When your cheapest single factory becomes your most expensive one
Supplier diversification and backup sourcing in China - single-factory risk, finding, vetting and keeping a backup vendor, splitting volume and costing it, and MOQ flexibility clauses

Hello, this is GreenFrog Seoul.

A WeChat message arrives from the factory you have worked with for three years. It is ten days since you placed your peak-season order. "Sorry, we cannot free up the line this month. We can ship at the end of next month." It turns out a large European buyer dropped a pile of orders at once and the whole line went to them.

You try to find another factory, but there is no time. Getting and approving new samples alone takes a month, and the mold sits at the current factory. In the end you wait six weeks, and in that gap your stock on Smart Store and Coupang runs out and your search ranking drops. Even after the goods come back in, the ranking does not recover easily.

The factory did nothing especially wrong in this story. From its side, it simply looked after a bigger customer first. The problem is that you had no other option at all. Today's topic is how to fill that gap: supplier diversification and backup sourcing.

Spreading supply risk came up briefly in Episode 37 as one part of stable reordering, country-level diversification into Vietnam or India was covered in Episode 56, and moving production out of a factory entirely was covered in Episode 99. This article sits between them. We dig into how to run two factories inside China at the same time, as a matter of routine.

When we talk to importers after a supply failure, we hear almost the same line every time. "We did look into a backup factory, but we never actually placed an order." A backup that exists only as a contact and a quotation does not work in a crisis. A backup factory is only a backup if it is already taking some of your volume.

1. How single-factory dependence breaks

Everyone knows that relying on a single factory is risky. Far fewer people have pictured exactly how that setup falls apart. The patterns we see again and again in consultations come down to roughly six.

The column to look at is the warning period. Finding a new factory, getting samples, vetting them and reaching mass production takes six to ten weeks even when it goes fast, and almost none of the six situations above gives you that much notice. If you start looking for a backup after the incident, you are already too late.

External shocks in particular are often not one factory's problem. When an environmental inspection hits an industrial park, every coating and plating plant in that park stops at once. Who carries the loss under the contract when force majeure delays delivery was covered in Episode 107, but however good the contract is, the goods still do not ship.

2. Why "the single cheapest factory" ends up costing more

Lining up quotations and giving all the volume to the cheapest one is a natural choice when picking a factory. Concentrating volume can bring the unit price down further. The problem is that this calculation assumes a year in which nothing goes wrong.

Let us widen the calculation a little. Say you have one core product with annual purchases of about KRW 500 million.

On unit price alone the single factory clearly wins. Add the cost of one stock-out and the picture changes. When a core product is out for three months, you lose that period's sales, and on top of that an online seller loses search ranking and review momentum while a B2B supplier shakes its customers' trust. Rushing to find another factory often means going into mass production with thin vetting, so quality problems stack on top.

The bigger cost, though less visible, is negotiating leverage. Factories know more accurately than you might think whether a buyer has an alternative. They guess from whether sample requests have gone out to other factories, whether your order pattern wobbles, and how hard your buyer pushes on price. Once a factory knows you have no alternative, its price is set not by its cost but by your switching cost. The levers for bringing prices down were covered in Episode 38, and those levers only really work when you have an alternative too.

3. Dual sourcing comes in several forms

Dual sourcing tends to conjure up an even 50/50 split. In practice the forms are much more varied, and which one fits depends on the product and scale.

For most small and mid-sized importers the starting point we recommend is the warm backup: 70-85% to the main factory and 15-30% to the backup, ordered steadily. Because the backup factory keeps producing, it remembers your specs and standards, and when trouble hits you only need to raise its volume.

Split sourcing only makes sense with large volume. Split a 3,000-unit-a-month product in half and both factories work around their MOQ, so prices rise and your priority falls at both. An SKU split, on the other hand, costs almost nothing. Put white and black at factory A and the colored versions at factory B and you keep two relationships without losing MOQ. If one factory stops, though, those SKUs stop entirely, so check in advance that each factory can make the other's items.

