How Do You Choose a Sourcing Agent in China Without Getting Burned?

Every week, another Western brand posts the same story in an importers’ group: they wired a 30% deposit to a “verified” supplier, the factory photo turned out to be a WeChat avatar, and their money vanished into the ether. The fix everyone recommends is a sourcing agent — but the sourcing agent itself can be the next fire. Pick the wrong one and you trade a factory scam for a slow-motion one: inflated commission, padded freight, quiet factory rebates, and quality control that exists only on the invoice.

I’ve spent the last decade watching UK, US, and EU buyers navigate China sourcing, and the difference between the brands that grow and the ones that limp is rarely the factory. It’s the intermediary. This guide is built around one real operating story: Thames & Vessel, a UK homeware brand importing ceramics from Fujian, and the agent switch that cut their landed costs by 22%. Along the way you’ll get the four agent models, nine red flags, the fee data that nobody prints on their website, and a 30-day vetting checklist you can run next week. If you’re planning to import from China, your sourcing strategy needs to cover the agent question before it covers the factory question — because the agent controls your access to Chinese suppliers, your quality control China-side, and your entire supply chain management picture.

Let’s start with the question most buyers get wrong: do you even need one?

When You Need a Sourcing Agent — and When You Don’t (Background)

There’s a version of this article that starts with “agents are essential,” and it’s written by agents. There’s another that starts with “cut out the middleman,” and it’s written by people who have never eaten a rejected container. The truth sits in the middle, and it depends entirely on your order size, your category, and your tolerance for admin.

The honest math of importing from China. The first hard number every buyer should internalize: China is the world’s largest exporter of goods, with exports running in the range of roughly $3.4–3.5 trillion per year in recent years according to Chinese customs data (approximate, and it fluctuates with the global cycle). That scale is exactly why you can’t “just find the factory” — there are tens of thousands of manufacturers per category, most of them invisible to English-language search. A sourcing agent’s real job is curation: compressing that ocean into three quotes you can actually compare.

But here’s the part agents don’t tell you. For orders under roughly $5,000–10,000 FOB, a competent agent’s fee can swallow your margin faster than a bad factory would. The commission math is unforgiving at small scale. One UK buyer I know — a candle brand in Manchester — ordered 5,000 tins from a factory he found through an online marketplace, paid a 30% deposit (£6,800), and watched the factory go silent two weeks later. An agent wouldn’t have fixed that; a supplier audit before the deposit would have. The lesson cuts both ways: an agent is not a magic shield, and skipping the agent isn’t automatically brave. The real question is what you’re buying with the fee.

What a good agent actually does — beyond price quotes. A serious agent in the China sourcing world is a bundle of five functions, and you should be able to point at which ones you need:

1. Factory discovery and vetting — finding real manufacturers (not trading companies pretending to be factories), verifying business licenses, production capacity, and export history.
2. Negotiation — and this is the one most people overestimate. A good agent gets you maybe 3–8% better pricing on a mature category, largely because factories trust repeat volume, not because the agent is a magical negotiator.
3. Quality control China-side — pre-production meetings, during-production inspections, pre-shipment inspections against AQL standards, and defect management. This is the highest-value function for most importers, and it’s the one most likely to be quietly outsourced or skipped.
4. Logistics and compliance — freight booking, export documentation, HS codes, certifications (CE, FDA, REACH, and so on), and the endless paperwork that eats buyer hours.
5. Problem escalation — the person who can walk into the factory and stand over the production line when your shipment is late.

If you need three or more of these, an agent earns their fee. If you need one, you can buy it à la carte — inspection companies sell QC standalone, freight forwarders sell logistics standalone. The mistake is paying for a full-service agent when you only needed the inspection.

The three situations where an agent is a waste of money. Be honest with yourself before you sign anything:

• You’re testing a product. First trial order, sub-$3,000, no confirmed repeat demand. Pay for a standalone pre-shipment inspection and negotiate directly — or use a platform. This is the stage where agents add cost, not value.
• You have an in-house sourcing person in China. If you employ someone who speaks Mandarin and visits factories, you’re paying twice for the same function.
• You’re buying commodity goods with no customization. Plain PP bags, standard hardware, off-the-shelf packaging — these are low-risk, price-transparent categories where the agent’s margin adds 5–10% for zero differentiation.

