Sourcing

When to Split One Promo Order Across Two Factories

Start with manufacturing fit, not administrative convenience

One supplier feels easier: one PO, one deposit, one balance payment, one shipment plan. For low-value, low-complexity promo orders, that simplicity is often worth protecting. But custom promo items do not belong to one manufacturing family. A 35 mm soft enamel pin, a zinc alloy bottle-opener keychain, a high-density woven patch and a 20 mm sublimation lanyard each run on different machines, materials, tolerances and defect controls.

That matters because many "full-line" promo suppliers are strongest in only one process group. Metal factories typically control mold engraving, stamping or die casting, trimming, polishing, plating, color fill and attachment assembly. Textile factories control loom gauge, embroidery stitch density, thread selection, edge finish, backing lamination and webbing conversion. A supplier that is excellent in metal but average in textile can still quote the whole set, but the weaker SKU often shows up later as frayed borders, fuzzy text, unstable color, loose hardware or inconsistent packaging.

Use four variables to frame the decision: process complexity, value concentration, schedule buffer and QC bandwidth. If one SKU represents 40-50% of total FOB value, is the campaign hero item, or has tight visual standards, specialization usually matters more than convenience. If the total order is below about USD 3,000 FOB, under five SKUs and tied to a hard in-hands date, single-source consolidation often wins because coordination risk becomes a larger share of total risk than manufacturing risk.

A useful threshold is this: if at least 70% of order value sits inside one process family, keep that family together unless there is a clear technical reason not to. If a mixed order includes mostly metal items plus one low-value textile add-on, bundling may be sensible. If the add-on requires fine woven text below 1.0 mm letter height, laser-cut clean edges, exact Pantone alignment or branded hardware, treat it as a separate manufacturing program rather than a convenience line item.

Run the RFQ by SKU, not by campaign

Buyers think in terms of a launch set. Factories think in terms of process capability. A stamped iron soft enamel pin, a die-cast keychain with a hinge, a woven patch with 0.8 mm lettering and a sublimated lanyard with a safety break do not share the same production logic. One is a mold and plating job, one is casting plus assembly, one is a loom-resolution problem and one is print plus sewing plus accessory attachment.

The quickest RFQ test is to ask each supplier the same four process-level questions for every SKU: is it truly in-house, what is the real MOQ, what tolerance can you hold, and what sample lead time applies to this exact process. A capable metal factory should answer with specifics such as 1.2 mm or 1.5 mm nominal thickness, overall size tolerance of ±0.20 mm, thickness tolerance of ±0.10 mm, plating options like bright nickel, bright gold, black nickel or antique brass, and attachment wire gauge or clutch type. A capable textile factory should state minimum line width, minimum readable letter height, backing options, edge style, webbing thickness and whether colors are matched visually or to Pantone reference.

If the supplier claims they can make everything but answers with broad phrases like "no problem" or "standard quality," assume they are consolidating through subcontractors. That is not automatically bad, but it is a different service model. Price it as a coordination service, not as evidence of deep process control. Ask whether they will disclose the actual factory type, sample sequence, domestic transfer path and who inspects before export.

A concrete example makes the split logic clearer. Suppose your order includes 1,000 soft enamel pins, 1,000 flat zinc alloy keychains, 150 woven patches and 150 lanyards. If the pins and keychains account for 85% of total FOB value and the patch and lanyard are simple giveaways, a strong metal supplier can lead. If the woven patch carries 0.8 mm letters, a heat-cut contour and premium hook backing, while the lanyard must hit a strict corporate blue on 300D polyester with a black safety break and gunmetal swivel hook, those textile SKUs should be quoted separately even at low quantity.

Split when the cost of failure is higher than the cost of coordination

The cleanest decision method is not highest unit price. It is highest failure cost. A lanyard with slight tone drift may still be usable. A challenge coin with shallow relief, pitted plating, unreadable edge text or an uneven antique finish is often a full remake. A keychain with poor rivet setting or weak split-ring wire can pass visual inspection but fail after distribution, which is usually more damaging because the defect reaches end users.

Adding a second factory creates predictable overhead. On typical China-origin promo orders, dual sourcing often adds USD 80-300 in bank charges, sample couriers, export paperwork and domestic trucking or transfer fees. It commonly adds 2-5 working days of coordination, and sometimes more if the goods must move to a third-party warehouse for recounting and repacking. That overhead is real, but it is usually smaller than the cost of remaking one visible SKU or expediting replacement freight.

A practical rule is to split when the expected loss from one failed SKU could exceed 8-12% of total order value. On a USD 8,000 FOB order, a defective hero SKU can trigger a USD 900-1,400 remake plus USD 250-700 in expedited air freight, plus launch damage if the event date does not move. In that situation, paying a few hundred dollars more in coordination cost to use a specialist is usually rational.