4. Where and how to find a backup factory

You find a backup factory the same way you found the first: 1688 and Alibaba searches (Episode 11), industry trade shows (Episode 17), and visits to industrial clusters (Episode 31). The difference lies in what you filter on.

A factory that does not share the main factory's weak points

People often pick a backup right next door to the main factory. Same cluster, similar skill level, easy logistics. But that way you take on the regional risk together. When power rationing or an environmental inspection hits the area, both factories stop at once.

Check whether the two factories use the same raw material supplier as well. If a special fabric or a particular brand of component comes from one source, both factories stop the moment that supplier stops. You have split into two factories, but at the raw material stage they are still tied to one line.

A factory similar in size to the main one, or slightly smaller

The size of the backup matters more than you might expect. A factory that is too big finds a deal worth 20% of your volume a nuisance and pushes it down its priority list. One that is too small lacks the capacity to absorb all your volume when the main factory stops. The ideal is a factory with the capacity to take your whole annual volume that still finds your 20% worth caring about. How to verify capacity was covered in detail in Episode 98.

First, make sure it is really a manufacturer

It is not rare for a backup candidate to turn out to be a trading company reselling the main factory's products. When you search the same region for the same item, several trading companies often list goods from one factory. In that case, while you believe you have diversified, you are buying goods from the same line with a middleman's margin added. How to tell factories from trading companies is in Episode 50, and business licenses and credit checks are in Episode 62.

5. Vetting so that two factories make the same product

There is something harder than finding a backup factory: getting two factories to make the same product. You sent the same sample and got the same price, yet when the goods arrive the color is subtly different, dimensions differ by 1-2mm, and the packing is different too. Your customers bought the same product and receive a slightly different item each time.

The cause is almost always that the standard lived only in conversations and a sample. The main factory is building to three years of unwritten understandings, and since those understandings were never written down, there is no way to pass them to the backup. That is why spec sheets, drawings, the BOM and a golden sample need to be put in order before you start dual sourcing. How to write them is in Episode 59, and how to lock in quality with a golden sample is in Episode 34.

Stage three, the cross comparison, is the one that gets skipped. People compare the backup sample only with the golden sample and stop there, but the golden sample may be years old and already differ from what the main factory makes today. What your customers actually receive is current production, so put recent goods from both factories side by side.

Be sure to put pre-shipment inspection on the pilot order. The backup factory's first production run is made by different people from the sample, using a different material lot. For inspection scope and AQL levels, see Episode 100.

6. How to keep a backup factory alive

Surprisingly often, a fully vetted backup factory is useless a year later, because no orders were placed. To a factory, a buyer with no orders for a year is effectively not a customer. The sales contact has changed, nobody remembers your specs, and the quotation has expired. When you call in a hurry, the answer is "let's start again from samples."

The way to keep a backup alive is, in the end, steady orders. Concretely, run it like this.

Small volume, but never interrupted

15-30% of total volume is about right for the backup factory. Less than that and your order becomes something the factory can postpone without a second thought; more than that and the main factory's volume discount breaks badly. More important than the ratio is the rhythm. Placing a small order every quarter does far more to keep the line staff and production standards alive than one big order every six months.

Do not tell the backup factory it is the "backup"

The moment you tell a backup factory "you are for emergencies", it treats you as an emergency customer. In practice, framing it as "we plan to grow your volume step by step as our second supplier" works well. If the main factory runs into trouble you really will raise their volume, so it is not a lie either.

Do not hide it from the main factory, but do not broadcast it

Many buyers worry that the main factory will be offended if it learns about the second supplier. Chinese factories know it is normal for buyers to use several suppliers. What matters is how you put it. Not "your prices were too high so we looked elsewhere", but "company policy requires at least two suppliers for core products", and most will accept it. The main factory often tightens up, and lead times and quality control actually improve.

7. What dual sourcing costs, and which products to start with

Dual sourcing is not free. First, a list of the costs involved.

What these costs should be compared against is the loss if an incident happens × the probability it happens. A rough calculation with the KRW 500 million product from earlier:

In this example the cost of dual sourcing is a little lower than the expected loss. Add the effect of blocking price increases and the gap widens. Stopping just 3% a year of the main factory's increases is worth KRW 15 million. Probabilities and losses differ by product, of course, so rather than reuse these numbers, redo the calculation with your own sales and margins.