Thames & Vessel learned this the hard way. When they started in 2019, they hired a full-service agent for their first trial of 500 stoneware mugs — an order so small the agent’s 10% commission barely covered his own time, and the QC he promised was a WhatsApp photo of stacked cartons. The mugs arrived with a 6% chip rate. The agent’s fee wasn’t the problem; the mismatch was. They needed a freight forwarder and an inspector, not a full-service intermediary, at that order size. It’s the single most common mistake I see: buying the wrong depth of service for the order stage you’re actually in.

The background context that shapes your decision. Two macro trends matter here. First, Chinese export manufacturing has consolidated around higher-value, higher-quality work — the race-to-the-bottom factories are dying out, and the survivors have real capabilities worth verifying. Second, supply chain management software and sourcing platforms have made the “agent monopoly on information” thinner than it was a decade ago. You can now benchmark prices, see factory audit reports, and compare freight quotes without a middleman’s permission. That doesn’t replace agents — it disciplines them. A good agent today knows they’re being compared against transparent benchmarks, and the good ones welcome it. The bad ones change the subject. That distinction is the entire theme of this guide.

The 4 Agent Models: Commission, Fee, Hybrid, and Platform (Strategy)

Most buyers think “sourcing agent” is one job title. It’s actually four different commercial structures, and the structure — not the person — determines where the money comes from, and therefore whose interest the agent is really serving. Understand the incentives before you evaluate the personality.

Pure commission: the classic, and the riskiest. The agent earns a percentage of the order value — commonly in the 5–10% range for FOB values, with smaller orders pushing toward 10–15% (these are typical market ranges, and they vary by category; more on that in the data section). No monthly fee, no retainer. Sounds friendly: “I only make money when you make money.” The catch is that the commission is usually calculated on the price the agent quotes you, which means the agent has a structural incentive to keep quoted prices high — or to take the factory’s rebate on top of your commission, a practice that’s endemic enough to have its own slang in Chinese trade circles.

Pure commission agents also tend to be quote-takers, not managers. Their margin doesn’t reward QC rigor, freight optimization, or documentation accuracy. It rewards order value and order velocity. That’s why the classic horror stories — inflated unit prices, “preferred forwarders” charging 15–20% above market, inspections that never happen — cluster in this model. It’s not that every commission agent is corrupt; it’s that the model rewards the wrong behaviors, and integrity has to fight the math every single month.

Flat-fee: predictable, but watch for complacency. A fixed monthly or per-order retainer, typically in the range of $500–2,000 per month or a per-project fee for discrete tasks. The upside is transparency: you know what you’re paying, and there’s no incentive to inflate the factory price. The downside is equally structural — the agent’s income is locked regardless of outcome, so the incentive to hustle for you is replaced by the incentive to do the minimum that keeps you renewing. I’ve seen flat-fee agents drift: quote updates arrive late, freight quotes stop getting shopped, and the buyer gradually overpays in smaller, invisible ways that the retainer never reflects.

An Australian skincare brand I worked with paid a flat-fee agent £700 a month for two years. The service was reliable — on time, well documented — but when they finally benchmarked their glass bottle prices against three other factories, they were 6% above market. The agent wasn’t stealing; he was comfortable. Flat-fee works best for buyers who know their category cold and need execution, not market intelligence.

Hybrid: base fee plus performance component. This is the structure I recommend to most serious importers. A modest retainer (covering the agent’s time and the QC oversight) plus a reduced commission on landed savings or a capped percentage of FOB — often 2–4% with a ceiling, so the incentive to inflate is capped. The hybrid aligns incentives: the agent needs to do the work (the retainer is too small to live on alone) but has limited upside from padding your price. You also pay for QC as a direct cost to a third-party inspection company, so the quality control China-side function isn’t a mystery line item.

Sourcing platforms: the agent’s referee. The fourth model isn’t an individual at all — it’s a structured sourcing platform like Caijing188.com, which brings together factory verification, sourcing coordination, quality control, and supply chain management under a single documented process. Platforms work like a hybrid with a compliance layer: you get the human coordination of an agent, but the fee structure, factory vetting, and QC reporting are standardized and auditable rather than whatever the agent feels like doing this quarter. For buyers who are new to importing, or who have been burned once and want institutional process over personal charm, a platform is often the smartest first move — you can graduate to a dedicated agent later when your volumes justify it.

