Sourcing from China for a Shopify Brand: MOQ, Variants and Fulfilment Math (2026)

A marketplace seller buys inventory and lets the platform carry the customer. A Shopify brand buys inventory, pays for the customer, ships the parcel, absorbs the return, then has to sell that customer again. The sourcing decisions are not the same, and the ones borrowed from FBA playbooks are usually the expensive ones.

The pattern repeats constantly. Someone builds a decent store, sources a product the way a course told them to, lands a container of one thing in three colours, and then discovers the money did not go into marketing — it went into cartons aging in a 3PL.

That is not a supplier problem. It is a sizing problem decided months earlier, in a spreadsheet nobody built.

1. Why DTC sourcing is a different problem

The structural difference is who holds the risk. On a marketplace, unsold stock is a storage fee. On a DTC store, unsold stock is cash you cannot spend on acquisition, in a channel where acquisition is the constraint.

Read the variant row twice. On a marketplace, one listing absorbs demand for every colour and size and the platform manages the split. On your store, each variant is a page with its own traffic, conversion rate and stock-out date — which is why an MOQ that feels fine for FBA feels suicidal for DTC.

2. MOQ is a cash-flow number, not a supplier quirk

Buyers treat minimum order quantity as a negotiation obstacle. It is more useful as a unit of time. Divide it by your daily sales rate and you get days of cover — the number that says whether the order is a business decision or a bet.

The same 600-unit MOQ is a comfortable two-month commitment for a brand doing twenty orders a day and a four-month millstone for one doing five. The supplier quoted the same number to both, because the supplier does not know your sell-through. Only you do.

• Negotiate against a forecast, not a wish. Factories often accept a lower first MOQ when you show a reorder plan and commit to a second order inside a stated window. What they want is a relationship with volume, not a one-off.
• Pay for the privilege in price, not in silence. A small first run costs more per unit because set-up is spread over fewer pieces. Accept that, keep cash exposure low, and renegotiate at the second order.
• Confirm whether the MOQ is per SKU. "1,000 units total" and "1,000 units per colour" are different requests, and quotations often mean the second. Get it in writing.

If demand is a handful of orders a day, a short run through a trading company is the right first step even at a worse unit price.

3. The variant trap

Variant proliferation is the most common self-inflicted wound in DTC sourcing, and it always looks reasonable when planned: three colours because the range looks better, four sizes because customers expect them, two materials because the premium version might sell.

The first order then arrives as twelve stock positions. Four sell, three trickle, five are dead by month three — not because marketing failed, but because you bought a range instead of a hypothesis.

1. One hero variant, bought deep enough to sell. One colour, or a sensible size curve in one style. This is the SKU you market, photograph and advertise.
2. One second variant, bought shallow — enough to test the page, not enough to hurt. Some factories will fold it into the same MOQ as a colour split.
3. Everything else as pre-order or waitlist. A waitlist is free demand data and tells you which variant earned an order rather than a compliment.
4. Add variants only after the hero SKU has a stable sell-through rate across a full reorder cycle. Then you are buying against your own data, not a competitor's review page.

4. Private label, white label, OEM — pick deliberately

The words get used interchangeably in supplier conversations, which is how buyers end up paying tooling money for a stock product. The distinction that matters is how much of the product is yours.

Most first launches should be private label. A steel injection mould is a five-figure line item with a schedule measured in weeks before a single saleable unit exists. Packaging, inserts, labels, a considered colourway and a good photograph deliver most of the perceived-brand benefit for a fraction of that, and can be changed after launch.

Custom tooling is a reward for proven demand, not a substitute for it. It also raises ownership questions — who owns the mould, what happens if you leave, whether it is amortised into the unit price — covered in the OEM and white-label guide.

5. Landed cost per order, not per unit

DTC brands die on a spreadsheet that stops at "product cost plus freight". The costs that decide profitability arrive after the product does. Here is a worked example — the numbers are illustrative, the structure is not.

Contribution before marketing is 23.49, or roughly 48% of revenue — the number your acquisition budget actually competes with. Note that the product is less than a third of the total: freight, postage, fulfilment and fees together outweigh it, so shaving five cents off the unit price is low-leverage against fixing an oversized carton. The returns reserve is not optional either; a store that does not budget for returns is not more profitable, just measuring later.

Duty is assessed on the customs value, which normally tracks what you paid. Misclassification is not a clever saving; it is how importers acquire back-duties and penalties.

6. Fulfilment: four models, four stages

There is no best fulfilment model, only one that matches your volume — the mistake is staying on the launch model too long.

