The apparel purchase order lifecycle

Why the purchase order is the object to follow

Most objects in an apparel stack have a home. The style lives in PLM, the plan in the planning system or a workbook, the stock in the warehouse system, the sale in the commerce platform or the point of sale. Each has one team that cares about it most and one system built around it. The purchase order is different: it is touched by more functions than any other object, and no function owns it end to end. Planning approves the money behind it, buying shapes it, sourcing places it, production chases it, logistics moves it, the warehouse receives it and finance pays for it.

That is why the purchase order is the best single test of how connected a stack actually is. A brand can have an excellent PLM, a well-run ERP and a careful planning team and still have a purchase order whose quantity by size is different in three places, whose ship date is right only in a production coordinator’s inbox, and whose landed cost is known to finance and to nobody who planned the margin. None of those systems is failing. The order is simply passing through more hands than any one system was designed to follow.

It is also the object where the plan meets the commitment. Before the order is issued, every number about it is a plan — changeable, approvable, a scenario until it is locked. After it is issued, the core numbers are a commitment to a third party, and changing them is a negotiation with a factory rather than an edit to a spreadsheet. The lifecycle is the record of how a planned number becomes a committed one, and how the commitment then moves while the plan is still reading it.

Which system is authoritative for the purchase order, and for every other object it touches — the style, the bill of material, the cost, the inventory position — is set out object by object in the system of record map. This page follows the order through time rather than across objects: what it is at each stage, where it lives, and what goes wrong as it moves.

Open-to-buy approval

A purchase order’s life begins before the order exists, with the open-to-buy that authorises it. Open-to-buy is the running balance of what a buyer can still commit in a period: for each delivery month or week and each category, the planned receipts the merchandise plan calls for, less the units or value already on order for that period. The planned receipts come from the standard retail identity — planned sales, plus planned markdowns, plus planned end-of-period inventory, minus beginning-of-period inventory, calculated at retail and converted to cost where open-to-buy is held at cost — and the on-order comes from every purchase order already placed against the period.

A purchase order is open-to-buy turned into a commitment. When the order is approved, it consumes budget in its delivery period; when it is cut, it releases budget; when it slips into the next month, it moves budget with it. That is why the first stage of the lifecycle is a planning object rather than a transaction. The order that eventually sits in the ERP is only correct if the budget it was drawn from is still correct when it arrives, and the open-to-buy only stays correct if every later stage reports back to it.

The flagship’s guide to open-to-buy planning sets out the identity and how the balance is maintained. For the lifecycle, the point is narrower: the open-to-buy is the first reader of every event that follows, and the later sections of this page are, in effect, a list of the events it needs to hear about.

Where the PO lives

The planning system or the open-to-buy workbook, by month or delivery, by category or class, usually at cost and often at retail as well.

Who owns it

Merchandise planning, with the budget approved by the head of merchandising or finance at the seasonal sign-off.

What it should carry

The money available to commit by delivery period and category: planned receipts for each period, less what is already on order for it.

The typical break

The approved budget sits in a workbook no order is linked to, so it is consumed by memory during the buy and reconciled to the ERP at month end.

The buy plan: style-color-size quantities

The buy plan converts money into units. For each style-color, the buyer sets a depth — how many units in total — and planning applies a size curve and a delivery phasing to it. The result is a grid: style-color by size, by delivery, sometimes by channel where the direct site and wholesale are bought separately. It also fixes the decisions the factory will need: whether the order is bulk eaches, prepacks or both; the target or quoted cost; which factory or agent; and the date the goods are needed in the distribution centre.

The size split is a planning decision that the purchase order inherits. Once the order is issued, every size is a separate commitment, and correcting a size curve means negotiating with a factory that may already have cut the fabric. That makes the buy plan the last point at which the size split is cheap to change — and the point at which it is most often decided in a hurry, after the totals have been approved.

Which level each of these decisions belongs at — depth at style-color, the curve at size, phasing at delivery — is the subject of the planning grain. The buy plan is also the layer of the plan of record that becomes a purchase order: once the ERP creates the order, the quantity belongs to the commitment, and the plan reads it back rather than editing it.

Where the PO lives

The planning system or a buy sheet, built from the line plan and the assortment, with target cost carried from costing in PLM.

Who owns it

Buying or merchandising for the style-color choices and depth; planning for the size curve, delivery phasing and the open-to-buy check.

What it should carry

Units by style-color, by size, by delivery and by channel; the pack configuration; the target or quoted cost; the factory or agent; the required in-DC date.

The typical break

The buy total is checked against open-to-buy while the size split is decided later, in a separate file, by someone else.

PO creation and terms

The purchase order is created when the approved buy is turned into ERP order lines and issued to the factory or agent. From that point it is a commercial document with legal weight: it fixes what the brand has asked for, at what price, on what terms and by when. In most apparel stacks the ERP is the system of record for it, and the plan becomes a reader of it rather than a writer.

The terms decide who carries cost and risk between the factory and the warehouse. The common ones in apparel sourcing are Incoterms rules, published by the International Chamber of Commerce in their current Incoterms 2020 edition. Under FOB (Free On Board), the factory’s obligation ends when the goods are on board the vessel at the named port of shipment; the brand arranges and pays for the main carriage, insurance, import clearance and duty. Under DDP (Delivered Duty Paid), the factory delivers the goods to the named destination, cleared for import with duty paid. Between them sit terms such as FCA, which the ICC’s guidance points containerised goods toward, although FOB remains the everyday shorthand in much of apparel sourcing.

For the plan, the term matters because it decides what the unit price contains. An FOB price is a first cost: freight, duty, insurance, brokerage and inland delivery all have to be added before it becomes a landed cost the margin can be planned on. A DDP price already contains most of them. Two orders at the same unit price on different terms are not the same cost, and a buy plan that compares them on unit price alone is comparing different things. Incoterms also allocate cost and risk, not ownership: when title to goods in transit passes is a matter for the sales contract, and finance should state when in-transit stock enters the brand’s books.

The order also carries several dates, and they are not interchangeable. The ex-factory date is the date the factory commits to ship. The ship window — a start-ship date and a cancel date — is the period inside which the brand will accept the goods. The in-DC date is when the goods are expected in the distribution centre, and some brands carry an on-floor or launch date beyond it. The factory works to the ex-factory date; the receipt plan works to the in-DC date; the open-to-buy is consumed in the period of the in-DC date. When one moves and the others are not recalculated, the order is late in one system and on time in another.

The break at this stage is almost always transcription. The buy sheet and the ERP order lay out the same information differently — sizes across columns in one, down rows in the other; packs as a ratio in one, as a pack SKU in the other — and the order is keyed or uploaded across the gap. A transposed size or a default date entered here is carried, faithfully, through every stage that follows.