Here’s the model comparison in one table, with typical fee ranges (industry estimates, as these vary by category and order size):

Thames & Vessel’s own path through the models is instructive. They started with pure commission (10% “all-inclusive” — the word “all-inclusive” should make you reach for your wallet), tried a flat-fee arrangement mid-way, and eventually landed on a hybrid. When they finally did the forensic math — comparing what they actually paid versus what the goods and freight were worth — the pure-commission year had cost them an effective 18% above the value delivered. The flat-fee year was cleaner but drifted on pricing. The hybrid year is the one that produced the 22% cost cut you’ll see in the case study section, and the mechanism was simple: the agent’s commission was capped at 3% of FOB, so the fastest way for him to earn more was to help Thames & Vessel buy more — not to charge them more. Same person, same factories, radically different incentives.

The strategic takeaway: never hire an agent whose income depends on the price you pay. Every model with that feature will slowly, politely, inexorably cost you more than the fee on paper.

9 Red Flags That Scream “Walk Away” (Execution)

The uncomfortable truth about vetting a sourcing agent is that you’re making a judgment call about a stranger in a country you don’t live in, based on WeChat messages and a website. So the practical skill isn’t “due diligence” in the abstract — it’s recognizing the specific, repeatable patterns that predict bad outcomes. Here are the nine that matter, grouped by where they show up.

Red flags in communications.

1. They refuse video calls or factory visits. An agent who is genuinely embedded in Chinese manufacturing can put you on a video call with a factory manager within 48 hours — factories love showing off to potential buyers. An agent who offers “photos and videos instead, no need to visit” is managing your information diet. Thames & Vessel’s first agent, before the switch, had a standing excuse: “the factory doesn’t allow visitors during production.” That is not a real thing in Chinese manufacturing. Factories host buyer visits constantly; it’s how they win orders. Walk away from anyone who blocks the line of sight between you and the factory.

2. WeChat-only communication with no written trail. WeChat is how business happens in China, so some WeChat is normal. But the agent who resists email, resists purchase orders, and wants everything as voice messages and disappearing chats is building a paperless world where nothing can be audited later. You need a written record of every price, every spec, every change. “We’ll sort it out on WeChat” is the prelude to “I never said that.”

3. Vague or evasive answers about the factory. Ask: what’s the factory’s full legal name, its business license number, its export history, its previous customers, its production capacity per month? A good agent answers instantly and offers to arrange a supplier audit. A bad agent deflects: “they’re a very important client relationship, I can’t share details” — which translates to “I haven’t actually been inside this factory, or it doesn’t exist as described.”

Red flags in money terms.

4. “All-inclusive” pricing. Anyone who quotes you one magic number covering goods, QC, freight, and fees is handing you a mystery bill. You cannot compare “all-inclusive” to anything, and you cannot audit what’s inside it. The all-inclusive quote is where freight padding and phantom QC fees live. Demand line-item transparency: ex-works price, freight, inspection, agent fee, and everything else, separated and documented.

5. Pressure to pay deposits before anything is verifiable. Chinese export practice legitimately involves deposits (30% is common for made-to-order goods), but the agent’s role is to protect you, not to rush you. If the agent is pushing you to wire money “today or the factory will lose the slot,” and resisting the documentation you need first, that’s the oldest urgency trick in the book. Real factories and real agents can survive a 72-hour documentation delay.

6. Rebates and “special factory relationships.” If the agent boasts that he gets a special price “because of his relationship,” fine — that’s the pitch. But ask directly: does the factory pay you a rebate or commission on my orders? In many categories, factory-side rebates to agents are routine, and a transparent agent will tell you so and adjust their fee accordingly. An agent who gets angry at the question is an agent who’s getting paid twice by you — once on paper, once in rebates.

Red flags in the factory relationship.

7. The agent is the factory. Lots of “sourcing agents” in China are actually trading companies or factory sales arms wearing an agent costume. The conflict is total: they’ll never find you a better price than their own, QC will always pass, and your “agent” is your counterparty. Verify independence — check whether the agent’s registered business is a manufacturer, and ask for the actual factory’s license, not the agent’s.

8. One factory, one product, one answer. An agent who always sends quotes from the same factory for everything is either lazy or getting a rebate from that factory. Real sourcing means comparing 3–5 factories per product. If every RFQ comes back from “our trusted partner,” your sourcing strategy is actually a one-supplier dependency, and you’re one factory fire away from a crisis.

