Once the balance is paid your leverage is gone
Quality deposits and retention with Chinese factories - setting the percentage and period, release condition wording, money that never comes back, and how it combines with other safeguards

Hello, this is GreenFrog Seoul.

The container has arrived at Busan. The shipping documents were clean so the balance went out already, and the goods are in your warehouse. Then incoming inspection turns something up. Out of 8,000 pieces, more than 400 have surface finish defects. Rework or scrap, either way it costs money.

You send photos to the factory on WeChat. The reply comes back. "It passed inspection. Could it have happened in transit?" Three days later you send again. "We are looking into it." Two weeks on it has drifted to something like "we will take a bit off your next order."

What is galling here is not the defects themselves. It is that all the money has already crossed over. The same conversation would have been over in three days if the balance were still unpaid; now that it is gone, it has become a request that takes weeks. Whether money is still in your hands changes the entire character of the exchange.

Today is about keeping a little of that money back. In practice it goes by retention, quality deposit, or held-back balance. Payment terms as a whole were covered in Episode 106, so here we dig into one thing only: how long and on what conditions you hold on to the last slice of the balance.

Collect the cases where a defect got resolved well and they share one feature. At the moment of negotiation the buyer still had unpaid money in hand. The amount does not need to be large. Even 5% left outstanding changes how fast a factory replies. At 0%, however obvious the defect, you are relying on the factory's goodwill.

1. Retention exists because of a gap in timing

Standard payment with a Chinese factory is 30% deposit and 70% balance either before shipment or against shipping documents. Under that structure the point at which a buyer sees the real condition of the goods is almost always after the balance has been paid.

Laying it out along a timeline shows why.

The table makes it plain that on the two rows where you are most likely to find a problem, the buyer holds nothing. However thorough a pre-shipment inspection is, it is a sample inspection. AQL sampling pulls something like 1-3% of the total, and while it catches visible things like surface scratches, plating discolouration that shows up weeks later or breakage after repeated use are structurally beyond its reach. The scope and limits of pre-shipment inspection were set out in Episode 100.

Retention fills that blank. Carve 5% out of the 70% balance and you still hold 5% after shipment. That 5% is what keeps the conversation alive through incoming inspection and the early selling period.

2. Why factories dislike retention

Propose a retention and most factories will push back. You need the real reasons to argue against them, so here they are one at a time.

The first is cash flow

Chinese factories, especially small and mid-sized ones, buy the raw material for the next order with the balance from this one. The rotation is tight. On a USD 50,000 order, 5% is USD 2,500, and situations where the absence of that USD 2,500 delays the next fabric purchase genuinely happen. Around the fourth quarter or just before Chinese New Year, when cash gets squeezed, the resistance is far stronger.

The second is the sense that it is not how things are done

In construction and plant work a defect-liability retention is standard and nobody blinks at it. In consumer goods manufacturing for export, though, T/T 30-70 has hardened into the default, so raise retention and the first answer is "none of our customers do that." Plenty of large buyers do trade on retention terms, but if that factory has never dealt with one, it lands as a novel demand.

The third, and the most delicate, is that it reads as a question of trust

Some contacts hear a retention request as "we do not trust your goods." Face is a more practical variable in Chinese business than people expect. Which is why, when you raise retention, how you phrase it decides whether it lands.

Say "we are worried about quality so we will hold money back" and you will almost certainly be refused. Say "our internal policy runs new suppliers on post-inspection settlement for the first three orders" and the acceptance rate jumps. The terms are identical; the difference is that one judges the other party and the other describes your own procedure.

The cards that actually work in negotiation

The second row works well in practice. Lift the deposit a little and the cash flow burden the factory feels genuinely eases. Go to 40% deposit, 55% at shipment, 5% after incoming inspection and the factory receives the same total with only the timing adjusted. What the buyer takes on is the risk of 10 percentage points more deposit; what the buyer gets is 5% still in hand after seeing the goods. With a new supplier it is worth working out whether that is a trade you want.