Where the PO lives

The ERP, which creates the order, numbers it and holds its lines; the issued order then leaves for the factory as a PDF, a portal record, a workbook or an 850.

Who owns it

Buying or sourcing — whichever team places orders with the factory — with the order approved against open-to-buy before it is issued.

What it should carry

Factory or agent; style-color-size quantities; pack instructions; unit price and currency; Incoterm and named place; payment terms; ship window; ex-factory and in-DC dates; ship mode; destination.

The typical break

The buy is re-keyed from the buy sheet into ERP lines: sizes transposed, prepacks entered as eaches or eaches as packs, and date fields left at the system default.

Factory confirmation

The factory’s confirmation is its acceptance of the order: the quantity it will make by size, the price it accepts, and the ex-factory date it commits to. It arrives in whatever form the factory works in — an 855 purchase order acknowledgment where the factory exchanges EDI, a status change on a vendor portal, a countersigned order, or an email with the order attached and a line of text that changes one size.

An unconfirmed purchase order is a request, not on-order. Until the factory has accepted it, the brand does not know whether the quantity, the price or the date will hold, and treating it as committed supply makes every later number optimistic. Brands differ in when they count an order as on-order — at issue or at confirmation — and either rule can work, but it has to be one rule, written down, so that the planning view and the ERP count the same orders.

The harder case is the confirmation that differs from the order. A factory may accept the order but confirm a smaller quantity on one size because the fabric lot is short, or confirm an ex-factory date two weeks later than asked because its line is full. A confirmation that differs from the order is an amendment in disguise, and it needs the same treatment as any other amendment: recorded, approved, written into the ERP, and carried into the plan’s on-order and open-to-buy. When it arrives as a sentence in an email, it is usually none of those.

The confirmation is also the first point at which the order’s date can be checked against the plan. A confirmed ex-factory date, plus transit and receiving time, gives a working in-DC date; if it falls after the plan’s receipt week, the order is already late before a single unit has been cut, and the time to decide what to do about it is now, not when the ship notice arrives.

Where the PO lives

The factory’s reply — an 855, a portal status, a signed order or an email — with any difference keyed into the ERP order by sourcing.

Who owns it

Sourcing or production, who chase the confirmation and decide whether a confirmed difference is acceptable.

What it should carry

The confirmed quantity by size, the confirmed price, the confirmed ex-factory date, and every exception to the order as issued.

The typical break

The factory confirms short on one size or later than asked, in a reply email; the ERP keeps the ordered quantity and date, and the plan never sees either difference.

Materials and WIP milestones

Between confirmation and shipment, the order becomes physical. Fabric and trims are booked against the bill of material held in PLM; lab dips, strike-offs and trim approvals are submitted and signed off; fit and pre-production samples are approved; fabric arrives at the factory; the order is cut, sewn, finished and packed; a final inspection is carried out, often by sampling against a stated acceptance quality limit under ISO 2859-1; and the goods leave the factory. Each of these is a milestone with a planned date back-scheduled from the ex-factory date, usually held in a time-and-action calendar.

The working delivery date lives here, not on the purchase order. The order carries the date it was raised against. The milestones carry the date the factory is actually working to, and they move first: a late lab dip approval pushes fabric in-house, which pushes cut, which pushes ex-factory. In a disconnected stack the milestone update reaches the production coordinator by email or a weekly WIP report, and it reaches the ERP order and the plan’s receipt week only if someone carries it there.

The bill of material is the second link to PLM. The tech pack and the BOM describe what the factory is making — the fabric, the trims, the labels, the construction. A change to the BOM after the order is placed, such as a substituted trim or a changed fabric weight, can change the cost and the date of an order that the ERP still shows as unchanged. The order and the product record have to be read together; neither is complete alone.

How the time-and-action calendar is built and read is in the flagship’s guide to the time-and-action calendar in apparel, and what a good WIP report between a factory and a brand contains is in its guide to WIP reporting between factories and brands. For the lifecycle, the rule is simple to state and hard to keep: a slipped milestone is a date change to the order, and it should move the order’s in-DC date and the plan’s receipt week in the same change.

Where the PO lives

PLM for the tech pack and the bill of material; a time-and-action tracker, vendor portal or WIP workbook for the milestones; the factory’s own production system.

Who owns it

Production or sourcing for the milestones; product development and tech design for the approvals the milestones wait on.

What it should carry

Milestone dates planned and actual — approvals, materials in-house, cut, sew, finish, pack, final inspection, ex-factory — and the working delivery date they imply.

The typical break

A milestone slips in the tracker and never re-dates the ERP order or the receipt plan, so the order looks on time until the ship notice says otherwise.

PO changes and amendments

An order rarely reaches the warehouse in the shape it was issued. Quantities are cut or chased as the season reads, size splits are rebalanced, dates slip with milestones, ship modes change from sea to air to recover time, prices move with material costs, and lines are cancelled when a style is dropped. Some changes start with the brand, some with the factory. All of them change a commitment that the plan has already counted.

Amendments are where the plan and the ERP part company, which is why they have their own section further down this page. The short version is that an amendment is agreed in one place, written into another, and expected to reach a third — and the third is usually the plan.

Where the PO lives

The ERP order, which should carry every amendment; in practice the amendment is agreed in email, a call or a vendor portal first.

Who owns it

Buying, as the single owner of every amendment — quantity, price, cancellation, date and ship mode — with sourcing and production supplying the working date and the factory’s proposed changes, and planning consulted on the effect on open-to-buy.

What it should carry

Each change as a new version: what changed, the prior value, the reason, who approved it and when, and its effect on on-order and open-to-buy.

The typical break

The order is overwritten in one system and left unchanged in the other, so the plan and the ERP each hold a quantity or a date the other does not.

ASN and packing list

When the goods leave origin, the shipment is described in a ship notice — the advance ship notice, or ASN — and a packing list. Together they say what is coming: which purchase order lines, which sizes, how many units, packed in which cartons, in which container or on which flight, leaving when. The commercial invoice that travels with them says what the factory is billing for. Where the factory or agent exchanges EDI, the ship notice is an 856; elsewhere it is a portal record, a spreadsheet or a PDF.

The ship notice is the first hard evidence of what will land. Everything before it is a promise; the ship notice is a record of what physically left. A size shipped short on the ship notice will be short at the dock weeks later, and the plan can act on it now — re-phasing the receipt plan, warning allocation, deciding whether the shortfall is worth chasing — rather than when the warehouse finds it. That is only true if the ship notice reaches the plan at size grain, which in many stacks it does not: it reaches the warehouse, which needs it to receive, and stops.

