U The Cycle: Eighteen Months of a Garment

Every chapter in Part 1 sliced the business by topic — money, factories, retailers, warehouses. This one slices it by time. We follow a single season, Fall/Winter 2027, from the afternoon it is a list of twenty-four style names on a whiteboard to the morning its last leftover unit ships to an off-price buyer — eighteen months later. At every step: what physically happens to the clothes, who acts, and precisely which record the ERP writes. When you finish, the whole machine should run in your head as one loop, because that is what it is.

What you need to know first

Three clocks run at once in this business, and the whole art of wholesale is that they are offset:

The selling clock. Buyers commit to a season roughly six months before it reaches their stores — chapter H's "selling forward." FW27 sells in showrooms and trade booths in January–February 2027 for delivery in August–September 2027.

The making clock. Factories need three to five months from a confirmed production order to finished goods on a boat, plus transit — and their inputs (fabric especially) have lead times of their own that sometimes have to be gambled on before the season sells.

The cash clock. Money leaves first and returns last. You pay for samples a year before revenue; you pay factories months before buyers pay you; terms and deductions stretch the gap further. Chapter K called this the working-capital cycle; this chapter shows you exactly where each payment lands on the timeline.

Dates below use FW27 as the worked example. Real calendars shift by category and geography — a swim brand and an outerwear brand run the same loop months apart — but the sequence never changes, and the sequence is the lesson.

The map: one season, end to end

Hold on to the loop's grammar as you read: SELL creates promises out. MAKE creates promises in. FULFILL converts stock into revenue by keeping the first promise with the second. COLLECT converts revenue into cash. CLEAR disposes of the error between what you guessed and what was true — and PLAN, next season, starts from what CLEAR taught you.

Months 1–3 · The line is planned (Aug–Oct 2026)

What happens. Merchandising (chapter E) decides what FW27 is: how many styles, in which categories, at which target prices, informed by what FW26 actually did — which is why the loop's end feeds its beginning. The output is the line plan: perhaps 24 styles — 9 carryovers (proven sellers returning, maybe re-colored), 15 new — each with a target wholesale price and a target margin before a single sketch exists.

Who acts. The owner and whoever does merchandising; the coordinator sanity-checks factory capacity; finance sanity-checks how much season the cash can carry.

What the ERP records. The season itself as a first-class record with its dates and its channel; placeholder styles — a name and a category slot, often before they have a style number. This sounds trivial and is not: the line plan is the season's budget, and every later screen ("how does the book compare to plan?") needs the plan in the database, not in a deck. A subtle schema consequence: styles must be allowed to exist before their numbers and codes are final, without breaking every uniqueness rule — a "slot on the line plan" is a real entity.

Months 3–6 · Design, protos and the cost sheet (Oct 2026–Jan 2027)

What happens. Each planned style becomes a tech pack — the engineering drawing of a garment: measurements by size, fabric, trims, construction, labeling (chapter G). Factories sew proto samples; protos get fitted, revised, re-sewn. In parallel the factory quotes, and the cost sheet (chapter F) fills in: fabric, trims, labor (CMT), freight estimate, duty — a landed cost per unit, against which the target wholesale price from the line plan either survives or forces a redesign ("value-engineering": that lining becomes cheaper, that fifth colorway dies). Styles that cannot make margin get cut from the line here, when cutting them costs a sample, not an inventory position.

Simultaneously, the selling artifacts get made: salesman samples (one full size-set of every style-colorway per rep or showroom — a real production run in miniature, and a real cost: a 24-style line shown by four reps is hundreds of garments made months before any revenue) and the linesheet, the priced catalog buyers order from.

What the ERP records. The product matrix becomes real: styles → colorways → SKUs (color × size), the three-level structure chapter E insisted must be rows, never naming conventions. Cost sheets as line-item records per style (never a single "cost" number — the components matter, because freight and duty move independently later). Sample inventory, ideally tracked as its own stock class — samples that vanish into reps' trunks are a real annual loss. And the linesheet should be generated from the database, not drawn in a layout tool; chapter H's exercise makes you do exactly this, because a hand-made linesheet is a price-disagreement generator.

Months 6–7 · Market: the book fills (Jan–Mar 2027)

What happens. The selling season — showroom appointments, trade shows, reps on the road, the whole apparatus of chapter H and chapter T. Orders arrive in every format the industry knows: tapped into a portal, scribbled at a booth, emailed as a buyer's own PO document, transmitted as an EDI 850 by a major. Each is a futures order: quantities by SKU, wholesale prices as shown (or negotiated, recorded, and approved), and a ship window — start-ship to cancel date — inherited from the season's delivery groups.

What the ERP records. The order book — and the grain of it decides what your company can know later. An order is a header (account, terms, window, buyer's PO number); lines at the SKU level (a "size run" of a style-colorway is a row per size, however the UI displays it); and, for orders that ship in parts, deliveries — "400 units in the August window, 600 in September" is one line with two delivery records, not two orders. Credit gates run at confirmation: an order from an over-limit or factor-declined account parks as a draft with a reason, and confirmation is a distinct verb from entry, done by someone with the authority the credit decision requires (chapter K).