3. What sets the percentage and the period

Retention terms are set on two axes, percentage and period. They have to be looked at together. Factories will rarely take 10% over 90 days, and 3% over 30 days barely does anything.

On percentage, 5% is a sound starting point. Go below 3% and it becomes an amount the factory will simply write off, which kills its function as leverage; go above 10% and the negotiation stops happening at all. On a USD 50,000 order, 5% is USD 2,500 — the band where a factory will not walk away from the money but does not feel crushed by it either.

The period comes down to what the retention is meant to verify. This is where practice goes wrong most often.

"30 days after shipment" is the term a factory accepts most readily and simultaneously the most useless one. Ocean freight alone from southern China to Busan is typically 12-20 days, and with clearance and inland transport it routinely runs 20-30. Thirty days from shipment means the retention releases the moment the goods land in your warehouse. If a factory agrees to that one without hesitating, you can guess why.

4. Release conditions - "if there are no quality problems" is where disputes start

Plenty of deals get the percentage and period right and then collapse here. The contract reads like this.

The usual wording
5% of the balance shall be retained as a quality deposit and refunded if there are no quality problems.

It looks fine until you take it apart, at which point nothing is actually settled. Who decides there are no quality problems? What is the threshold for a problem? By when must the decision be made? What happens if no decision is made? All four are blank.

So here is how it really goes. The buyer found defects and will not release the retention; the factory says that level is within tolerance and wants the money. With no standard, whoever is louder wins, and the fight drags into the next order.

The four things release conditions must contain

The fourth row is decisive for persuading a factory. Part of why factories resist retention is the worry that the money will never come back, and an automatic release clause removes it. It costs the buyer nothing either — you were going to inspect and notify within the deadline anyway. It is a trade: certainty for the factory, a deadline for you.

Sample wording
5% of the order value (USD 2,500) shall be held as a quality retention and paid within 5 business days where inspection carried out by the Buyer or the Buyer's nominated inspection body within 21 days of arrival at the Buyer's nominated warehouse meets AQL 2.5 (Major) / 4.0 (Minor). Where the Buyer gives no written notice of rejection within that period, the retention shall be deemed automatically released. Where notice of rejection is given, the parties shall agree the remedy — rework, replacement, or price reduction — within 14 days of the notice date, and the balance net of the agreed reduction shall be paid.

It looks long, but all four elements are in there and it carries through to what happens on rejection. Drop the last sentence and it stops at "no payment if rejected," which produces a different deadlock. How to document the quality standard itself was covered in detail in Episode 103, worth reading alongside this when you draft a retention clause, since it makes the standards part easy to fill in.

5. Situations where the retention never comes back

Sometimes the retention is in place and the money still does not do its job. In practice it shows up in three patterns.

Pattern 1 — A request to offset it against the next deposit

The most common one. As the release date approaches the factory says, "wire fees are a waste, just deduct it from the next order's deposit." It sounds reasonable and it does save the fee.

The trouble is when it repeats. Offset, then offset again, and the retention becomes a bookkeeping figure with nothing behind it. Two years later when you wind the relationship up, each side's calculated balance is different. Try to go order by order and the records have diverged from some point onward, and reconciling them is hard.

If you are going to offset, confirm it in writing every single time. One line of email is enough: "The USD 2,500 retention on order 110 has been released and deducted from the USD 15,000 deposit on order 115, with USD 12,500 remitted." Get "confirmed" back from the factory and that is your record.

Pattern 2 — Not returned when the relationship ends

You have decided to change factories and the last order is done. Then the final retention does not come back. From the factory's side there is no next order, so the incentive to return it has evaporated. And there is no next order to offset against either.

Why this position is structurally weak is obvious. Pursuing legal process in China over a few thousand dollars does not pencil out, and the factory knows that. Which is why it can be better not to apply a retention on the final order at all. Instead, strengthen pre-shipment inspection toward a full check and restructure so that the balance goes out after the inspection passes. Whether to tell the factory in advance that this is the last order is a judgement call, but doing so while a retention is outstanding is not something we would recommend.