Running this calculation also shows that not every product needs a second source. There are two tests: how much it hurts when that product is cut off, and how hard it is to move to another factory.

The top right cell is where to start. A large share of sales and hard to move means that when something goes wrong it hurts the longest and the most. The bottom left cell, by contrast, gives almost no reason to spend on dual sourcing. For products tied to certifications that must be redone when the factory changes, such as Korea's KC certification, rate the switching difficulty one level higher.

8. Clauses to put in the contract

Dual sourcing is an operational question, but it is a contract question too. If a few clauses are missing from your contract with the main factory, the moment you split volume the price can jump or the factory can claim a breach. OEM contracts as a whole were covered in Episode 12; here we look only at clauses tied directly to dual sourcing.

Of these, the MOQ flexibility clause has the biggest practical effect. Say the main factory's terms are an MOQ of 3,000 units per order, with a price of USD 1.20 at 3,000 units or more. A buyer who ordered the full 36,000 a year at 3,000 a month moves 25% to the backup, and monthly orders to the main factory drop to 2,250, below MOQ. Naturally the factory raises the price or asks you to consolidate orders.

Switching to an annual cumulative basis solves this. You commit to 27,000 units a year and place individual orders flexibly from a minimum of 1,500. The factory has its annual volume secured, so there is room for it to agree, and the buyer gets to break orders into smaller units.

Sample wording (MOQ flexibility)
The minimum order quantity under this contract shall be a cumulative 27,000 units per contract year. The minimum quantity per individual purchase order shall be 1,500 units, and the unit price of USD 1.20 shall apply where cumulative orders in the contract year reach 27,000 units or more, even if an individual order is below 3,000 units. If the cumulative quantity at the end of the contract year is below 27,000 units, the parties shall agree within 30 days on either settling the price difference for the shortfall or carrying the shortfall over to the next contract year.

There is a reason for the last sentence. If nothing says what happens when the annual commitment is missed, the factory will resist the clause from the start. Setting the shortfall settlement in advance eases the factory's worry and makes the clause easier to get through.

Sample wording (non-exclusivity and advance notice)
This contract is non-exclusive, and the buyer may source products identical or similar to the contract products from third-party suppliers. The supplier shall give written notice at least 60 days before the effective date of any change in unit price, and the existing price shall apply to purchase orders confirmed before such notice. Where the supplier decides to stop production, relocate the factory, or outsource a key process, it shall notify the buyer in writing without delay and no later than 30 days before implementation.

9. Three real cases

These are anonymized versions of patterns we meet often in consultations. Amounts and timelines are illustrative.

Case 1 - One environmental inspection, four months of downtime

The importer brought in metal household goods. It had taken 100% of a surface-coated product from one factory in Zhejiang for four years. One day the factory called: its coating line had failed a local environmental inspection and been shut down.

Upgrading the equipment and getting permission to restart took four months. The importer rushed to find another coating factory, but factories in the same area were going through the same inspection, and vetting samples from a factory in another region took six weeks. In the end the core product was out of stock for ten weeks.

Afterwards the importer set up a second source in Guangdong for the coating process only. It ordered 20% of volume there continuously and matched color standards, and a year later, when a similar situation came around again, it raised the backup factory's volume within three weeks and got through without a stock-out. The same incident happened twice, and what made the difference was the 20% placed in normal times.

Case 2 - One backup quotation cut a price increase to a third

This was a kitchenware importer. The main factory announced a 12% price increase, citing higher material costs. The importer had already been placing 15% of its volume with another factory for a year.

As soon as the increase came in, the importer asked the backup factory for a current quotation on the same spec. It came in 3% above the main factory's existing price. With that number as its basis the importer reopened talks with the main factory, and the increase was settled at 4%.

The importer never once said "we will move if you do not agree." The factory already knew the importer was working with another factory, and that fact alone shifted the reference point of the negotiation. Material prices really had risen, so not dismissing the request but narrowing it to what the evidence supported was what worked.

