How Shein and Temu Changed Sweater Sourcing

By YZR Sweaters · 2026-11

industrial knitting machines in a Dongguan sweater factory running small-batch production

Quick answer

Shein and Temu did not invent cheaper sweaters — they removed the risk of guessing. Their model is small test batches, real-time sales data, and reorder-on-demand. Traditional buyers can borrow the logic without the disposability: test narrow, expand fast, and pre-agree the reorder terms that make speed possible.

Every buyer we work with has asked some version of the same question in the past two years: how can a platform sell a sweater for twelve dollars and still make money? The honest answer is that the sweater is not where the model wins. It wins on data, on inventory risk, and on a supply chain rebuilt around those two things. Understanding how it actually works — not the scare version, not the myth version — tells you what to borrow and what to deliberately ignore.

What the ultra-fast model actually does

The conventional season still works like this: design six months ahead, commit to a big order per style, hope the buy is right, then mark down what missed. Shein's model, which Temu's sellers effectively plug into, runs the opposite way. A style starts as a micro-batch — often 100 to 200 pieces, sometimes fewer. The listing goes live, and software watches what happens: click-through, add-to-cart, sell-through by size and colour. If the signal is weak, the style dies with almost nothing lost. If it is strong, production scales within days, and the winning style keeps being reordered in rolling small batches instead of one giant commitment.

The consequence is a complete inversion of where the money sits. A traditional retailer's biggest cost is markdown on the wrong buy. The ultra-fast platform's biggest cost is air freight and unit price — and it pays those willingly, because paying more per piece is cheaper than owning dead stock. That is the whole trade in one sentence.

Why knitwear is the awkward category

Sweaters resist this model more than most garments, and it is worth knowing why, because it is also why there is still room for specialists. A woven top can be cut and sewn in days from roll stock. A knitted garment usually needs yarn committed, greige knitted, linking, washing and finishing — each step with its own minimum sensible lot. A 3-gauge chunky cable knit cannot be turned around in a week at any price. So the platforms quietly concentrate on flat, fine-gauge knits — 12-gauge and up, synthetic or blend yarns held in stock — and they accept simpler constructions. Look at what actually trends on those platforms: basic crews, simple cardigans, brushed fleece pullovers. Almost no fully-fashioned silhouettes, almost no complex cables, very little real wool.

This is not a limitation they are embarrassed about. It is the design brief. When a buyer asks us to price "a Shein-style cardigan," the first question back is always: which parts of the specification are you actually willing to give up? Because the answer determines whether you are quoting a sweater or a different product that resembles one.

The parts worth borrowing

You do not need to become a fast-fashion platform to use the good ideas. Three of them transfer cleanly to a normal knitwear programme:

The parts worth refusing

Equally, there are things the model does that a brand with a reputation should not copy. The first is quality triage: a platform can tolerate a 20 percent return rate because the unit economics absorb it; a label that customers trust cannot. The second is specification drift — hitting a price by quietly downgrading yarn mid-season is normal in that world and fatal in ours; fibre content must stay within tolerance and on the label. The third is compliance shortcuts: the cheap end of the market is where RSL testing and traceability actually get skipped, and regulators in the EU and US are tightening exactly there. We have watched buyers learn this the expensive way: the platform price is real, but the platform risk profile is not one their brand can carry.

What a hybrid programme looks like

The strongest versions we see from mid-size brands look like this: a core range of fully-specified, tested sweaters — the brand's backbone, bought seasonally with proper inspection — plus a small test capsule of simple, fine-gauge styles in stock yarns, built to a four-to-six-week response, with reorder slots pre-booked. The capsule chases trend; the core carries margin and brand. When a capsule style wins, it graduates into the next season's core with a proper specification, or it dies quietly having cost almost nothing. That is the Shein lesson, minus the Shein cost.

If you want to run a test capsule like this, the conversation to have with us is early — before you design it, ideally — because the yarn choice and the construction decide whether a four-week response is physically possible.

Want a fast-response test capsule alongside your core knitwear programme? Tell us the target retail price and response window, and we will tell you what is realistic.

Get a free quote How capacity booking works

Frequently asked questions

Can any sweater be produced in the Shein style, in 100–200 piece batches?

No, and any factory that says otherwise is cutting corners somewhere. Fine-gauge knits in stock yarns with simple construction can run small batches economically. Chunky gauges, complex cables, fancy yarns and piece-dye programmes all have physical minimums. The honest approach is to design the capsule around what small-batch production suits, not to force every style into it.

Is it risky to source from factories that also supply ultra-fast platforms?

It depends on how the factory separates the work. The risk is specification drift and capacity contention — your order being bumped by a rush job. The protections are ordinary: a written specification, agreed lead times with penalty terms, and pre-shipment inspection. Our advice is to ask directly how a factory schedules platform work versus contracted brand work, and to walk away if there is no clear answer.

What should a small brand do first to borrow the test-and-reorder model?

Start with one style, two colours, a full size run, and a pre-agreed reorder slot. Agree the reorder price and lead time in the first purchase order so expansion does not need renegotiating. Then actually let the sell-through data make the call — that discipline, not the speed, is what makes the model work.