The ship notice is also the point at which shipped units should leave on-order and enter in transit. They are no longer an open commitment at the factory, and they are not yet stock. Treating them as either is a mistake: counting them as on-order hides that they have shipped, and counting them as received hides that they have not arrived. The available-to-sell rule on the sibling resource follows the same logic: a conservative definition counts inbound only once there is evidence it has left origin, which is exactly what the ship notice is.

Where the PO lives

The factory, agent or freight forwarder sends the ship notice — an 856, a portal record or a file — with the packing list and commercial invoice; the warehouse system loads the expected receipt from it.

Who owns it

Logistics, with sourcing answering for any difference between what was ordered and what was shipped.

What it should carry

Shipped quantity by PO line and by style-color-size; cartons and their contents; packs and their ratios; the container or booking reference; the departure date.

The typical break

The packing list is keyed carton by carton into the warehouse system, or the ship notice references the wrong PO line, so the expected receipt is wrong before the goods arrive.

In transit

In transit is the stage the brand controls least and sees least. The goods are on a vessel, a flight or a truck; their position and estimated arrival are held by the forwarder or the carrier; their clearance is handled by a customs broker; their delivery to the distribution centre is booked against a dock slot. The brand’s own systems usually hold the shipment as a single status — shipped — and a date that was set before the goods left.

A vessel delay is a date change to every purchase order on board, and the plan needs it as one. A container that rolls to a later sailing, or waits at port for clearance, moves the in-DC date of every order line inside it. The receipt week moves, the open-to-buy period may move, and the available-to-sell for any promise depending on those units moves with them. Where the update lives only in the forwarder’s portal, the plan keeps expecting the goods in the original week until the receipt does not happen.

Transit is also where the cost of the order starts to change. Freight is charged per shipment, not per unit, and has to be allocated back to the order lines it carried; duty is assessed on entry. Both are inputs to landed cost, and both arrive as separate invoices, weeks apart, from parties the plan never deals with.

Where the PO lives

The freight forwarder’s or carrier’s systems for bookings, vessel and flight status, arrival and customs clearance — often outside every system the brand runs.

Who owns it

Logistics, working with the forwarder and the customs broker.

What it should carry

Departure and estimated arrival at port, customs clearance status, inland delivery booking, and the revised in-DC date they imply.

The typical break

The arrival estimate moves in the forwarder’s portal while the ERP keeps the original in-DC date, so units in transit are neither on hand nor visibly late.

Receipt and inspection

Receipt is where the order meets the building. The warehouse counts what arrived against the expected receipt loaded from the ship notice, carton by carton or by scan, and records what it found: the quantity by size, any cartons short or over, any wrong sizes or colors, any damage. Quality may hold some or all of the receipt for inspection at the distribution centre, even where a final inspection was carried out at origin. Only units received, counted and released from any hold are stock the business can sell.

On-order should be relieved by what was received, not by what was shipped or ordered. The receipt is the moment in-transit units become on hand, by size, and the plan’s receipt flow, the ERP’s open order and the warehouse’s stock should all change in the same event. Where the ERP relieves the order line by its full quantity on the first receipt, or the plan relieves on-order from the ship notice, the brand loses sight of exactly the units that are missing.

Discrepancies need a home. A short carton, a size received that was not on the ship notice, a quantity over the order’s tolerance — each has to be recorded against the order, assigned to someone to resolve with the factory, and held somewhere visible while it is resolved. A disputed unit is neither on order nor on hand, and a planning view that cannot show it as its own line will quietly count it as one or the other.

Where the PO lives

The warehouse system, which receives against the expected receipt and posts to the ERP; quality records for any inspection at the distribution centre.

Who owns it

The warehouse for the count; quality for inspection and grading; sourcing for resolving discrepancies with the factory.

What it should carry

Received quantity by PO line and by style-color-size, discrepancies against the ship notice and the order (shorts, overs, wrong sizes, damage), and the inspection result.

The typical break

The receipt is posted at the line total rather than by size, or a discrepancy is settled in email, so the plan, the ERP and accounts payable each hold a different received quantity.

Landed cost and three-way match

Before the factory is paid, accounts payable runs a three-way match: the purchase order, the receiving record and the factory’s invoice are compared, line by line. The price billed should agree with the price on the order, and the quantity billed should agree with the quantity received. A difference holds the invoice until it is resolved — a credit for a short shipment, a corrected price, an approved over-shipment. The match is a finance control, but it is also the last place a quantity discrepancy left unresolved at receipt will surface.

Landed cost is assembled alongside it. The unit price on the order is the first cost; freight, insurance, duty, brokerage and inland delivery are added from their own invoices and allocated to the order lines they belong to. The result is the actual landed cost per unit — the number the margin is really made on. At buy approval the plan used an estimate of it; at this stage the actual replaces the estimate, and the difference is the cost variance on the order. The flagship’s landed cost glossary entry sets out the components and why planning decisions run on landed cost rather than on first cost.

The break is that the actual stays in finance. The season’s initial markup was planned on the estimate, the open-to-buy was consumed at the estimate, and the in-season margin read is built on the estimate — while the actual is known, a few weeks after receipt, to a team that does not own any of those numbers. A landed cost that never returns to the plan leaves the plan’s margin permanently one estimate away from the truth.

Where the PO lives

The ERP and accounts payable for the match; the costing record for landed cost; separate invoices from the factory, the forwarder and the customs broker.

Who owns it

Finance, with sourcing answering for price differences and the warehouse for quantity differences.

What it should carry

The purchase order, the receipt and the invoice matched on price and quantity; actual freight, duty and other costs allocated to the order lines; the actual landed cost per unit.

The typical break

The actual landed cost is calculated in finance and never reaches the plan, so the season’s margin is planned on the estimate and reported on the actual.

Close

An order is closed when every line is either fully received or cancelled, and every invoice against it has been matched and settled. Most orders close cleanly. The ones that do not are the orders with a small unshipped balance — a few units of one size the factory never made, a line partly received with the remainder waiting on a decision — and those are the ones that cause trouble.

An open balance that nobody will ship is phantom on-order. It consumes open-to-buy in a period that has passed or is passing, it shows as inbound to anyone computing available to sell from open orders, and it sits on every open-order report until someone closes it by hand. The close rule — what happens to an unshipped balance, by when, and who decides — is a small decision with a long tail, and it is one of the steps in the ownership agreement at the end of this page.

Where the PO lives

The ERP, which closes the order or its lines once every unit is received or cancelled and every invoice is settled.

Who owns it

Buying for the decision to cancel an unshipped balance; finance for the financial close.

What it should carry

A closed status; any unshipped balance cancelled with a reason; the open-to-buy released by the cancellation; the final actual landed cost.

