its currency, its date and its owner, rather than one field called cost — and when a cost event written by any one of them re-flows into the planned margin of the plan of record as a new version. In practice that means a cost change anywhere re-prices the plan, the margin is read at the colorway and size the cost actually varies at, and a landed cost that lands above target by more than an agreed threshold opens a review while the decision it affects can still be taken. A stack where every cost figure is accurate in its own system and the plan still carries the figure that was true when it was last touched is integrated, not connected.

Does one costed record replace PLM or the ERP?

No. PLM stays the system of record for the bill of material, the quoted cost and the costing versions behind it; the ERP stays the system of record for the purchase order cost, the standard cost and the landed-cost accruals; finance stays the owner of the actual. One costed record means the plan of record reads those figures with their bases, dates and owners attached and holds the planned landed cost and planned margin beside them, so that a change in any one of them becomes a new version of the planned margin rather than a number someone re-keys. The systems keep their jobs. What changes is that the figure the season’s margin is read on is the same figure in every room.

• ✓Apparel costing is the set of cost figures a style carries through its life — target, quoted or FOB, bill-of-materials, standard, landed and actual invoiced cost — and the margin each implies against the ticket price. They are six facts about one garment, not six estimates of one number.
• ✓Each has a system of record: PLM for the bill of material, the target as objective and the quote; the ERP for the purchase order cost, the standard and the landed-cost accruals; the plan of record for the planned landed cost and planned margin; finance for the actual; retail and ecommerce for the ticket price and the realised average unit retail.
• ✓The figures disagree for seven reasons, none of them an error: timing, currency, duty and the value it is assessed on, freight allocation, minimum-order surcharges and price breaks, sample and development cost treatment, and roll-ups at style, colorway and size.
• ✓Cost changes at five events — line-plan target, costed sample, purchase order cost locked and amended, actual landed cost, post-season true-up — and at each one a different team writes the number into a different system, and the plan learns of it only when someone re-keys it.
• ✓In the illustrative worked example, one style ticketed at 125.00 reads 70.4% initial markup on FOB, 63.0% on the planned landed cost of 46.25 and 61.6% on the actual landed cost of 48.00; the colorway below its dye minimum reads 62.2% while the style average reads 62.8%.
• ✓Connected means one costed record every function reads, cost events that re-flow into the planned margin as new versions, and a review that opens when landed cost exceeds target by an agreed threshold while a decision is still available. A stack where every figure is right in its own system and the plan still carries a stale one is integrated, not connected.
• ✓Agreeing one definition is a sequence: write the ladder down, one owner per figure, landing assumptions as data, the figure that consumes open-to-buy, the grain, the currency rule, the sample cost treatment, the return path for each event, and the review threshold.
• ✓Footwear adds tooling amortised across pairs, home and furniture makes container freight the cost base, jewelry rests on a metal price that moves daily, and outdoor and sporting goods fix the dealer price before the cost is final — each changes which cost event matters most, not whether the ladder applies.

• See which system owns cost and every other apparel data object — the system of record map →
• Follow one purchase order from open-to-buy approval to landed cost and close →
• Read what the plan of record holds and who holds the pen →
• Read why the margin is made before the buy →
• Read why ERP and Excel still leave apparel margins exposed →
• Read why the style-color-size hierarchy is the hard part of the data model →
• Read why workflows break at handoffs rather than inside functions →
• See which decision belongs at which level — the planning grain →
• Read what the ERP that holds the order cost and the standard is — the apparel and footwear solution →
• Read how the 810 invoice and the 820 remittance change the margin an account delivers — apparel EDI →
• See which clock is allowed to change each number — the operating cadence →
• Sequence a connected implementation one seam at a time →