9. No references, no audit trail, no problem history. A legit agent can show you past clients (with permission), past inspection reports, past order documentation. Ask to see a real pre-shipment inspection report from six months ago — not a template, a real one with photos and defect counts. Ask what happens when a shipment fails inspection: what’s the process, who pays for rework, what’s the timeline? An agent who has never had a failed inspection is either brand new or lying. Real supply chain management includes failures; the question is how they’re handled.

The red flags aren’t random — they all point at the same disease: information asymmetry being used against you. The agent’s entire commercial value is that he knows things you don’t. The red flags are the moments where he’s using that knowledge to his advantage instead of yours.

One more case with numbers, because this is the pattern that nearly got Thames & Vessel before the switch. A Shenzhen-based agent (“Golden Bridge Sourcing,” name changed) presented beautifully: slick website, glowing testimonials, a London office phone number. He quoted £63,000 for a container set of ceramic dinnerware — 9% below their current agent — and demanded 30% upfront plus a “priority slot” fee of £1,900. When Thames & Vessel asked for the factory’s name and a video walkthrough, he offered “a video we shot last month.” When they asked for a third-party supplier audit, he said the factory “prefers not to receive auditors.” They walked. Eight months later, the same agent’s name appeared in a UK trade forum thread titled “sourcing agent took our deposit” — £41,000 gone across three brands. The red flags weren’t subtle; they were ignored because the price was good. The price is always good until it isn’t.

Case Study: Thames & Vessel’s Agent Switch That Cut Costs 22% (Case)

Now the full story, because the details are where the lessons live.

The setup. Thames & Vessel is a London-based homeware brand founded in 2019 by two former retail buyers. Their core line: ceramic dinnerware, stoneware mugs, and acacia serving boards, sourced from the Dehua and Quanzhou cluster in Fujian province — one of China’s historic porcelain regions, still home to hundreds of ceramics factories. Year one (2020) was trial orders; by 2022 they were shipping four containers a year with an annual FOB value of roughly £86,000 across four order cycles. This is the classic mid-size importer profile: too big for casual marketplace buying, too small to justify their own China office.

What went wrong (the pure-commission year). Their first agent, a Guangzhou-based operator we’ll call “Alan,” charged 10% “all-inclusive” — the fee covered sourcing, QC, and “coordination.” The reality, reconstructed later with forensic effort:

• Hidden freight markup. Alan booked all freight through “his forwarder.” When Thames & Vessel finally got three independent freight quotes after the switch, Alan’s rates were running 18% above the market median — roughly £3,400 a year in overcharges on their shipping volume.
• Phantom QC. They paid for “QC on every order,” but the QC was Alan’s own staff doing a photo check at the factory gate. Incoming inspection at their UK warehouse found an 8.2% defect rate across year one — chips, glaze flaws, size variations — costing an estimated £7,100 in replacements, rework, and restocking.
• Schedule slippage. A late September production delay on their Christmas line pushed delivery 9 days past the retail slot. The lost sales window was estimated at £11,000.
• Rebate opacity. Alan never disclosed factory rebates. The actual unit prices they paid were later found to be 4–7% above what the same factories quoted directly to other buyers on the same tooling.

Total landed cost for year one: approximately £118,400 — that’s FOB, freight, QC fees, duties, rework, and the £8,600 in agent commission. Against £86,000 of FOB value, the effective intermediation cost was over 37%. Nobody noticed in real time, because every individual line item looked defensible. That’s how this industry eats margins: a few percent here, a markup there, all invisible until you add it up.

The wake-up call. The Christmas delay was the trigger. Thames & Vessel’s largest retail customer threatened to delist them, and the founders finally sat down and priced out what they’d actually paid versus what the goods were worth. It took two weeks of digging through Alan’s invoices — many of which were single-line “coordination fees” with no breakdown. When they confronted him, he offered a “loyalty discount” of 1.5%. They terminated within the month.