The typical break

Small open balances are never closed, and sit as on-order for months — consuming open-to-buy and showing as inbound that will never arrive.

Stage, system of record, owner, and the typical break

The eleven stages in one view. The system-of-record column names where the order should be held at that stage in a typical apparel stack, not where it always is; the last column is the diagnostic. If a brand recognises the break in a row, the order is not reaching that stage, or leaving it, on agreed terms.

Seven places the order is typed again

A purchase order that is held in one system at each stage still has to cross from one system to the next, and in many apparel stacks it crosses by being typed again. Each crossing is a translation — from one layout, grain and set of codes to another — and each is a delay between the moment something changes and the moment the next system knows. Re-keying is not an error in itself; it is a place where an error, a delay or a missing owner can enter the order, and the seven below are where it most often happens.

1. 01
   The buy sheet into the ERP

   The approved buy — units by style-color, size and delivery, with pack configuration and target cost — is typed or uploaded from the planning workbook into ERP purchase order lines. Size order, pack versus bulk and the date fields are the usual casualties, because the buy sheet and the ERP rarely lay them out the same way.

2. 02
   The ERP order out to the factory

   The issued order leaves the ERP as a PDF, a portal record, a shared workbook or an 850, and the factory or agent keys it into its own production system. Every factory now holds a copy of the order in its own shape, and every later change has to be made in both.

3. 03
   The factory confirmation back into the ERP

   The confirmation — accepted quantity by size, price, ex-factory date — comes back as an 855, a portal status or a reply email, and someone in sourcing keys any difference into the ERP. A confirmation that differs from the order on one size is the easiest thing in the lifecycle to miss.

4. 04
   Milestone dates into a tracker, and the tracker into the PO

   Time-and-action milestones are reported in a WIP tracker, a portal or a weekly workbook. When a milestone slips, the new ex-factory date has to be carried from the tracker into the ERP order and from there into the plan’s receipt week. It is often carried into neither.

5. 05
   The packing list into the warehouse

   The packing list and the ship notice describe cartons — which PO lines, which sizes, which packs, in which carton. Where they do not arrive as an 856 or a structured file, the warehouse or the logistics team keys the expected receipt by hand, carton by carton.

6. 06
   The receipt into the ERP and the plan

   The warehouse posts the receipt; the ERP relieves the open order; the plan’s on-order and receipt flow should move in the same instant. In many stacks the plan finds out from a weekly receipts report, at style-color rather than size, and its on-order is wrong by everything received since.

7. 07
   Freight, duty and invoices into landed cost

   The factory invoice, the freight forwarder’s invoice and the customs entry arrive separately and are keyed into accounts payable and the costing record. The actual landed cost they produce is the number the plan’s margin was built on an estimate of, and it rarely travels back to the plan at all.

The three mechanisms behind a failure at any of these crossings — translation, latency and ownership — are set out in the handoff problem, which describes the same pattern across the whole workflow. The purchase order is the object that makes the pattern easiest to see, because it crosses more seams than anything else and carries the same numbers across all of them.

Five ways size and prepacks complicate the order

A purchase order for fasteners has one line per item. An apparel order has one line per size of every color of every style, often split between packs and loose units, often on a size scale the factory labels differently from the brand. Each of the five complications below is a place where an order that looks right at style-color is wrong at the grain the business actually receives and sells at.

The order line is a size, not a style

Every sellable size of every color is its own commitment on the order, with its own confirmed quantity, its own shipped quantity and its own received quantity. A factory that ships the right total in the wrong size split has not filled the order, even though the style-color line reconciles. An order tracked at style-color can be complete in total and short in the size that sells, and nothing in a style-color report will say so.

This matters most at the stages where quantities change without anyone deciding to change them: a confirmation that shorts one size, a shipment that leaves a size behind, a receipt that finds a carton of the wrong size. Each produces a size-level difference that a style-color view nets away. For bottoms sized by waist and inseam, the grain is the cell of the grid, and an order flattened to waist alone cannot tell the plan whether the 32 by 30 is coming.

Prepacks: the pack is the unit the factory ships

A prepack is a sealed pack holding a fixed ratio of sizes — say one extra-small, two small, four medium, three large and two extra-large — that is ordered, packed, shipped, received and allocated as one unit. Wholesale accounts often order in packs because they receive and allocate in packs. For the order, a pack changes three things.

First, the factory cannot short one size inside a pack without breaking the ratio. If the medium fabric is short, the factory either makes fewer packs or asks to change the ratio, and either is an amendment to every size in the pack. Second, the order has to be held at two grains at once: packs for the factory and the warehouse, which handle them as units, and eaches for the plan, which reads size. The pack ratio has to be held once and referenced everywhere, or the plan’s exploded size quantities and the ERP’s pack quantities drift apart the first time a ratio changes. Third, packs fix how the received stock can be used: units inside a sealed pack are not available to a channel that sells single units until the pack is broken.

Bulk and packs on one order

Many orders carry both: prepacks for wholesale accounts and bulk eaches for the direct site, stores or replenishment. The two halves of the order behave differently. A shortfall is absorbed by bulk first, because bulk can be shorted size by size without disturbing a ratio, so the direct channel tends to carry the cut while the wholesale packs ship complete. A partial shipment usually carries whole packs and some of the bulk, leaving bulk balances open. The worked example later on this page has exactly this shape.

For the plan, the consequence is that the order’s size totals are not one number but two, with different destinations. A medium cut on bulk is a direct-site in-stock problem; the same cut spread across packs would have been a wholesale fill problem. The order has to be read by size and by half, or the plan cannot tell which channel is carrying the shortfall.

Size scales that do not match

The brand’s size scale is defined in PLM, the ERP holds sizes as codes on the item, the factory may label sizes in its own convention, and a retail partner may expect a different size description again. Where the mapping between them is rebuilt by each team — a lookup in the buy sheet, another in the factory workbook, a third in the warehouse — a size can be ordered as one thing and received as another. The mapping belongs to the product record, held once, as the apparel data model describes, and every system that touches the order should read it rather than keep its own.

Over- and under-shipment by size

Factories rarely ship exactly the ordered quantity. Cutting yields vary, seconds are pulled at inspection, and many orders carry a tolerance — stated in the terms — inside which a short or over shipment is accepted as complete. The tolerance is usually stated for the order or the line, and applied to the total. Applied at style-color, it can accept a shipment that is within tolerance overall and well outside it on a single size. A tolerance that is checked only on the total is a tolerance on the wrong number; the brand should decide whether it applies by size as well, and the receipt should be checked against it at the grain the order was placed at.