Two disciplines separate a real order book from a spreadsheet of wishes. Bookings are not revenue — the book is a ledger of promises, and nothing in it touches the income statement until goods ship. And the book must stay editable without losing history: buyers revise, cancel, extend, and swap for months, and every revision needs its who/when/why, because the book is what production is about to bet real money on.

Month 7 · The book closes and the cut is decided (Mar 2027)

What happens. The most consequential arithmetic of the year. With the book substantially in, the coordinator and owner decide what to actually manufacture. For each SKU: confirmed demand from the book, plus a stock bet (extra units for immediates, reorders, and DTC — pure judgment, informed by last season's sell-through), minus anything already covered (carryover stock, earlier runs), tested against factory MOQs. Styles that sold below minimum face the season's hardest calls: eat the minimum, renegotiate, or cancel the style and unwind its orders — each a phone call nobody enjoys.

What the ERP records. The cut sheet as a computed screen — demand minus coverage by SKU, margin visible per row — and then the decision itself as production orders (purchase orders to full-package factories, or production runs staged through dye/cut/sew vendors, per chapter G's sourcing models). Each carries quantities by SKU, the agreed cost, and dates back-scheduled from the buyers' cancel dates: goods must finish, sail, clear, land, and ship to stores inside the windows sold at market. From this moment the pipeline has a second book — the inbound book — and the system's core identity becomes keeping the two books honest against each other: on-order rises now, so that available-to-sell (on-hand + on-order − allocated, chapter 1) can tell reps the truth about late-season immediates.

Months 7–11 · Making: fabric, dye, cut, sew (Mar–Jun 2027)

What happens. The physical sequence, compressed from chapter G: fabric is milled or drawn from booked greige; lab dips (small dyed swatches) go back and forth until the brand approves each color; fabric is dyed, relaxed, spread and cut against markers that minimize waste; bundles are sewn, finished, pressed, tagged and packed — each stage possibly at a different vendor with its own dates, which is why a "run" is naturally modeled as stages, not as one factory and one date. Along the way: TOP samples (top-of-production — first units off the line, approved before the bulk proceeds), in-line QC, and a steady drip of reality: the dye house is a week behind, a trim missed its boat, sizes are running long.

Who acts. The coordinator, daily — chasing, recording, re-planning. Every factory update lands as revised stage dates; every revision re-asks the same question: which orders does this endanger? That query — join the endangered run lines to the order deliveries they cover, compare dates to cancel dates — is the coordinator's heartbeat, and it is only answerable if the cut sheet recorded which demand each run covers.

What the ERP records. Runs with stages (vendor, cost, dates each); revisions with reasons; and payments going out — deposits at placement, balances at ship, or letters of credit, per the factory terms. The money loop's outbound half runs through here, months before the inbound half exists.

Month 11 · Inspection, the boat, the border (Jun–Jul 2027)

What happens. Finished goods pass a final inspection — commonly an AQL sampling plan (ANSI/ASQ Z1.4): a statistically sized random sample is inspected and the lot passes or fails on the count of defects. Pass, and cartons go to port; the ex-factory date is hit or missed. Then freight — for FW27's main deliveries, ocean (weeks) with air (days, at multiples of the cost) reserved for whatever slipped too far — then customs: the entry filed against each style's HTS classification, duty paid on the declared value. Chapter G covered Incoterms — who owns the goods and the risk on each leg — and chapter F covered why duty and freight belong in landed cost per unit, not in an overhead bucket.

What the ERP records. Shipment-level events on the inbound book: inspection results, ex-factory confirmations, vessel ETAs, entry numbers, and the actual freight and duty amounts that turn the cost sheet's landed-cost estimate into the season's real number. If actuals never flow back, every margin report for FW27 is a guess wearing a suit.

Month 12 · Receiving day (Jul 2027)

What happens. The container backs up to the dock and the two books finally touch. The warehouse counts cartons in against the run: 4,772 units landed where 4,800 were ordered — a short shipment on two SKUs, an overage on one, and one carton of the wrong colorway. The coordinator rules on the gap (chapter T): accept short and close, or hold the run open for the balance.

What the ERP records. The pivot on which the whole data model turns: receipt movements — on-hand rises, on-order falls, both in one transaction, each unit released against the run at no more than what was outstanding (over-receipts are their own explicit event, never a silent negative). From this moment the goods are real, ATS is backed by shelves instead of promises, and every downstream act is chapter J's territory.

Months 12–14 · The window: allocate, pick, ship (Aug–Sep 2027)

What happens. The season's second crescendo. As each order's start-ship date arrives, its stock is allocated — physically still on the shelf, logically spoken for. When supply is short (it always is, somewhere), allocation is a fairness policy wearing math: who gets the 48 units that landed against 60 sold? Ship complete versus ship short, key accounts first or pro-rata, call the buyer or just ship — decisions made in bulk, order by order is for amateurs. Then the floor: pick, pack to each retailer's spec, label (GS1-128 for the majors), ASN before the truck, tracking numbers back onto the order (chapters I and J). Goods that miss a cancel date do not ship at all without a begged extension — they become instant leftovers.

