A first glove order fails more often on size mix than on price. This guide gives practical S to XXL ratios, EN ISO 21420 checks, carton maths, MOQ limits and PO rules for knitted and dipped work gloves from China.
Start with a default curve, not equal quantities
If your first PO is 3,000 pairs of 13-gauge polyester, nylon or HPPE palm-coated work gloves, do not split the order evenly across sizes. Equal 25 percent blocks look fair on a spreadsheet and usually leave you long on the smallest and largest sizes for 6 to 12 months. For a general industrial line sold in Europe or North America, a practical opening ratio is size 8 at 20 percent, size 9 at 40 percent, size 10 at 30 percent and size 11 at 10 percent. In wholesaler channels, size 9 is usually the first stock-out. Add size 7 only when the channel already serves smaller hands or mixed-gender assembly teams. This advice is for knitted and dipped work gloves, which are a Yiwu factory's core OEM category. It does not transfer directly to complex leather fashion gloves, batting gloves or motorcycle gloves, where finger pattern and panel shaping change the fit curve. Under EN ISO 21420, size 9 is referenced at about 229 mm hand circumference and 192 mm hand length, so ask the factory for the actual size chart used on the knitting floor instead of accepting a generic S, M, L and XL label set.
Build the ratio from wearer data, not the last supplier's carton
The right curve depends on who will wear the glove and what they do. A builder merchant selling 10-gauge or 13-gauge latex or nitrile work gloves often skews larger, with size 8, 9, 10 and 11 closer to 10, 40, 35 and 15 percent. A light-assembly distributor selling 15-gauge nylon or nylon-carbon PU gloves may need size 7 included and can open at 10, 25, 35, 20 and 10 across sizes 7 to 11. If your customer has no history, sample at least 30 wearers and record either their existing glove size or hand circumference. Task also changes the size acceptance window. A close-fit PU palm glove used for electronics or inspection is rejected quickly if the fingertips run 5 mm too long. A brushed-acrylic winter liner with latex crinkle has more tolerance because the glove is already bulky and often worn over a thin inner glove. Seasonal winter lines can also shift one size up when worn over a liner. Buyers who reuse one company-wide ratio across warehouse, assembly and cold-room work nearly always misforecast at least one size.
Know what size mixing does to MOQ and production cost
On the factory side, sizes are not just different cuff stamps. Size 8 and size 10 usually run on different hand formers, different knit programs or both, and the back logo position often shifts because knuckle spacing changes by several millimetres. A normal first-order MOQ for a knitted and dipped OEM glove might be 1,200 to 3,000 pairs per colourway, but that does not mean every later top-up can be tiny. If you run out of one size only, many factories still need 600 to 1,200 pairs of that size to justify knitting, coating, printing and packing setup. The penalty is sharper on expensive yarns. A 13-gauge cut shell using HPPE with glass fibre or steel fibre costs far more to rerun than a simple polyester shell, and waste on short runs is harder to hide. If the glove also carries a silicone transfer logo, heat-print patch or size-coloured overlock, the factory cannot easily swap stock between sizes after production. Bad size planning looks like a sales problem at the importer end and a setup problem at the factory end. It is both.
Pack by size or your warehouse will create the next problem
Most distribution errors start after production, not during knitting. For a standard 13-gauge nitrile palm glove, a common export pack is 12 pairs per inner bag and 120 pairs per outer carton. On thinner 15-gauge PU styles, some buyers go to 240 pairs per carton, but you should confirm the gross weight stays inside your warehouse limit, often 15 to 18 kg for manual handling. Whatever pack you choose, keep each carton single-size. A 3,600-pair order at 20, 40, 30 and 10 percent becomes 6 cartons of size 8, 12 of size 9, 9 of size 10 and 3 of size 11 if you pack 120 pairs per carton. Trouble starts when buyers ask for odd ratios that produce broken cartons, such as 17 or 23 percent. Then the factory is repacking partials, inner-bag counts get messy and the receiving warehouse has more chance to book stock into the wrong bin. Put the size code on the glove print, inner bag and outer carton. If you use GS1 labels or ERP scanning later, clean size separation at packing saves more time than arguing over one cent on unit price.
Validate with trial wear data and size-level QC
Before you place a full 20-foot or 40-foot order, run 300 to 600 trial pairs through two or three customer sites for 30 days. Record three numbers by size: exchanges, user complaints and actual monthly consumption. If exchanges run above about 3 percent, either the ratio is wrong or the pattern is inconsistent between sizes. Ask what failed. Palm width, finger length, thumb crotch and knit wrist tightness are different factory corrections, and they do not get solved by just moving more stock into size 10. At final inspection, do not stop at workmanship under AQL 2.5 or 4.0. Pull finished pairs from every ordered size and measure total length, palm width and middle-finger length against the approved spec sheet. On dipped gloves, fingertip coating creep can make a shell feel half a size shorter even when the knitted liner passes the tape measure. That is why a single pre-production sample in size 9 is not enough if your bulk order covers sizes 7 through 11.
Write size rules into the PO so repeats stay consistent
Your purchase order should state the ratio, the numeric size mapping, the pack by size and what happens if one size falls short. A workable commercial rule is plus or minus 5 percent on total quantity but no more than 2 percent variance on any single size without written approval. Also say clearly that the factory may not replace missing size 8 with extra size 10 just to hit the total pair count. This sounds obvious, but shortages are often hidden inside a correct grand total. For repeat business, keep a 12-month size matrix by SKU, country and channel. The UK, Germany and North America do not always pull the same curve even on the same 15-gauge nylon or PU assembly glove. After two or three replenishment cycles, most importers can narrow the size forecast better than they narrow the price. Buyers who manage size data properly place cleaner reorders, hold less dead stock and spend less time blaming the factory for what was really a forecasting error.
Coming to Yiwu or Hangzhou?
We host roughly 40-60 buyer visits a year. Workshop A & B run Mon-Sat; Workshop C (cut-resistant) Mon-Fri. Book a slot two weeks ahead and we can pull random samples from any active production line for you to inspect.
Sourcing is messy work. If this article saved you a 90-minute call with a trader, share it with one other importer who needs to see it.