PO amendments, and why they are where the plan and the ERP diverge

The plan and the ERP usually agree on the day an order is issued, because the order was keyed from the plan. They disagree later, and the cause is almost always an amendment. An order is amended many times between issue and shipment, by different people, for different reasons, through different channels — and each amendment has to reach the ERP, the plan, and often the factory’s own system. Every amendment that reaches only some of them is a permanent difference between the plan’s on-order and the ERP’s open order, until someone finds and reconciles it.

Definition — Purchase order amendment

A purchase order amendment is any change to a placed order’s commitment — quantity by size, size split, price, ship date or ship window, ship mode, terms or destination, or the cancellation of a line — agreed between the brand and the factory or agent after the order was issued. It is a new version of a commitment rather than an edit to a plan, so it should be agreed in one place, recorded with its reason and the prior value, written into the system of record for the order, and carried into the plan’s on-order and open-to-buy in the same change. An amendment that reaches only one of them is where the plan and the ERP begin to disagree.

Used by: Buyers, sourcing and production teams, merchandise planners and ERP owners at apparel brands

Related: Purchase order, EDI 860, factory confirmation, on order, open-to-buy, working delivery date, plan of record, version history

The kinds of amendment, and what each moves

Amendments come in a small number of kinds, and each moves something different in the plan. A quantity cut releases open-to-buy and reduces the receipt; a chase consumes open-to-buy and raises it; a size-split change leaves the total alone and moves the curve; a date slip moves the receipt, and possibly the open-to-buy, into a later period; a ship-mode change usually moves the date and the landed cost together; a price change moves the cost and the margin; a cancellation releases everything the line held. The table sets them side by side.

Four reasons amendments split the plan from the ERP

1. 01
   Agreed outside both systems

   Most amendments are agreed in an email thread, a call or a factory portal before they are written anywhere a planner can see. The agreement is real; the record is a message in one person’s inbox. Until someone writes it into the ERP and the plan, both are wrong, and they are rarely updated by the same person on the same day.

2. 02
   Written into one system

   A buyer who cuts a quantity updates the ERP because the factory needs a revised order. The planner, not copied on the email, still counts the original quantity against open-to-buy. Or the reverse: a planner edits the receipt in the plan when a slip is reported in the WIP meeting, and the ERP order keeps its original date because nobody in sourcing was told to change it.

3. 03
   Overwritten, not versioned

   When an amendment does reach the ERP, it usually replaces the order’s quantity or date with the new value. The original lives in a change log, if anywhere. The plan, which needs to know both what was ordered and what is ordered now — to measure how far the season has moved from the approved buy, and why — gets only the current state. An order that has been cut three times looks, in the ERP, like an order that was always that size.

4. 04
   No reason recorded

   A quantity that changed without a reason cannot be read. A cut because the fabric mill was short says something about the factory; a cut because the style is selling slowly says something about demand; a cut because open-to-buy ran out says something about the plan. Without a reason code, every amendment is just a smaller number, and the season-end review cannot learn anything from it.

The remedy is not to slow amendments down. It is one amendment path: one place where every change to a placed order is agreed and recorded, with the change, the prior value, the reason and the approver; written into the ERP order, which remains the single writer of the commitment; and carried into the plan’s on-order and open-to-buy in the same change. Who may approve which kind of amendment is a decision right — the decision rights map lists PO placement and PO amendment as two of its sixteen decisions, both owned by buying — and why the plan has to hold amendments as versions rather than edits is set out in the plan of record.

How open-to-buy consumption and on-order tie back to the plan

The plan does not need to hold the purchase order. It needs to hold what the order means for the season: how much budget it has consumed, in which period; how many units it will deliver, by size, in which week; and how far it has moved from the approved buy. On-order is the plan’s reading of every open purchase order, by size and by delivery period, and open-to-buy is the plan’s budget less that reading.

On-order is merchandise committed to a supplier and not yet received. For the lifecycle, the useful definition splits it by state: units still at the factory, units in transit after the ship notice, and units in dispute after receipt — with received units leaving on-order altogether. The flagship’s on order glossary entry sets out how on-order enters open-to-buy and why cancel dates matter to it. What the lifecycle adds is the list of events that should move it, and the rule that each event moves the plan and the ERP together.

Two choices in that table are the brand’s to make, and both have to be made once. The first is whether on-order counts from issue or from confirmation. The second is the value at which open-to-buy is consumed — usually the landed-cost estimate used at buy approval, sometimes the first cost, sometimes retail value as well — and when the actual replaces it. Either choice can work; two systems using two different ones will disagree by exactly the orders the choice affects.

The same events feed the in-season view of the business. Units in transit with a ship notice are the inbound that a conservative available-to-sell definition counts for promises falling after their arrival; units still at the factory are not. How that rule is set, and why the working delivery date decides whether inbound availability is honest, is in available to sell in apparel. Where on-order sits among the layers every function reads — line plan, assortment, buy, purchase orders, WIP and in-season position — is in the plan of record, whose boundary rule applies here directly: the ERP creates and owns the order, and the plan reads it so that open-to-buy nets what is on order.

The failure the rule prevents is familiar. A planner opens the open-to-buy report on Monday and sees budget in a month that buying knows is fully committed, because three amendments last week reached the ERP and not the plan. Or the reverse: the plan shows a month overspent because a cancelled line is still counted. Either way, the next buy decision is made against a number that is wrong by the amendments it did not hear about.

Where EDI touches the order: brand as buyer, brand as seller

A wholesale apparel brand sits in two purchase order lifecycles at once. Upstream, on its own buy, the brand is the buyer: it issues purchase orders to factories and agents, and receives their confirmations, ship notices and invoices. Downstream, on its wholesale business, the brand is the seller: it receives purchase orders from retail partners, and sends them confirmations, ship notices and invoices. The X12 documents are the same in both directions; the brand’s role, the team that answers for each document and the decision it should feed are not.

On the buyer side, the documents are only as standard as the factory. Many factories and agents do not exchange X12 or EDIFACT at all and work through a vendor portal, a shared workbook or email, so the same five facts — the order, the acceptance, the change, the shipment and the bill — arrive in several shapes across a brand’s supplier base. The lifecycle above is written so that it holds whichever shape they arrive in: what matters is that each fact reaches the system that holds the order at that stage, at size grain.

The two lifecycles meet in the plan. The retailer’s 850s and 860s are the wholesale order book that the brand’s own buy has to cover; the factory’s 855s and 856s are the supply that covers it. A retailer’s cancel date that falls before a factory’s working ex-factory date plus transit is a booked order the brand is about to miss. What each transaction set carries, the planning decision each should feed, and why EDI data so often stops at the ERP instead of reaching the plan is set out in full in apparel EDI in the connected stack; this page does not restate it.