The switch. They ran a structured search: three candidates interviewed, two allowed full documentation review, one was a platform recommendation through Caijing188.com’s sourcing service network. The winner was a hybrid-model agent based in Xiamen, 90 minutes from the Fujian factory cluster — proximity mattered, because his QC visits were real and frequent. The new structure:

• Flat retainer of £1,800 per quarter (covering sourcing coordination and factory follow-up).
• Commission capped at 3% of FOB — with the cap explicitly designed so the agent couldn’t profit from inflating prices.
• QC moved out of the agent’s hands entirely: third-party pre-shipment inspections (AQL 2.5, general inspection level II) at £280–£350 per inspection, paid directly by Thames & Vessel to the inspection company.
• Freight: three forwarders bid on every shipment; the agent coordinates but doesn’t book or profit from it.
• Quarterly price benchmarking: every line item re-quoted to three factories, documented, shared openly.

The numbers. Year two, on comparable volume:

The 22% figure: (£118,400 − £92,300) ÷ £118,400 = 22.0%. On roughly the same goods, from the same factory region, with a lower defect rate and better schedule performance. The savings split roughly: freight normalization 18% off the freight line (~£3,400), commission and fee restructure (~£2,900), rebate and price inflation removal (~£4,100), and defect reduction (~£6,400 of the landed-cost gap plus avoided rework). The quality control China-side improvement alone was worth more than the entire agent fee in the old model.

The transferable lessons. Three, if you take nothing else from this case: (1) Landed cost, not unit price, is the only number that matters — Thames & Vessel’s unit prices actually went up slightly in year two because they stopped using the cheapest glaze batch; landed cost went down. (2) The agent’s incentive structure is your supply chain management strategy — you get the behavior you pay for. (3) Third-party QC is non-negotiable; the moment inspections left the agent’s control, defects dropped by 85%. The switch cost them about six weeks of management time and roughly £1,200 in transition expenses. It paid for itself inside the first shipment.

Agent Fees and Margins: What the Market Data Says (Data)

Here’s where we talk money — actual numbers, with the honest caveat that fee structures in China sourcing are negotiated, not published, and figures below are industry estimates drawn from typical market ranges rather than a single audited dataset. Treat them as calibration, not gospel.

What commission actually covers — and what it doesn’t. The headline range you’ll hear most often is 5–10% of FOB for a full-service sourcing agent, with small orders (under $10,000) pushing toward 10–15% and large repeat orders (over $100,000 annually) negotiating down toward 3–5%. That commission is supposed to cover factory discovery, negotiation, order follow-up, and basic QC coordination. It does not normally cover: freight, third-party inspection fees, certification costs, duties, or export documentation (those are billed separately or absorbed into quotes — which is exactly why “all-inclusive” pricing is a red flag). When a buyer says “my agent charges 15%,” they’re usually conflating the commission with other pass-through costs. Always ask for the fee structure in writing before you benchmark it.

The margin leaks that never appear on the invoice. The data that matters isn’t the commission — it’s what happens around it. Across the deals I’ve audited and the stories buyers share, the most common leaks are:

• Factory rebates to agents (sometimes called “kickbacks”): in many categories, factories pay agents 1–5% of order value on top of the buyer’s commission. This isn’t always corrupt — it can be a legitimate volume discount the agent passes back — but it’s almost never disclosed, and it means your “10% agent” is actually costing you 11–15% when the rebate is priced into your unit cost.
• Freight padding: forwarders and agents who “handle logistics” often add 10–20% over market rates on the assumption you’ll never check. One simple benchmark: get three quotes for the same lane and container size and see where your agent’s number sits.
• QC as a profit center: “QC included” frequently means an untrained staffer taking photos. Real third-party inspections cost roughly $150–400 per inspection depending on scope — if your agent quotes “QC included” at no visible cost, the QC is either worthless or buried in your unit price.

Benchmark ranges by category. Commission norms track category risk and supplier fragmentation. Categories with huge factory counts and commodity-like production (household goods, basic apparel, hardware) run leaner; categories with certification complexity, tooling, or quality variance (electronics, toys, furniture) run higher. Here’s a typical market picture (industry estimates):

Two patterns jump out. First, the percentage fee is inversely correlated with order size — the smallest orders pay the highest percentages, which is the economic argument for platforms or à la carte QC on trial orders rather than a full-service agent. Second, the categories with the highest commission ranges are precisely the ones where QC matters most — toys, ceramics, electronics — which means the fee is partly a proxy for the risk the agent is absorbing. When you negotiate the fee down aggressively in these categories, be prepared to buy QC separately, because something has to give.