The transition point is not a magic order count. It is when the cost of slow delivery — refunds, chargebacks, customers who never return — exceeds the cost of holding inventory.

7. Packaging is a sourcing decision, not an afterthought

For a DTC brand the parcel is the first physical brand touch, and often the only one before the product. It also has an MOQ, a lead time and a specification, so it belongs in the sourcing plan rather than the week before launch.

• Decide packaging before the product order is confirmed. Carton dimensions drive freight and mailer size drives postage. Boxes two centimetres oversized are a permanent surcharge.
• Buy packaging separately where possible. A specialist supplier has lower MOQs, better print and shorter lead times than a factory printing boxes as a sideline. The trade-off is coordinating two suppliers instead of one.
• Specify the insert, not just the box. A printed card with a QR code to a reorder offer costs little per unit and does more for repeat rate than most acquisition tactics at the same spend.
• Test the parcel before the run. Drop it, ship one to yourself, open it as a customer would. Damage in transit is a returns line.

If product and packaging come from the same factory, annex the dieline, print colours and material to the purchase order. Verbal packaging approvals are the most common source of "we thought you wanted the other finish". The specifics are in the custom packaging guide.

8. Compliance sits on you, not on the factory

As brand owner and importer of record you are the party a regulator or marketplace will contact. The factory can supply test reports and declarations, but it cannot take your liability or manage your labelling.

• Product safety and the applicable standard. Which standard applies depends on the product, market and age group. "Our product passes CE" is not a statement of anything: a self-declared declaration is not a test report, and neither is a certificate for a different model. Confirm scope, model and standard on the document itself.
• The responsible economic operator. Selling consumer products into the EU requires a responsible party established in the EU, with an address on the product or packaging. If you ship from China direct to EU customers, that entity has to exist in your chain somewhere.
• Labelling and language. Warnings, importer identity, batch marking, age grading, battery and material declarations. All of it is print — cheap before the run, expensive after.

None of this is exotic, but it is specific and it changes. Treat any summary, including this one, as a map rather than the territory: check the current official text for your product and destination before ordering. EU-facing requirements are set out in the EU product compliance guide.

9. The first-order sequence

Ninety to a hundred and twenty days, honestly scheduled. Compressing it is possible; skipping steps is what costs money.

1. Write the specification before you shop. Dimensions, materials, tolerances, finish, packaging, labelling, test requirements. One page with photographs — this is what you quote against, inspect against and argue with.
2. Get three quotations on identical scope. Different scopes produce incomparable prices. Ask each supplier to price the same specification and state what is excluded.
3. Buy samples from two suppliers, at your cost. Free samples create obligation and hide who is serious. Test them, do not just look at them.
4. Approve and sign a golden sample. One sealed unit with both signatures, attached to the specification. Your only defence against definition drift.
5. Place the order with staged payment. Deposit up front, balance against a passed inspection or a bill of lading — never against a promised ship date.
6. Book packaging and freight in parallel. Print proofs and the forwarder's booking both overlap with production, not with shipping.
7. Inspect before the container closes. Pre-shipment inspection against a defined defect standard, plus a during-production check if the product is new.
8. Onboard the 3PL before the goods arrive. Receiving appointments, SKU labels and carton marking agreed while the vessel is still at sea.
9. Launch on one hero SKU. Spend acquisition on a single page you can optimise, not a range you cannot afford to restock.

If this is your first order overall, the sequencing matters more than the supplier you pick — the three-step first-order plan is a shorter version of the same path.

10. Reorder points, calculated backwards

Most stock-outs are arithmetic failures, not demand surprises. Build the lead time from its parts.

Then reorder at: daily sales rate × (lead time + review period) + safety stock. At twenty orders a day and a ninety-day lead time, you need 1,800 units in the pipeline before you have sold the first one — which is exactly why the first MOQ conversation matters, and why a product that cannot support that volume is not yet a DTC product, however good the sample was.

Air-freighting a partial quantity works as gap cover, at a higher cost per unit than planning correctly. Insurance, not a strategy.

The Bottom Line

A Shopify brand's sourcing problem is not finding a factory. It is committing cash to the right quantity of the right number of variants, with a landed cost that leaves room for acquisition. MOQ divided by daily sales rate, a deliberately small variant set, and a per-order cost model will tell you more about whether a launch works than any supplier negotiation.

Start narrow, buy one thing well, pay a little more per unit for a smaller commitment, and let your own sell-through data size the second order. The brands that scale are the ones that still had cash when the second order was due — not the ones that bought the biggest first container.

— Jack Zhang, ChinaBizGuides