One order, five sizes, packs and bulk, a partial shipment

The numbers below are illustrative, chosen to show the arithmetic rather than to describe any brand or factory. A brand orders one style-color — a woven shirt in one color — from one factory on FOB terms, for a single delivery planned into its distribution centre in week 32. The order has two halves: 50 prepacks for wholesale accounts, each holding 12 units in a 1-2-4-3-2 ratio from extra-small to extra-large, and bulk eaches for the direct site. The plan’s landed-cost estimate is 17.60 a unit: the FOB price of 14.00, plus illustrative allowances of 1.10 for freight, 2.10 for duty and 0.40 for insurance, brokerage and inland delivery.

The order as issued

The pack column is 50 times the ratio: 50, 100, 200, 150 and 100, which is 600 units. The bulk column is 40 + 90 + 140 + 110 + 70 = 450. The order is 600 + 450 = 1,050 units, and at the 17.60 estimate it consumes 18,480 of open-to-buy in week 32 (1,050 × 17.60).

Confirmation, a slip, a partial shipment and a receipt

Confirmation The factory confirms the order with one exception: the fabric lot for the bulk mediums is short, and it confirms 120 bulk mediums rather than 140. It cannot short medium inside the packs without breaking the 1-2-4-3-2 ratio, so the cut falls on bulk. The confirmed order is 1,030 units, and the 20-unit cut releases 352 of open-to-buy (20 × 17.60).

A slip Weeks later, the WIP report says 60 of the 120 bulk mediums will miss the vessel and follow on the next sailing, landing in week 35. On-order by delivery week is now 970 in week 32 and 60 in week 35 — still 1,030 in total — and 1,056 of on-order (60 × 17.60) has moved from week 32 to week 35.

Shipment one The ship notice for the first shipment lists all 50 packs (600 units) and bulk of 40 extra-small, 90 small, 60 medium, 110 large and 70 extra-large (370 units): 970 units in all. The factory invoices 970 units at 14.00, which is 13,580.

Receipt The warehouse receives all 50 packs intact. One bulk carton of 10 large is missing from the container, so bulk large is received at 100 rather than 110. The receipt is 960 units: 600 in packs and 360 loose (40 + 90 + 60 + 100 + 70).

Each column sums. Confirmed: 90 + 190 + 320 + 260 + 170 = 1,030. Shipped: 90 + 190 + 260 + 260 + 170 = 970. Received: 90 + 190 + 260 + 250 + 170 = 960. Open is confirmed less shipped, and only medium has a balance: 320 − 260 = 60. Large is not open: the factory says it shipped 260 and the warehouse counted 250, so the 10 units are a discrepancy to resolve, not an order to wait for.

Where the 1,050 approved units are now

The units reconcile: 960 + 60 + 10 + 20 = 1,050. So do the values: 16,896 + 1,056 + 176 + 352 = 18,480, the open-to-buy the order consumed when it was approved. Against the plan’s week 32 receipt of 1,050 units, the week delivered 960, and the 90-unit variance is three different things — 20 cut, 60 late and 10 in dispute — each with a different owner and a different next step.

The three-way match and the landed cost

The price agrees on all three documents. The order quantity of 1,030 is not a mismatch with the invoice: the first shipment is a partial, and 60 units remain open. The mismatch is between the invoice and the receipt — 970 billed, 960 received — so the invoice and the receipt differ by 140 (10 × 14.00), and the 10 units are held in dispute until the factory, the forwarder or the warehouse accounts for the missing carton.

When the forwarder’s invoice arrives, the freight allocated to this order comes to 1.30 a unit rather than the 1.10 estimate. With the other components unchanged, the actual landed cost is 17.80 (14.00 + 1.30 + 2.10 + 0.40), and the 960 received units carry 192 more cost than the plan consumed for them (960 × 0.20). That difference is a margin fact about this order. It reaches the plan only if someone carries it there.

The same order, as each system sees it

Suppose the amendments did not reach every system. The buyer entered the 20-unit cut in the ERP because the factory needed a revised order, but the planning workbook was not told; the 60-unit slip was reported in the WIP meeting and never entered in the ERP or the workbook; and the warehouse knows only the ship notice it received against.

The workbook shows 1,050 − 960 = 90 still open in week 32, because it heard about neither the cut nor the slip; it also still counts the 352 of open-to-buy the cut released. The ERP shows 1,030 − 960 = 70 open in week 32 — the 60 mediums at their original date plus the 10 large it cannot yet tell apart from an open balance. The warehouse expects 970 − 960 = 10 more units and knows nothing of the mediums. Four views of one order, four open quantities — 60, 90, 70 and 10 — and each is computed correctly from what its system was told.

What the size read says that the total does not

At style-color, 960 of 1,050 units arrived on time, and the delivery looks nearly complete. At size and by half, it does not. Bulk medium — the direct site’s half of the size that sells most — was ordered at 140 and received at 60: 20 were cut and 60 are three weeks late. The wholesale packs arrived complete, and cannot be broken for the site without a decision to do so. So the direct site starts week 32 with 60 bulk mediums against a plan for 140, and under a conservative available-to-sell rule the 60 still to come count toward its promises only once they have a ship notice. That is a reallocation question this week, a chase question if the medium is selling, and a buy question for the next order — none of which is visible in a report that says 960 of 1,050 arrived.

What the planning side needs back from the PO stream

The purchase order stream is usually reported for the teams that run it: an open-order report for buying, a WIP report for production, an inbound schedule for logistics, an unmatched-invoice list for finance. The planning side needs something different from the same data. Planning needs the order stream at size grain, by delivery week, with the reason for every movement visible, because the decision depends on why the supply moved — a cut, a slip, a short receipt — and not only on by how much.

On-order by delivery week

The basic read is on-order by style-color-size, by the week it is now expected in the distribution centre, split into at the factory, in transit and in dispute. It is what open-to-buy is netted against, what the receipt plan is checked against, and what available to sell reads for future promises. It has to be built from the working in-DC date — the confirmed ex-factory date as it moves, plus transit and receiving — and not from the date the order was raised against, or it describes the season the brand planned rather than the one that is arriving.

Late-PO exposure

Late-PO exposure is every open order line whose working in-DC date falls after its planned receipt week, valued and by size, with the weeks of delay beside it. In the worked example it is 60 bulk mediums, three weeks late, worth 1,056 at the plan’s cost estimate. The more useful version adds what the late units were meant to cover: a launch date, a promotion, a wholesale cancel date. A late order is a planning problem only in proportion to what was waiting for it, and the exposure report is where the planner sees which late orders matter.