Case 3 - Dual sourcing that increased returns

This was an apparel accessories importer. To stabilize supply it moved 40% of volume to a second factory. It sent the main factory's product as the sample and asked for "exactly this", and the sample approval went through.

Once sales began, size-related returns rose noticeably. On checking, the two factories had interpreted the pattern differently, and the same size M differed by 2cm in chest width. One sample had matched, but with no size chart or tolerances on paper, each factory produced in its own way.

The importer paused dual sourcing, documented the size chart, tolerances and fabric spec, and got new samples from both factories against the same documents. It learned the expensive way that the premise of dual sourcing is not two factories but one set of standard documents.

10. Common mistakes

These come up again and again in consultations about dual sourcing.

• Vetting a backup factory and then placing no orders for over a year, so it could not be used when needed
• Choosing a backup in the same industrial park as the main factory and taking the regional risk together
• Both factories using the same raw material supplier and stopping together at the material stage
• A backup candidate turning out to be a trading company selling the main factory's goods
• Sending only a sample without spec documents, so color and dimensions differed between the two factories
• Comparing only with the golden sample and not with the main factory's current production
• Splitting a small-volume product 50:50 so prices rose at both factories
• Splitting volume while MOQ stayed per order, so the main factory's price jumped
• No mold release terms in the contract, so volume could not be moved to the backup
• Telling the backup factory it was "for emergencies" and falling down its priority list
• Playing the two factories' quotations off against each other until both relationships soured
• Not arranging certification separately for the backup factory's product on a certified item
• Dual sourcing even small-share products, so only management costs went up
• Seeing signs of a factory's cash trouble and still putting off backup preparations

11. Dual sourcing checklist

When choosing which products to dual source

• Listed each product's share of sales and contribution to profit
• Marked each product's switching difficulty (mold, certification, special process)
• Prioritized products with a large sales share that are hard to switch
• Chose a form for each product (cold, warm, split, SKU split, component)
• Calculated added cost and expected loss with our own numbers

When finding and vetting a backup factory

• Confirmed it is in a different region from the main factory
• Confirmed its key raw material suppliers do not overlap with the main factory's
• Confirmed it is a real manufacturer through the business license and on site
• Confirmed it has the capacity to take our full annual volume
• Documented and handed over spec sheets, drawings, BOM and golden sample
• Compared the backup sample side by side with the main factory's current production
• Put pre-shipment inspection on the pilot order
• Planned certification for the backup factory's product where required

While running dual sourcing

• Placing an order with the backup factory at least once a quarter
• Keeping the backup's share of volume within 15-30%
• Tracking both factories' defect rates and on-time rates in the same sheet
• Added non-exclusivity, annual cumulative MOQ and advance notice clauses to the main factory contract
• Wrote ownership and release conditions for molds and drawings into the contract
• Regularly checking the main factory for signs of cash trouble
• Refreshing the backup quotation at least every six months

Closing - the 20% is not an insurance premium, it is an option

When dual sourcing comes up, a common reaction is "for that money I would rather squeeze the main factory's price further." In a year when nothing happens, that is right. A 3% price difference is real money.

But what dual sourcing buys is not compensation after an incident; it is options in normal times. A buyer with options can meet a price increase notice with evidence in hand. It can tell a factory whose quality is slipping that volume will be reduced and mean it, and when its line is taken it runs another line. A buyer without options has to wait on the factory's goodwill in every one of these situations.

The start does not have to be grand. Pick one top-selling product, put its spec documents in order, and get samples from two or three backup candidates in a different region. Placing 15% of next quarter's volume with one of them is already dual sourcing. At your next contract renewal with the main factory, bring up the annual cumulative MOQ and advance notice clauses together.

GreenFrog Seoul supports the whole process: choosing which products to dual source, finding and auditing backup factories, aligning quality standards between two factories, and negotiating MOQ flexibility and non-exclusivity clauses. If your main factory has already stopped and you urgently need a replacement, we start with emergency sourcing and vetting; if nothing has gone wrong yet, we start with a product analysis and contract review.