What the ERP records. Allocation as reserved units at the delivery grain (so partial windows work); pick/pack/ship as ledger movements; the shipment document itself — cartons, contents, ship-to snapshot, tracking — which is simultaneously a logistics record, the trigger for the invoice, and (chapter I) the evidence file for the deduction fight to come. The ship confirmation closes the goods loop: promise kept.

Months 12–16 · The money loop closes (Aug–Nov 2027)

What happens. Each shipment births an invoice — for what shipped, not what was ordered — on the account's terms. Invoices route to the factor (assigned, advanced against, collected by them) or to house AR dunning. Buyers pay on their own clocks, minus chapter I's deductions; finance matches remittances, fights the fightable, eats the rest, and the season's gross becomes the season's net.

What the ERP records. Invoices traceable to shipments traceable to orders (the custody chain of chapter T); payments; deductions with reason codes and outcomes; commission accruals on whichever trigger the rep agreements name. When the dust settles, the ERP can answer the only question that grades the season: net collected dollars per style, against landed cost per style — true margin, the number the next line plan is built from.

Months 14–18 · Returns, markdowns, the true margin (Nov 2027–Jan 2028)

What happens. The error term gets disposed of. Authorized returns come back (RA numbers, inspection, restock or write off). Leftover stock — the gap between the stock bet and reality — cascades down the clearance ladder: reorders and immediates first, then DTC promotion, then off-price (Chapter D's jobbers and closeout chains, at or below cost, with the channel-conflict rules chapter H explained), then donation or destruction — increasingly regulated, per chapter M. Meanwhile the season review happens: sell-through by style, fill rate by account, deduction rate by retailer, margin by everything — the CLEAR stage feeding the next PLAN.

What the ERP records. Return receipts as inbound movements with dispositions; markdown sales in their channels; write-offs as explicit inventory adjustments with reasons (shrink hidden inside "adjustment" is a lie your balance sheet tells you); and the season's final scorecard, which is not a new record at all — it is a read over everything the previous seventeen months wrote, which is the entire argument for having written it carefully.

Three seasons at once

Everything above described one season alone. The company never experiences it that way. In any given week, three seasons overlap: FW27 is shipping, SS28 is selling at market, and FW28 is being line-planned — three positions of the same loop, running through the same ten people and the same database simultaneously.

This is why everything in the ERP must carry its season. A stock number, an order total, a margin report — meaningless unless scoped to a season, because the building contains three of them at all times. It is also why "the busy season" is a myth: the tides of chapter T come from different seasons' stages landing on different desks in the same week.

What this means for your ERP

The cycle, restated as design requirements:

• Two books, one arithmetic. The order book (promises out) and the inbound book (promises in) are the system's spine, and ATS = on-hand + on-order − allocated is the equation that joins them. Every term must have exactly one writer and a ledger behind it.
• The season is a first-class entity — with dates, delivery groups, and a foreign key from nearly everything else. Three live at once; unscoped numbers are noise.
• Grain decisions are destiny. Orders at SKU grain with delivery-level splits; runs with stages; costs as component lines; counts as deltas. Every place this chapter said "recorded as," the grain was the real decision — coarser grains cannot answer the questions later stages ask, and no report can recover detail that was never written.
• Commitments before events. The riskiest money (greige bookings, sample spend, factory deposits) moves before the demand exists. A system that only records events-that-happened, and not promises-outstanding, is blind exactly where the company is most exposed.
• Back-schedule from cancel dates. Every production date is derived from a buyer's window. Model the dependency, and slippage becomes a query ("which orders does this endanger?") instead of a panic.
• Actuals must flow back. Freight, duty, deductions, commissions — the estimate lives in the plan, the actual lives in the event, and margin reports read the actuals. An ERP that never reconciles estimate to actual grades every season on a curve.
• The custody chain is sacred. Order → coverage → receipt → allocation → shipment → invoice → payment. Each link traceable to the last. This chain is simultaneously your operations, your audit trail, and your dispute evidence; break it anywhere and three departments feel it.

Field notes & further reading

• ASQ on acceptance sampling and AQL: what an AQL inspection actually is, from the standards body behind ANSI/ASQ Z1.4. Read it before you accept or write "AQL 2.5" in a PO, so you know what statistical promise those characters make.
• ICC Incoterms® rules: the official home of FOB, CIF, DDP and friends — who pays and who bears risk on each leg between the factory and your dock. The inbound half of this chapter's timeline changes shape with each term.
• USITC Harmonized Tariff Schedule: already cited in chapter H and it belongs here too — the duty rate that lands in month 11 was determined by a classification decision made back in the costing months. Look your garment up; do not trust hearsay.
• American Apparel & Footwear Association: the U.S. industry association; its publications and seminars are where sourcing calendars, compliance changes and trade-policy shifts get discussed by the people living them.
• Chapters E, F, G, H, I, J and K each expand one or two stages of this loop to full depth. This chapter is the film; those are the frames.