For wholesale, the critical comparison is the factory’s working ex-factory date, plus transit, against the cancel date of the retailer orders the goods are meant to fill. An order that will land after a retailer’s cancel date is booked demand the brand is about to lose, and the time to act — expedite, change ship mode, negotiate the date — is while the goods are still at the factory.

Receipt flow versus plan

Receipt flow versus plan compares what actually arrived each week, by size, with what the plan expected, and explains the variance. In the worked example, week 32 planned 1,050 units and received 960, and the 90-unit variance is 20 cut, 60 late and 10 in dispute. A receipt variance with no breakdown is a number to worry about; one broken down by cause is three decisions with three owners. Receipts also drive the in-season position: weeks of supply, sell-through and the markdown read all assume the stock arrived when the plan said it would.

Size-level shortfall by channel

Where an order carries packs and bulk for different channels, the shortfall has to be read by size and by channel. A medium cut on bulk is a direct-site in-stock problem; the same cut on packs would have been a wholesale fill problem. The plan needs to see which half of the order carried each difference, so the reallocation, the decision whether to break packs, and the conversation with the wholesale accounts start from the right facts.

Cost variance

Finally, the plan needs the actual landed cost against the estimate it planned on, by order and by style-color, as soon as the actual is known. A freight or duty difference across a season’s orders changes the margin the season is delivering, and a planner reading the in-season margin on estimates is reading a number finance already knows to be wrong.

Where these reads sit on the weekly and monthly clocks — and which of them the weekly trade meeting should open on — is set out in the operating cadence.

Agreeing PO data ownership across teams, step by step

The work of agreeing who owns the purchase order’s data is mostly not technical. Every team that touches the order already believes it owns the part it touches, and in a sense it does. What is missing is a written agreement about which team may change which field after the order is issued, which events move on-order, and where the order’s exceptions are reconciled. The output is one writer per field, one amendment path, and an on-order figure the plan and the ERP both compute from the same events. The nine steps below are a sequence, not a menu: each depends on the ones before it.

1. 01
   Map every state the order passes through

   Write down the eleven stages as they actually run in the brand — not as the ERP manual describes them — and for each, name the system that holds the order and the person who answers for it. The map usually shows two or three stages where the order is held in email or a personal workbook, and those are where the work starts.

2. 02
   Name one writer for every field

   Quantity by size, unit price, Incoterm, each date, ship mode, pack configuration, status. Each field gets exactly one team that may change it after the order is issued, and every other team reads it. Buying usually writes quantity and price; sourcing or production writes the working ex-factory date; logistics writes the shipped and in-transit status; the warehouse writes the received quantity; finance writes the actual landed cost.

3. 03
   Agree the date set, and which date the plan reads

   Ex-factory date, ship window start and cancel date, in-DC date, on-floor or launch date. Agree which ones every order must carry, which one the factory commits to, which one the receipt plan reads, and who updates the working date when a milestone slips. The plan should read the working in-DC date, not the date the order was raised against.

4. 04
   Set one amendment path

   One place where every change to a placed order is agreed and recorded: the change, the reason, the prior value, who approved it and when. The amendment is written into the ERP order and carried into the plan’s on-order and open-to-buy in the same change, so that no amendment reaches one of them and not the other.

5. 05
   Agree the grain

   The plan reads the order at style-color-size by delivery. Packs are exploded to eaches for planning and kept as packs for the factory and the warehouse, with the pack ratio held once and referenced everywhere. Where the factory’s size labels differ from the brand’s, the mapping is held once, not rebuilt by each team.

6. 06
   Agree when on-order moves

   When an approved buy becomes on-order (at issue, or only at confirmation), when on-order becomes in transit (at the ship notice), and when in transit becomes on hand (at receipt, by size). Write the rule down. Open-to-buy is consumed and released on the same events, so the planning view and the ERP move together.

7. 07
   Agree the discrepancy rules

   What counts as a short, an over or a wrong-size receipt; the tolerance the terms allow; who resolves each kind; and where disputed units sit while they are resolved — neither on order nor on hand, as their own visible line. A discrepancy settled in email leaves the plan, the ERP and accounts payable each believing a different quantity.

8. 08
   Agree how actual landed cost returns to the plan

   The estimate used at buy approval, the cost components it includes, the point at which the actual replaces it, and who carries the difference back into the plan’s margin and open-to-buy. Without this step, the season is planned on estimates and reviewed on actuals, and nobody can say where the difference arose.

9. 09
   Define close, and review exceptions on a clock

   When a PO is closed, what happens to an unshipped balance — cancelled and released to open-to-buy, or kept open against a stated date — and who decides. Then review the exceptions on a stated clock: late confirmations, slipped dates, unresolved discrepancies, open balances past their cancel date. The exception list is the evidence that the ownership agreement is being followed.

The agreement itself fits on one page. The table below is the set of decisions it has to record, with the failure each one prevents. A brand that can fill in every row has an agreed purchase order; one that cannot has a set of habits that each team follows differently.

Who should own the agreement is an operating-model question before it is a systems one. It is usually best held by whoever owns the buy — a head of buying or merchandising operations — with planning, sourcing, logistics and finance each signing off the fields they write. The owner of the agreement is the person who settles a dispute about which quantity on the order is true, and in many brands that dispute currently has no owner, which is why it is settled by whichever report is opened first.

The sequencing logic for the systems work — one seam at a time, one owner per phase, starting where the cost is most visible — is in the connected planning implementation playbook. For the purchase order, the seam to start with is usually the amendment path, because it is where most of the disagreement between the plan and the ERP comes from.

The same lifecycle in other merchandise verticals

The eleven stages hold in every merchandise vertical that buys finished goods from factories; the grain of the order and the stage where it breaks do not. This follows the logic of the connected stack by vertical: each vertical has an attribute its orders are placed against, and an order tracked one level above it hides the shortfall. In apparel that attribute is size within style-color, split by pack and bulk. In the verticals below it is something else.

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

Follow ten orders through one delivery

The tempting first project is a new purchase order tracker. The useful one is smaller. Pick one delivery — ten or so orders, across two or three factories, with at least one carrying packs and bulk — and follow each order from approval to close, at size, recording what each system said about it every week: the plan’s on-order, the ERP’s open order, the WIP tracker’s ex-factory date, the ship notice, the receipt, the invoice. Write down every week a system disagreed with another, and which stage caused it.

That log is the business case, and it is specific to the brand in a way no outside benchmark could be. It also tells the brand which step of the ownership agreement to do first. If most of the disagreement starts at confirmation and amendments, the amendment path comes first; if it starts in WIP and transit, the date set and the working-date owner; if it starts at receipt, the discrepancy rules and the on-order relief.