This is why experienced buyers compare risk-adjusted cost, not just piece price. A quote of USD 0.82 for a 35 mm soft enamel pin may beat USD 0.70 if it includes cleaner die lines, more even enamel fill, lower plating reject rates and tighter clutch alignment. The same applies to woven patches and assembled keychains. The specialist often wins not by being cheapest, but by being less likely to fail on the SKU that cannot fail.

Use MOQ, tooling and FOB math to stress-test the split

Splitting does not automatically increase cost. The real test is whether each SKU can be bought at a rational MOQ without forcing dead stock, duplicate tooling or inefficient freight. In 2026, custom promo categories still have relatively accessible entry MOQs, but pricing rises quickly below 300 pieces because setup, labor and inspection are spread across fewer units.

Product type	Typical MOQ	USD FOB range	Common tooling/setup	Useful technical benchmark	When splitting usually makes sense
Soft enamel pins, 30-40 mm	100-200 pcs	USD 0.45-1.10/pc	USD 45-90 mold	1.2-1.5 mm nominal thickness; size tolerance ±0.20 mm; thickness tolerance ±0.10 mm	Keep with one metal supplier unless plating quality, custom attachment or premium packaging is critical
Challenge coins, 45-50 mm	50-100 pcs	USD 1.90-5.20/pc	USD 80-180 mold	Relief depth usually 0.30-0.60 mm; text strokes above 0.20 mm; edge text needs clean registration	Split when coin is the hero SKU or includes 3D relief, cutouts, edge text or antique finish
Metal keychains with moving parts	100-300 pcs	USD 0.95-3.20/pc	USD 60-160 mold	Check hinge play under 0.30 mm, ring wire about 1.8-2.2 mm, and rivet pull strength consistency	Split when assembly durability, opener function or hardware finish matters more than lowest unit price
Woven patches, 60-80 mm	100-300 pcs	USD 0.38-1.30/pc	USD 20-50 setup	Lettering below 1.0 mm requires high loom resolution; border should show low fray and clean contour	Split when fine text, exact border style, hook backing or shape accuracy is important
Embroidered patches, 75 mm	100-300 pcs	USD 0.40-1.40/pc	Low or no setup	Thread coverage often 70-100%; stitch density and merrow consistency drive perceived quality	Split when raised embroidery, 3D foam, twill color or edge finish is visually important
Sublimation lanyards, 20 mm x 900 mm	100-500 pcs	USD 0.32-1.05/pc	USD 15-45 setup	Webbing thickness commonly 0.8-1.2 mm; width tolerance about ±1 mm; print registration and buckle spec matter	Split when color repeat, safety break specification or custom accessory finish is required

The biggest cost trap is duplicate setup on low-value add-ons. A 100-piece lanyard line worth only USD 45-100 FOB rarely justifies its own factory unless it supports a badge-access program, safety requirement or strict brand launch. By contrast, a 300-piece challenge coin order worth USD 700-1,500 FOB often does justify specialist sourcing because mold quality, polishing and plating variation affect perceived value far more than on a simple textile extra.

Also watch for MOQ inflation hidden inside convenience quotes. Some trading companies bundle attractively, then raise textile MOQs, downgrade premium backing to standard adhesive, or standardize hooks and split rings to protect margin. Ask for FOB by SKU, tooling listed separately, packaging listed separately, and accessories broken out by type. If the quote is not transparent at line-item level, you cannot compare bundled versus split scenarios honestly.

Build the timeline around two approval clocks, not one ship date

Two suppliers do not just mean two invoices. They create two artwork cycles, two sample queues, two production plans and often one final shipping date controlled by the slower factory. Typical production lead times after sample approval are about 10-18 calendar days for standard soft enamel pins, 12-20 days for assembled metal keychains, 14-22 days for challenge coins with deeper relief or specialty edges, 7-14 days for woven or embroidered patches and 7-15 days for sublimation lanyards. During plating backlog, pre-holiday rush or material shortages, add another 3-7 calendar days.

The hidden delay is usually sampling. If Factory A returns digital artwork in 24-48 hours and a pre-production sample in 5 days, while Factory B takes 3 days for artwork and 8-10 days for sampling, the project is already misaligned by nearly a week before bulk starts. Add 1-3 days for domestic transfer, then 1-2 days for receiving, recounting and repacking at the consolidation point, and your nominal buffer can disappear even without a technical failure.

For first-time importers, avoid splitting when the in-hands date leaves less than 10-12 calendar days of buffer after the planned ex-factory date. If you do split, assign one supplier or one forwarder as the formal consolidation point, lock carton mark format before production, and confirm the latest acceptable transfer date in writing. Also decide who pays domestic freight, who bears risk during supplier-to-supplier transfer and whether final carton count is checked before export release.

If one SKU is extremely schedule-sensitive, do not force full consolidation just for appearances. It can be smarter to ship the hero SKU separately by air if missing the launch would damage the campaign. Paying USD 180-400 extra freight for 200-500 pieces is often cheaper than a missed event, a partial rollout or dead stock that arrives after the promotion has passed.