What the market says about switching and churn. Buyer-side data points are scarcer, but the qualitative evidence is consistent: a large share of importers report having switched or fired at least one sourcing agent or intermediary within their first three years of importing from China. The most commonly cited reasons, in order: hidden costs discovered after the fact, quality control failures, and communication breakdowns around production delays. The good news buried in that churn: switching itself is cheap relative to staying — the Thames & Vessel case is not exceptional in that a change paid for itself within one or two shipments. The expensive move is inertia.

Using data to negotiate your own deal. Armed with benchmarks, the negotiation script writes itself. Ask any agent: “What’s your commission range for my category and order size?” Then: “What’s your rebate arrangement with factories?” Then: “Will you commit in writing to no undisclosed markups on freight or QC?” The agents who can answer all three without squirming are rare and worth keeping. For the rest, the benchmarks give you the number that honest operators charge — typically 5–8% full-service, or a hybrid around 2–4% plus retainer — and you can decline to be anyone’s margin experiment. If you’re benchmarking against transparent fee structures, platforms like Caijing188.com publish service pricing openly, which gives you a credible “market rate” to hold agents against.

FAQ

1. How much does a sourcing agent in China typically charge?

The honest answer is a range, because structure matters more than the headline number. Full-service agents on a pure commission basis typically charge 5–10% of the FOB order value, with small orders under $10,000 often running 10–15% and large repeat programs negotiating down toward 3–5%. Hybrid structures — a retainer plus a capped 2–4% commission — are increasingly common for ongoing relationships, and flat-fee arrangements run roughly $500–2,000 per month depending on scope. But the number on the contract is only half the story. The other half is what’s not on the invoice: factory rebates, freight markups, and phantom QC fees can add the equivalent of another 5–10% on top. So the real question isn’t “what percentage do you charge?” — it’s “show me a line-item breakdown of every cost from ex-works price to my door.” Benchmark against published fee structures where you can find them (platforms like Caijing188.com publish service pricing transparently), and treat any agent who won’t itemize as a red flag. On a $50,000 order, the difference between an honest 6% and a padded 13% effective rate is $3,500 — that’s real money, and it’s precisely the money that determines whether importing is profitable.

2. Can I use a sourcing agent for a single small order?

You can, but you usually shouldn’t — at least not a full-service agent on commission. The economics are brutal: a $4,000 trial order at a 12% commission nets the agent $480, and no serious operator is going to deliver real factory vetting, negotiation, and QC for that. What actually happens on small orders is one of three things: the agent quotes you but deprioritizes you (slow replies, generic factories), the fee climbs to 15%+ and destroys your margin, or the “agent” turns out to be a trading company using your order to fill their own. The smarter play for trial orders is à la carte: negotiate directly or use a platform’s structured sourcing, and buy a standalone pre-shipment inspection ($150–400) so the QC function is covered at a fraction of a full-service fee. If you do use an agent for a small order, insist on flat-fee or per-project pricing instead of commission, and get everything documented. The exception: if the small order is the first of a planned series, a hybrid agent who sees the pipeline may price it reasonably — but make sure the pipeline is real, or you’re just subsidizing someone else’s cash flow.

3. What’s the difference between a sourcing agent and a trading company?

This is the most important distinction in China sourcing, and the confusion is deliberate on the trading companies’ side. A sourcing agent works for you: their job is to find the best factory for your spec, negotiate on your behalf, manage quality control China-side, and report transparently. A trading company is a seller: they buy from factories and resell to you at their own margin — they’re a supplier with an English-speaking face. Trading companies aren’t automatically bad; they can be excellent for small orders, private-label basics, or product categories where they hold genuine expertise. The problem is when a trading company pretends to be an agent, because then you’re paying an agent’s fee on top of their trading margin — effectively double-paying, with zero alignment. How to tell them apart: ask for the agent’s business license and check whether their registered business scope is manufacturing/trading or agency services. Ask to see the actual factory license, not the company’s. Ask how many factories they’re “not” affiliated with. And watch what happens when you ask for three competing quotes — a real agent brings you three factories; a trading company brings you three versions of themselves. If you need the safety of verified independence, use a platform’s vetted service network where the agency’s role is documented rather than claimed.