Pattern 3 — The defect is confirmed but the amount is not

Four hundred defective pieces, and the factory accepts it. Then you split on the compensation figure. The buyer argues USD 4,000 including rework labour and warehouse costs; the factory says USD 1,200 on a manufacturing cost basis. The retention is USD 2,500.

Dig in and simply withhold the retention here and the factory will decline the next order or raise the price. The relationship ends. So it is better to fix the basis for calculating reductions in advance. One line in the contract about whether a defect is valued on unit cost, on selling price, or with actual rework expense added makes most of this fight disappear. Calculating claim amounts and running dispute procedures were covered in Episode 36.

6. How it compares with other safeguards

Retention is not the only answer. Several tools address the same problem and each covers a different stretch. Try to cover everything with one of them and none of it works properly.

The row worth pausing on is the L/C. An L/C reads as a safeguard, but what the bank examines is documents, not goods. If the bill of lading, invoice and packing list match the credit terms, the bank pays. That holds even if the container is full of defective product. What an L/C protects against is the risk of taking payment and not shipping, not quality risk.

Trade Assurance works when you order and pay inside Alibaba. Receive the goods, raise a dispute, and Alibaba arbitrates, so it does extend to quality issues. As volume grows, though, most trade moves off-platform to direct dealing, and at that moment the protection vanishes. How it works and how disputes actually run were covered in Episode 91.

The combinations we recommend in practice

Layer the tools and the gaps get covered. Set it by stage of the relationship and you will not go far wrong.

• New factory, orders 1-3 — pre-shipment inspection mandatory plus 10% retention for 30 days after arrival. The inspection fee stings, but this is the period to spend it.
• Orders 4-10 — drop pre-shipment inspection to once every two or three orders and ease the retention to 5% for 14 days after arrival.
• Stable relationship — 3% for 7 days after arrival, inspection only on new items or large orders. Easing the terms is itself something you can use as a bargaining chip.
• First production off a new mold — regardless of how many orders in you are, apply both a 10% retention and inspection. Initial production runs wobble even at settled suppliers.

7. Three cases from practice

Patterns we meet often in consultations, anonymised. Figures and timelines are illustrative.

Case 1 — 5% turning a month into three days

A household goods importer. On the second order with a new factory they put in a 5% retention (about USD 2,100) for 21 days after arrival. Incoming inspection found handle joint defects at around 3%.

They sent photos and the inspection report, adding "please settle the remedy before the retention is released." The factory's answer came in two days, and on the third day they agreed free replacement of the entire defective quantity, shipped with the next order.

When the same company hit a similar defect at their previous factory, it took six weeks and ended in partial compensation. The difference was not the defect rate but the fact that money not yet handed over was sitting on the negotiating table.

Case 2 — No release conditions, so each side said something different

An electronics accessory business. They had an 8% retention, but the contract language was just "payable after quality confirmation." Incoming inspection passed, then from the second month of sales, returns for charging cable contact failures started coming in.

The buyer held the retention on the basis that quality confirmation was still under way; the factory demanded it on the basis that passing incoming inspection meant confirmation was finished. Since the wording fixed neither the decision-maker nor the deadline, both had a case of sorts. It dragged for four months and ended in a fifty-fifty split, during which new product development stopped.

The real loss here was not the retention but the four months of standing still. Their contracts afterwards carried the line "retention released 14 days after incoming inspection; defects from the same cause arising within 90 days of sales start to be handled as a separate claim." Separating retention from claims stopped the same problem recurring.

Case 3 — Never recovered at the end of the relationship

The most painful version. After three years they decided to change factories over pricing. USD 3,800 of retention on the final order was outstanding, and incoming inspection had passed without issue.

Then the factory stalled on payment. First it was the accounting close, then the contact had resigned, then there were no replies at all. Legal process did not add up at that amount and they never got it back.

Looking back there were signals. Word of their sample requests at a new factory circulated in the industry, so the intention to end the relationship reached the factory before the final order shipped. With future revenue gone, the factory had no reason to return USD 3,800. That last order should have been structured with strengthened pre-shipment inspection instead of a retention, with the balance paid after confirmation that the inspection passed.

