Sizes and colors do not move at the same rate.
A garment shop runs on three coupled variables that nobody in a single role gets to see at once: the buyer size curve that lands by EDI 850, the cut plan that has to nest against the fabric lots actually on the floor, and the multi-region sub-contract sewing schedule that turns the cut order into garments. Polymr keeps the three variables aligned continuously.
| XS | S | M | L | XL | XXL | total | |
|---|---|---|---|---|---|---|---|
navy | 42 | 96 | 156 | 188 | 110 | 28 | 620 |
sand | 48 | 102 | 168 | 192 | 124 | 32 | 666 |
sage | 38 | 84 | 138 | 156 | 96 | 24 | 536 |
rose | 40 | 78 | 138 | 160 | 126 | 36 | 578 |
Built for the garment shape.
Polymr generates AI tech packs from a design brief. Sketches, mood boards, sample photos in, fully typed tech pack out with construction notes, BOM, grading specs, point-of-measure tables, and trim sheet. The same parser reads incoming buyer tech packs and emits a typed cut list plus a size matrix the cutting floor can sign against.
Scrap optimization runs across cut bundles. Mixed-size markers nest against the live roll width, fabric utilization is scored per marker, and bundle grouping is suggested before the marker locks. The fabric lot pointer rides every sub-contract sewing PO, so traceability from the bolt to the finished garment is intact whether the cut runs in-house or at a Tier-2 partner.
Bundles map size, color, and fabric lot in one row.
WO-CUT-22 decomposes the 2,400-unit PO into five cut bundles. Lot L-446 carries the navy L and sage L bundles together because they nest cleanly against the 58-inch roll. Lot L-447 went isolated to a single marker so its +2.1 dE shade variance does not mix into a multi-lot garment.
The XS and XXL units bin-pack into a single marker rather than burning a separate roll. Sub-contract sewing on V-218 and V-244 receives the bundles with the lot pointer attached.
| Bundle | Size | Color | Units | Fabric lot | Status |
|---|---|---|---|---|---|
| B-44-01 | L | navy | 96 | L-446 | cut · sewing W22-V218 |
| B-44-02 | L | sage | 80 | L-446 | cut · sewing W22-V218 |
| B-44-03 | M | navy | 78 | L-447 | cut · isolated dE 2.1 |
| B-44-04 | XL | rose | 64 | L-451 | queued · sub-contract V-244 |
| B-44-05 | XS+XXL | sand | 32 | L-451 | queued · bin-pack marker |
W19 ran 14% over forecast. The re-cut went out Friday.
The retailer EPOS feed lands every Friday evening. The sell-through tile shows forecast against actual for PMR-GR-22 navy L. When W19 ran +14% over forecast, the re-cut order issued Friday at 22:14 before the W22 fabric pull locked the following Monday. The buyer kept ship-window adherence without a buy-up call to V-218.
Six failure modes a multi-region garment shop lives with.
Each is what a cutting-room manager, a sub-contract liaison, or a buyer-facing planner names when asked why the season did not hit the contracted fabric utilization.
- Size-curve change mid-cutBuyer B-2204 shifts the curve from 4-12-22-32-22-8 to 6-14-24-30-20-6 two weeks before cut
Style PMR-GR-22 was already marker-built against the original curve. The buyer's merchandising team revises the per-region size distribution two Fridays out from the cut date, and the four hours of marker-nesting work done by the cutting-room planner against 58-inch FAB-9118 has to be redone from a blank sheet. Nothing in the prior nest is reusable because every per-size marker shifts when the upstream ratio shifts.
- Dye-lot variance not modelledLot L-447 reads +2 dE from L-446 but the marker still treats both as one bin
Both fabric lots came in against the same FAB-9118 PO and the warehouse stacked them on adjacent racks. The colourist flagged a +2 delta-E variance against the lab dip on L-447, but the marker plan was built against an aggregated yardage figure rather than per-lot. The cutting room ends up with a 1,800-unit run where 18% of garments mix L-446 panels with L-447 panels and the QC pass at finishing kicks them back as shade mismatches.
- Sub-contract ASN never gets reconciledSewing partner V-218 in Region South ships a mixed-style ASN against three open WOs
V-218 holds three open work orders against styles PMR-GR-22, PMR-GR-31, and PMR-GR-44. Their Friday ASN bundles all three into one packing list because the truck consolidates back to the finishing facility on the same lane. Receiving has to hand-decompose the carton-level packing list against the three WOs because no inbound system keys on the styles inside the ASN. The split takes the receiving team a full shift and downstream finishing starves on Monday.
- Cut-yield drift on style PMR-GR-22Yield drops from 84% to 79% because pattern weight was reweighted but the marker was not
Pattern engineering reweights the front panel on PMR-GR-22 from a 220 GSM to a 240 GSM during a buyer-driven fit revision. The new pattern goes into the spec sheet rev v4, but the marker library still references the v3 nesting solve. Five percentage points of fabric waste compound across every cut order for that style for three weeks before a planner notices the yield drift on a Friday recap report.
- Spec-sheet rev miscommunicated to finishingFinishing reworks 9% of style PMR-GR-31 because rev v3 was unclear on the hem treatment
Spec sheet rev v3 for PMR-GR-31 was issued mid-week with a hem-detail note that the finishing supervisor read as a blind-stitch hem; tech-pack intent was a coverstitch hem. The first 1,200 units come off the finishing line stitched the wrong way, the buyer's QC inspector rejects the carton on Tuesday, and the rework loop costs four production days plus the freight to bring the units back to the finishing facility.
- Buyer EDI change order across regionsEDI 850 change arrives Friday 5pm in Region North; sub-contract V-244 in Region South is already on Saturday production
The buyer revises the upsize ratio on PO-30441 via an EDI 850 change order at the close of business in Region North. The change should freeze the cut plan at sub-contractor V-244 across the time zone in Region South, but V-244 starts the Saturday shift on the old ratio because the EDI middleware does not push to the partner portal until the Monday morning batch. Two days of mis-ratioed sewing have to be re-distributed against the corrected size curve, and ninety units land outside the buyer's tolerance window.
What this looked like at a four-plant apparel and uniform maker.
- Situation
- A new style spec arrived weekly with a per-SKU size curve (XS-3XL) and a fabric + trims BOM. Each plant cut, sewed, and finished a slice of the order against shared raw inventory.
- What was breaking
- Cut-plan generation was a spreadsheet pass that took 4-6 hours per style and was redone whenever the size curve changed. Fabric utilization ran 84% against an industry-realistic 91% because nests were not optimized against current roll-width inventory.
- BOM extraction + RFQ
- Planning + purchasing
- Margin and bottleneck analysis
Where this page neighbours.
- Make-to-order
Configurable style families behave like configured-product MTO once the size curve and the colour callout are pinned to a buyer PO. The same quote-to-cut compression Polymr brings to MTO fabricators applies when a style spec lands with a per-region size curve attached.
- Discrete manufacturing
Revision propagation on garment BOMs (pattern weight, trim count, finishing routing) behaves the same way as discrete-shop revision blast: a pattern change has to walk every open cut, every sub-contract WO, and every finishing routing or the floor consumes the wrong rev.
- Industrial / multi-region
Multi-plant cutting, multi-region sub-contract sewing, and consolidated DC finishing make garment a multi-site operations problem in production-system terms. The same per-site planner queues, central operations roll-up, and consolidated procurement that industrial multi-site shops need are the same primitives garment uses.