Sourcing

When to Split One Promo Order Across Two Factories

One PO can hide four different manufacturing risk profiles

A mixed promo order can look simple on paper: 1.25 inch soft enamel pins, 70 x 70 mm woven patches, 20 mm polyester lanyards, and 50 mm zinc alloy keychains all shipping to one warehouse for a launch, sales meeting, campus event, or retail reset. Marketing wants one visual system, procurement wants one PO, and operations wants fewer invoices and fewer cartons. The sourcing error is assuming a supplier that quotes all four SKUs also controls all four processes well.

These items do not share the same production logic. Pins and keychains depend on die tooling, die-struck or zinc alloy casting accuracy, polishing, plating chemistry, enamel fill, curing, and hardware assembly. Patches and lanyards depend on loom setup or embroidery digitizing, thread or polyester input quality, weave density or stitch count, border finishing, print registration, sewing, and hardware attachment. Many trading companies and mixed-category suppliers are strong in one family and subcontract the other to a partner workshop 50 to 300 km away.

That distinction matters most on date-critical programs. If ex-factory is required in 25 to 30 calendar days from artwork approval, one subcontracted SKU slipping by 5 to 8 working days can block co-packing, booking cut-off, or retailer intake. The real decision is not single-source versus multi-source in theory. It is whether one supplier truly owns the bottleneck process, or whether your "one factory" model is actually a brokered chain with limited visibility and weaker schedule control.

Start with dependency mapping, not product count

First ask whether the order is operationally linked or only bundled for convenience. If every SKU must match one Pantone family, fit one backing-card system, pass one final inspection plan, and arrive together for one launch date, then the order is operationally linked. In that case, the slowest production path determines the usable schedule. If pins are for media kits, lanyards are for an event team, and patches are for e-commerce add-ons, then the bundle is mostly administrative and can be split with much less risk.

Require each quoting supplier to map the production path by SKU and identify what is in-house versus outsourced. A credible metal-item answer should name artwork check, die or mold creation, die striking or casting, trimming, polishing, plating, color fill, curing, attachment soldering or assembly, inspection, and packing. A credible textile answer should name strike-off, weaving or embroidery, border finishing, backing application, sewing, hardware attachment, inspection, and packing. If a supplier says it can do "everything" but cannot state where woven patches are made or who prints the lanyards, treat that as a warning, not a minor wording gap.

A practical sourcing rule is simple: if one late SKU prevents release of the entire program, manage the order to the latest path and the highest-risk process. If SKUs can ship and deploy independently, split where capability diverges. For most promo merchandise mixes, the natural divide is metal items on one side and textile items on the other.

Compare landed risk, not only the bundled FOB quote

A bundled quote often looks attractive because the supplier gives a package discount and the buyer sees only one negotiation. FOB benchmarks are still useful. For a standard 1.25 inch soft enamel iron pin with butterfly clutch and 2 colors, typical China FOB pricing is about USD 0.46 to 0.82 at 300 pcs, USD 0.36 to 0.68 at 500 pcs, and USD 0.28 to 0.52 at 1000 pcs, plus die charges usually around USD 45 to 85. For a 50 mm zinc alloy keychain with split ring, 2D relief and standard nickel plating, expect roughly USD 1.05 to 2.10 at 300 pcs, USD 0.88 to 1.72 at 500 pcs, and USD 0.68 to 1.35 at 1000 pcs, with mold charges commonly USD 80 to 180 and higher for multi-level 3D detail.

For textile items, a 70 mm woven patch is commonly around USD 0.34 to 0.72 at 300 pcs, USD 0.25 to 0.55 at 500 pcs, and USD 0.18 to 0.39 at 1000 pcs. Merrow border and standard backing usually cost less than laser-cut shapes, hook backing, or retail carding. A 20 mm sublimation lanyard with swivel hook and safety break often runs USD 0.52 to 0.96 at 300 pcs, USD 0.41 to 0.78 at 500 pcs, and USD 0.31 to 0.60 at 1000 pcs. Jacquard lanyards are often 12 to 25 percent higher than sublimation at the same MOQ because loom setup, yarn color selection, and slower throughput increase cost.

FOB alone is a weak decision tool if one category is poorly controlled. A bundled order that saves 4 to 6 percent on unit cost can become more expensive if one outsourced SKU fails approval and needs remake, split shipment, or air uplift. On a 500-piece program, remaking one textile SKU and air-freighting 1 to 2 cartons can easily add USD 250 to 900. If missing an event date or retail intake window costs sell-through or distributor access, the commercial loss is usually far greater than the initial bundled discount.