8. Common mistakes

What comes up repeatedly in retention consultations.

• Trading on plain T/T 30-70 with no retention clause, leaving no leverage once defects appeared
• Starting the clock at shipment, so the retention released before the goods even arrived
• Writing only "payable if there are no quality problems," leaving the decision-maker and standard blank
• Setting no decision deadline, so the factory came to believe it would never see the money
• No automatic release clause for non-notification, leaving the amount floating on the books
• Not writing the procedure for rejection, so things deadlocked at "no payment if rejected"
• No basis for calculating reductions (unit cost / selling price / actual rework cost), so the parties split on the amount
• Offsetting against the next deposit without written confirmation, so balances drifted apart
• Applying a retention to the final order and never recovering it after the relationship ended
• Setting the retention at 2-3%, low enough that the factory simply wrote it off and it gave no leverage
• Demanding 10% or more over 90 days, so the negotiation broke down entirely
• Proposing it in terms of "we do not trust you," turning it into a question of face
• Skipping pre-shipment inspection on the grounds that a retention was in place, and missing obvious defects
• Withholding the retention over delivery or pricing issues unrelated to quality, and being refused the terms from the next deal onward
• Not fixing an inspection report format, leading to arguments about what the notice actually said
• Assuming an L/C made quality safe too and putting no safeguard in place

9. Retention checklist

When you set the contract and order terms

• The retention percentage and amount are stated as figures
• The clock starts at arrival, with a cap measured from the shipment date
• The decision-maker (buyer or nominated inspection body) is named
• Pass and fail criteria are written down to AQL level
• The decision deadline and the method of notice are set
• Language for automatic release on non-notification within the deadline is included
• The procedure on rejection (rework, replacement, reduction) is written
• The basis for calculating reductions — unit cost, selling price, or actual cost — is chosen
• The payment deadline on release is stated in business days
• The inspection report format is agreed and shared
• The retention terms appear identically in the Chinese-language document

While the relationship runs

• Retention balances by order are managed in a single table
• Every offset has been confirmed in writing
• Release dates are in the calendar and not missed
• Incoming inspection is completed before the release deadline passes
• Where there is no problem, release notice goes out first, within the deadline
• The point at which terms will be eased is planned against the stability of the relationship

When you request release or report a defect

• Quantity inspected, quantity defective and defect rate are set out as figures
• Photos and quantities are broken down by defect type
• It is written to the agreed inspection report format
• Notice was given in writing within the deadline
• The request (rework, replacement, reduction) is presented with an amount
• The basis for the reduction is calculated against the contract standard
• The agreement is confirmed by email and a reply obtained
• The remaining retention balance and payment schedule are stated

Closing - what the 5% you held back is doing

Raise the subject of retention and a common reaction is "what is an amount like that going to do?" Fair enough. USD 2,500 on a USD 50,000 order will not cover a major defect loss.

But covering the loss is not what a retention does. What it does is decide whether, when you raise a defect, the factory sees it as something to handle now or something that can wait. To a factory paid in full, you are a finished transaction; to a factory holding a retention, you are a transaction not yet settled. That difference in perception carries straight through into response times and the quality of the resolution.

When you build the terms, the period and the release conditions matter more than the percentage. A contract with 3% and 14 days from arrival, a decision standard and automatic release works far better than one with 10% and nothing but "payable after quality confirmation." The first leaves the factory mistrustful and the buyer unsure when to pay, uncomfortable on both sides; the second makes it clear who does what and when, so there is nothing to fight about.

If you are already trading without a retention, try proposing 5% for 14 days after arrival on the next order. Raise it alongside a slightly higher deposit and the odds of refusal drop noticeably. If the factory will not take it at all, that is information too. A factory confident its goods will pass incoming inspection does not make much of two weeks.

GreenFrog Seoul supports payment terms design with Chinese factories, drafting retention clause language, negotiating release conditions, and the practical work of notifying defects and settling reductions. Where a dispute over a retention payment has already started, we begin by organising the evidence and mapping settlement scenarios; where the terms are not set yet, we work with you from the contract language stage.