Standardize finish, dimensions and packaging before chasing color match

Buyers often assume dual-source inconsistency will show up first in Pantone mismatch. In mixed promo orders, the first visible problem is usually finish, thickness and visual weight. One supplier may deliver a 1.5 mm bright gold pin with crisp polished edges, while another supplies a 1.2 mm keychain in a flatter gold tone with softer edge definition. Both may be acceptable in isolation, but together they do not feel like one coordinated set.

Solve that with a shared appearance specification above the individual drawings. Define metal tone family such as bright gold, bright nickel, black nickel or antique brass. State whether the target look is glossy, satin or antiqued. Standardize backing card size, paper weight and finish if products will be carded together, for example matte 350 gsm card, 90 x 55 mm, with 5 mm rounded corners and one euro slot. Keep logo clear space, border line weight and typography treatment consistent across every SKU. For lanyards, specify width, webbing hand feel, hook finish, buckle color and stitch color, not just the print file.

Color expectations must also be process-specific. Woven thread, embroidery thread, sublimation ink and metal enamel do not reflect light the same way even when referenced to the same Pantone chip. The right target is visual harmony, not literal identical color under every light source. On higher-value sets, request one approval board with all pre-production samples assembled or at least photographed together under neutral lighting before mass production starts.

Concrete language works better than generic brand notes. Example: bright nickel plating; smooth front face; no exposed burrs; enamel fill visually level; nominal front thickness 1.5 mm; backing card matte 350 gsm; logo minimum line weight 0.25 mm on metal and 0.40 mm on woven items; lanyard width 20 mm ±1 mm with silver swivel hook and black safety break. That gives two suppliers the same finish target instead of two different interpretations.

Only split if the QC standard is explicit and measurable

Splitting works only when acceptance criteria are written clearly. Otherwise you double the ambiguity. For metal items, useful checkpoints include overall size tolerance of ±0.20 mm for small items, thickness tolerance of ±0.10 mm where realistic, plating appearance, enamel fill consistency, burr control, attachment alignment and hardware twist or pull resistance. For textile items, QC usually centers on weave clarity, stitch density, edge finish, print registration, accessory count and pack accuracy.

For inspection, AQL 2.5 for major defects and AQL 4.0 for minor defects remains a common commercial baseline on promo products. For high-visibility brand launches, many importers tighten hero SKUs to AQL 1.5 major. Define major defects before production begins: wrong plating family, wrong backing, missing attachment, unreadable text, visible deep scratch on the front face, count shortage beyond agreed tolerance, wrong accessory or severe staining. Define minor defects separately: light thread fuzz, tiny backside mark, slight edge softness, or limited color variation within the approved sample range.

• Use one master spec sheet with SKU code, size, material, plating or thread type, backing, packaging and carton mark.
• Approve pre-production samples against end use, including wear, assembly and hardware function, not desk appearance alone.
• State remake triggers before deposit: wrong color family, unreadable text, weak hardware, count shortage or packaging error.
• Require packed quantity breakdown by polybag, inner box and master carton before balance payment.
• If consolidating, confirm domestic freight, transfer address, repacking method and damage liability in writing.
• Ask for dated in-line and final QC photos with caliper measurements, hardware checks and carton count labels.

As a budgeting rule, if the total order is under roughly USD 3,000 FOB, adding paid third-party inspection to a split order can consume too much margin unless the event is mission-critical. Once the order reaches about USD 5,000-10,000 FOB, or includes a hero SKU that cannot fail, splitting becomes more defensible if at least one supplier can show disciplined in-house QC records, measurement photos, plating or assembly checks and final packing verification.

Use a weighted scorecard and default to the simpler path unless specialization clearly wins

If the decision is still unclear, use a simple weighted scorecard instead of instinct. Score each sourcing path from 1 to 5 on technical fit, MOQ efficiency, tooling efficiency, sample speed, production reliability, packaging control and consolidation complexity. Double-weight technical fit and lead-time reliability, because those two variables create the most expensive surprises.

A single-factory strategy usually wins when the order is modest in value, under five SKUs, packaging must be uniform, and schedule buffer is tight. A split strategy usually wins when one hero SKU has demanding specs, represents a meaningful share of spend and can justify one more approval cycle without jeopardizing the ex-factory date. If the scores are close, default to the simpler path unless the specialist materially reduces failure risk on the most visible item.

The strongest RFQ method is to request two structures at the same time: one bundled quote from a capable lead supplier and one specialist quote for the highest-risk SKU. Compare more than FOB price. Check tooling duplication, MOQ by process, sample lead time in days, domestic transfer cost, packaging consistency, inspection method and who owns final consolidation. In many real orders, the best answer is neither one factory nor two factories in absolute terms. It is one lead supplier plus one specialist for the SKU most likely to fail if treated as an add-on.

Before sending RFQs, prepare a one-page decision sheet listing SKU, process, quantity, MOQ flexibility, must-match brand elements, latest sample approval date, required ex-factory date and non-negotiable QC points. That forces suppliers to answer the operational question that actually matters: does bundling reduce complexity, or does it hide process risk inside a convenient quote?

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