Decision factor	One factory usually wins when	Two factories usually win when
MOQ and setup cost	All SKUs sit near 300 to 500 pcs, packaging is simple, and no item needs a special retail presentation	One family becomes cost-efficient only at 1000+ pcs, or one SKU needs unique packing, backing cards, or barcode work
Lead time	All SKUs can sample and produce within 15 to 22 working days after approval	One family runs 5 to 8 working days slower or depends on outside strike-off, plating, or print scheduling
Process ownership	Supplier clearly controls the relevant family in-house and can name each checkpoint	One or more categories are subcontracted and workshop details or timing remain vague
QC discipline	Supplier provides measurable QC standards for metal and textile items separately	Supplier uses one generic QC sheet and cannot define tolerances or defect levels by SKU
Color in use	Items are not displayed side by side and cross-material match is less critical	Items must sit together on one card, in one gift set, or on one launch display
Logistics and packout	All SKUs ship bulk with no synchronized assembly or labeling dependency	Items need co-packing, arrival sequencing, FNSKU labeling, or retailer-specific carton marks

Sample timing reveals real capability faster than the sales deck

The fastest way to test whether a supplier truly controls multiple categories is to compare sample paths, not brochure claims. For standard metal items, a capable factory should usually return artwork review in 1 to 2 working days, tooling layout in 1 day, and a pre-production sample in about 5 to 7 days for soft enamel pins or 7 to 10 days for zinc alloy keychains. Bulk production after approval is commonly 10 to 15 days for 500 to 1000 pcs. Add 2 to 4 days for dual plating, glow enamel, epoxy dome, spinner parts, chain attachments, or individual card-and-bag packing.

Textile milestones should be equally specific. Woven patches often require 1 to 2 days for layout approval, 4 to 6 days for strike-off or sample, and 7 to 12 days for production depending on border type, backing, and color count. Sublimation lanyards are usually faster: 1 day for proof, 3 to 5 days for sample, and 7 to 10 days for production at standard 15 mm or 20 mm width. Jacquard lanyards normally add 2 to 4 extra days because loom setup and yarn confirmation extend the front end.

If a supplier gives exact timing for pins but says the patch or lanyard timing is "about one week" or "depends on workshop," read that as operating data. It usually means the supplier is coordinating rather than producing. A useful threshold is this: if sample timing differs by more than 5 working days between metal and textile families, or if one family has an undefined milestone after deposit, splitting the order is usually safer than keeping the program under one vendor.

Use SKU-specific QC standards with numeric tolerances

AQL can sit at the order level, but defect definitions must sit at the SKU level. For many promo programs, AQL 2.5 for major defects and AQL 4.0 for minor defects is workable for final random inspection on lots from 500 to 3000 pcs. The important point is not the AQL label by itself. It is how each product defines pass and fail.

For pins and keychains, major defects usually include wrong plating tone, exposed base metal on the front, missing clutch or ring, bent post affecting wearability, cracked enamel, sharp unsafe burrs, or attachment offset outside tolerance. Minor defects may include light polish haze on the back, tiny enamel dust points outside the main visual area, or slight color drift versus the approved sample. Useful control points include overall thickness tolerance of +/-0.10 to 0.15 mm, attachment position tolerance of +/-1.0 mm, outline size tolerance of +/-0.30 mm, and minimum separated line width of 0.20 to 0.25 mm for reliable soft enamel fill. Decorative plating layers often fall around 0.03 to 0.08 microns for appearance finishes; if corrosion resistance matters, specify a salt-spray target such as 24 to 48 hours for standard decorative use rather than assuming finish color equals durability.

For patches and lanyards, the controls are different. Woven or embroidered patches are commonly held to finished size tolerance of +/-1.5 mm, shape symmetry within approved artwork, clean merrow or heat-cut edge, and no loose thread ends beyond the agreed allowance. If iron-on backing is used, require adhesion consistency on the approved sample and define whether the patch is decorative iron-on or intended for wash durability. Lanyards should specify width tolerance of +/-1.0 mm, print registration, seam alignment at the join, hardware function, and breakaway buckle fit if included. A swivel hook that binds under normal hand pull, a buckle that opens unintentionally, or sublimation print that ghosts beyond the edge is a major defect, not a cosmetic note.

When one supplier uses the same generic QC sheet for enamel fill, woven edge quality, and lanyard hardware, disputes later are likely. That is often the clearest technical reason to split the order, even if the headline unit price looks competitive.