Then write the agreement, name its owner, and move one stage at a time onto it — starting with the stage whose disagreements cost most. The structure the order depends on — one product record resolving style, color, size and pack, one calendar, one set of channels and locations — is described in the apparel data model.

What an agreed purchase order does not fix

An agreed purchase order — one writer per field, one amendment path, one on-order — makes every team read the same order. It does not make the factory ship on time. If the fabric mill is late, the order is late however well it is tracked; if a vessel rolls, the receipt moves; if a carton goes missing, it is missing in every system at once. What changes is when the brand finds out, and whether everyone finds out together.

Nor does it make the commercial decisions the order raises. Whether to air-freight the late mediums, whether to break wholesale packs for the direct site, whether to cut a retailer’s order or negotiate its cancel date — those are choices about cost, service and which customer to protect, and no record of the order makes them for the brand. What one agreed order does is put every such choice in front of the people who own it while there is still time to choose. That is enough to be worth doing.

Frequently asked questions

What is the apparel purchase order lifecycle?

The apparel purchase order lifecycle is the sequence of states a purchase order passes through between the open-to-buy approval that authorises it and the financial close that settles it: the buy plan by style-color-size, PO creation and terms, factory confirmation, materials and work-in-progress milestones, amendments, the ship notice and packing list, transit, receipt and inspection, landed cost and three-way match, and close. A different system and a different team hold the order at almost every stage, which is why it is re-keyed at the seams.

Which system is the system of record for an apparel purchase order?

In most apparel stacks the ERP is the system of record for the purchase order itself: it creates the order, holds its lines, receives against it and matches the invoice to it. But the order is shaped and changed elsewhere. The quantity by size comes from the buy plan, the materials from the bill of material in PLM, the working ex-factory date from the factory’s milestone reporting, the shipped quantity from the ship notice and the received quantity from the warehouse system. The ERP should be the one writer of the order, with every other system either feeding it through a defined path or reading from it.

What is the difference between FOB and DDP on an apparel purchase order?

FOB and DDP are Incoterms rules, published by the International Chamber of Commerce, that fix where the factory’s responsibility for cost and risk ends. Under FOB, the factory delivers the goods on board the vessel at the named port of shipment, and the brand arranges and pays for the main carriage, insurance, import clearance and duty. Under DDP, the factory delivers the goods to the named destination cleared for import with duty paid. For planning, the difference is what the unit price includes: an FOB price needs freight, duty and other costs added to reach landed cost, while a DDP price already contains most of them.

Why do purchase order amendments make the plan and the ERP disagree?

Because an amendment is usually agreed outside both systems — in an email, a call or a factory portal — and then written into one of them. A quantity cut entered in the ERP but not the plan leaves open-to-buy consumed by units that will never arrive; a ship-date slip recorded in a WIP tracker but not the ERP leaves the receipt plan expecting goods in the wrong week. Most systems also overwrite the order with its latest state, so the prior value and the reason are lost. One amendment path, written to the ERP and carried into the plan in the same change, is what keeps them aligned.

How does on-order tie back to open-to-buy?

Open-to-buy for a period is the planned receipts for that period minus what is already on order for it. Every purchase order therefore consumes open-to-buy in its delivery period, and every change to it moves open-to-buy: a cut releases budget, a chase consumes it, and a date slip moves it from one period to the next. In the illustrative example on this page, a 20-unit cut at confirmation releases 352 of open-to-buy at the plan’s landed-cost estimate of 17.60, and a 60-unit slip moves 1,056 of on-order from week 32 to week 35.

Should on-order be relieved when goods ship or when they are received?

Shipped units should move from on-order to in transit at the ship notice, and from in transit to on hand only at receipt, by size. Relieving on-order straight from the ship notice treats every shipped unit as received. In the illustrative example on this page, 970 units shipped and 960 were received: the 10-unit difference is neither on order nor on hand, and is held as its own disputed line until it is resolved.

What is a three-way match on an apparel purchase order?

A three-way match is the accounts-payable control that compares the purchase order, the receiving record and the supplier’s invoice before the invoice is paid: the price should agree with the order, and the quantity billed should agree with the quantity received. In the illustrative example on this page, the first shipment was invoiced at 970 units and received at 960, so at a unit price of 14.00 the invoice and the receipt differ by 140, and the 10 units are held in dispute until the shortage is resolved.

Which EDI documents does an apparel purchase order involve?

On the brand’s own buy, the brand is the buyer: it sends the 850 purchase order and any 860 change to the factory or agent, and receives the 855 acknowledgment, the 856 ship notice and the 810 invoice — where the factory exchanges EDI at all, since many work through a portal, a workbook or email instead. On its wholesale business the brand is the seller and the same documents run the other way: it receives the retailer’s 850 and 860, and sends the 855, 856 and 810.

• ✓The apparel purchase order lifecycle runs from the open-to-buy approval that authorises an order to the financial close that settles it, through eleven stages: OTB approval, buy plan, PO creation and terms, factory confirmation, materials and WIP, amendments, ship notice and packing list, transit, receipt and inspection, landed cost and three-way match, and close.
• ✓The ERP is usually the system of record for the order itself, but the order is shaped and changed in the plan, PLM, the factory’s milestone tracker, the forwarder’s systems, the warehouse system and finance — and re-keyed at each crossing.
• ✓Size and prepacks complicate every stage: each size is its own commitment, a factory cannot short one size inside a pack without breaking the ratio, and shortfalls tend to fall on bulk and on the channel bulk serves.
• ✓Amendments are where the plan and the ERP diverge: they are agreed outside both systems, written into one, overwritten rather than versioned, and recorded without a reason.
• ✓On-order should move on events — up at issue or at confirmation under one written rule, to in transit at the ship notice, to on hand at receipt by size — with open-to-buy consumed and released on the same events.
• ✓In the illustrative worked example, a 1,050-unit order ends its first delivery as 960 received, 60 open and three weeks late, 10 in dispute and 20 cut; the agreed view and three disconnected systems show 60, 90, 70 and 10 open.
• ✓Agreeing ownership is a sequence: map the stages, one writer per field, the date set, one amendment path, the grain, the on-order rule, the discrepancy rules, landed cost back to the plan, and a defined close with exceptions reviewed on a clock.

• Read how the 850, 855, 860 and 856 should feed the plan — apparel EDI →
• See how inbound purchase orders count toward available to sell →
• Read what the plan of record holds and who holds the pen →
• See which system owns the purchase order and every other object →
• See who is accountable for PO placement and amendment →
• See which decision belongs at which level — the planning grain →
• See the ten apparel software categories and where each one stops →
• Read why workflows break at handoffs rather than inside functions →
• Sequence a connected implementation one seam at a time →