The split model that works most often is metal together and textile together

For most mixed merchandise programs, two factories is the cleanest structure. Put pins, challenge coins, medals, badges, magnets, and metal keychains with one metal-focused supplier. Put woven patches, embroidered patches, jacquard or sublimation lanyards, and other sewn or woven goods with one textile-focused supplier. This follows equipment, operator skill, and inspection logic instead of forcing one supplier to coordinate an unrelated category through subcontractors.

The benefit is practical rather than theoretical. A good metal factory can run die lines, polishing, plating, enamel fill, and assembly across several SKUs with one process language and one QC team. A good textile factory can manage weave clarity, stitch density, thread matching, border finishing, sewing, and hardware attachment with shorter feedback loops. That usually reduces remake risk, shortens approval cycles, and produces better answers during pre-production because comments come from the process owner rather than a sales coordinator.

Avoid over-splitting modest orders. If you have four SKUs at 300 to 500 pcs each, using three or four factories often creates more sample traffic, more carton mismatch, more pickup coordination, and more labeling errors than it removes. For sub-1000 piece launches, two factories is usually the practical ceiling unless one SKU adds electronics, RFID, batteries, molded PVC, or complex retail kitting.

• Keep one factory when all SKUs are standard, quantities are low, and the latest acceptable ship date still has at least 7 to 10 days of schedule float.
• Split into metal and textile suppliers when one family is outsourced, sample timing differs by more than 5 working days, or one supplier cannot define item-specific QC points.
• Request in-house versus outsourced disclosure before deposit for tooling, plating, weaving or embroidery, lanyard printing, sewing, assembly, and packing.
• Compare MOQs at 300, 500, and 1000 pcs using the same packaging assumptions so setup-driven price distortion is visible.
• Lock SKU-level controls in writing: AQL, dimensions, plating or color standard, hardware function test, packaging method, and carton labeling.
• Avoid three or more suppliers unless one SKU needs a special process such as RFID, battery components, molded silicone, or custom retail display assembly.

If you split, centralize color control, packing standards, and freight ownership

The main objection to split sourcing is brand consistency. That risk is real, but manageable if specifications are centralized before production starts. Issue one master art pack covering Pantone references, logo lockups, critical dimensions, backing-card artwork, barcode position, polybag warning requirements where applicable, and export carton mark format. If the pin and patch will sit on the same card or in one gift set, include a display-relationship sheet showing relative scale and placement so each factory is not optimizing its own item in isolation.

For color-critical programs, approve color physically by material family. Enamel, woven thread, and printed polyester reflect light differently, so exact cross-material identity is unrealistic. Controlled alignment is realistic. If the key brand color is Pantone 186 C, approve one actual enamel sample, one patch strike-off, and one lanyard sample against that reference before bulk starts. PDF proofs are not enough for a program where items will be displayed side by side.

Freight and packout also need one owner. China-side consolidation works well when schedules are stable, one forwarder handles pickups, and export carton sizes are standardized. For a mid-size order, domestic transfer or consolidation fees often add roughly USD 80 to 250, but that can still be cheaper than multiple destination handling events. If schedules are uncertain, destination-side consolidation is often safer because it avoids supplier-to-supplier blame when one factory ships late. If no pre-assembled kits are required, separate shipments with harmonized carton labels and receiving instructions are often the lowest-friction option.

A practical decision framework for current orders

Request two quote structures on identical assumptions: one all-in quote from a single supplier and one split-by-category quote from specialists. Ask for the same MOQ tiers, usually 300, 500, and 1000 pcs; the same FOB basis; the same packaging method; the same sample charges; and the same required ship window. Then compare more than unit price. Compare sample days, production days, declared in-house processes, tooling cost, defect criteria, and who owns final packing.

Next, identify the bottleneck before placing the PO. If pins and keychains sample in 6 to 8 days and produce in 12 to 14 days, but woven patch strike-off depends on an outside mill and lanyard timing remains vague after deposit, the schedule risk is already visible. If all SKUs are standard, quantities are low, and the supplier gives precise process-level timing and QC for every category, one supplier can still be the lowest-friction choice. If the quote is broad but the technical detail is thin, splitting early is almost always cheaper than recovering late.

The best sourcing decision is rarely about who can quote everything. It is about who can control each process with the lowest risk to ship date, defect rate, and freight escalation. When one warehouse date links the program, the weakest SKU path often becomes the most expensive line on the PO. That is the point where two factories are not adding complexity. They are removing hidden operational risk.

Have a project? Send your artwork and target quantity and we’ll reply with a detailed quotation within 12 working hours.