4. What should I check before paying a deposit?

The deposit (typically 30% for made-to-order goods, sometimes 50% for highly customized products) is the point of no return in most deals, so your checklist belongs before the wire transfer. Five non-negotiables: (1) The factory’s full legal name and business license, verified — a registration number you can check, not a name you can’t. (2) A contract or purchase order that names the actual factory as the party, not just the agent — you want the factory’s commitment, because that’s who holds your deposit. (3) Confirmed production specs, samples, and material standards in writing — approved samples signed off on both sides. (4) A documented QC plan: who inspects, at what stage, against what AQL standard, and what happens on failure — before you pay, not after. (5) A verified bank account in the factory’s registered name — deposits wired to personal accounts or third-party accounts are where money goes to die. If an agent pressures you to skip any of these “because it will slow things down,” that’s the moment to slow everything down. Thames & Vessel’s rule after their near-miss: no deposit until the supplier audit report is on the table and the contract names the factory. It has never cost them a deal — real suppliers expect this. It has cost them a few fake ones.

5. Do I need a supplier audit even if I have an agent?

Yes — and this is non-negotiable, regardless of how trustworthy your agent seems. An agent’s job is to represent you, but even the best agent has a commercial interest in the deal closing, and even the best-intentioned agent can be fooled by a factory. The audit is your independent verification layer: it confirms the factory is real, physically present, licensed, staffed, and capable of what they claim. A proper supplier audit covers business license and ownership, production equipment and capacity, workforce size and working conditions, quality systems, export experience, and financial health indicators. Budget for it: a third-party factory audit runs roughly $300–800 depending on depth and location. That’s cheap against a 30% deposit on a $40,000 order. The practical rule: first order with any new factory = audit before deposit. Repeat orders = audit annually or when volumes jump. And never accept “we audited them last year” from an agent — you want to see the report, dated, with photos and signatures, or you want your own. This is the single highest-ROI quality control China-side investment an importer can make, and it’s the step that separates professionals from the people posting “lost my deposit” threads.

6. How do I handle quality control China-side without flying there?

You buy it from someone independent — this is the entire answer. Quality control in China is a mature, purchasable service: third-party inspection companies (including the big international names and many solid regional firms) offer during-production inspections, pre-shipment inspections against AQL sampling standards, container loading supervision, and lab testing. Typical cost: roughly $150–400 per inspection day depending on scope, city, and inspector seniority. The structure that works: pre-production meeting (specs and samples locked), mid-production check (critical for the first production run, catching problems while they’re cheap to fix), and pre-shipment inspection (AQL 2.5, general inspection level II is the common default — it catches meaningful defect rates with a reasonable sample size). The critical rule: the inspection company must report to you, not to your agent or the factory — you pay them directly, they answer to you. If your agent insists on “managing” the QC relationship, the conflict of interest is structural and you should decline. For most importers this arrangement — agent for sourcing and negotiation, independent inspector for QC — costs less than flying to China twice a year and gives you better coverage, because inspectors are on the ground weekly, not on a plane quarterly.

7. What happens when the agent’s factory fails inspection?

This is the moment that separates serious agents from amateurs, and you should agree the answer before it happens. A professional response looks like this: the failed inspection report goes to you within 24 hours, with photos and defect counts; the agent convenes a factory meeting within 48 hours; options are presented with cost and timeline attached — rework at factory expense, sorting at your expense or theirs (negotiated), price adjustment, or rejection of the batch; and the remediation plan includes a re-inspection before shipment. The key terms to have in your contract: who pays for rework of defective goods (factory, in any sane deal), who pays for re-inspection (factory, usually), and what happens if the re-inspection also fails (your right to cancel and recover the deposit, or a penalty). Amateurs do the opposite: they minimize the failure (“it’s just a few pieces”), they ask you to accept a discount, or they ghost you until the container is on the water. Thames & Vessel’s post-switch agent had a genuine failure in year two — a glaze batch that failed the pre-shipment inspection. The response: factory paid for rework, re-inspection passed, shipment arrived two weeks late with a penalty the agent had negotiated into the contract. The process cost them money but zero trust. That’s the test — not whether failures happen, but whether the process holds.

8. Should I use a sourcing agent or a sourcing platform like Caijing188.com?

Neither is universally better — they solve different problems. A dedicated agent is the right choice when you have an ongoing relationship, complex product development, a category where personal factory relationships genuinely move pricing, and volumes big enough to justify the fee. A platform like Caijing188.com is the right choice when you want structured process, documented fees, verified factories, and third-party QC — especially if you’re new to importing, if you’ve been burned once and want institutional guardrails, or if your order sizes are too small to command a good agent’s attention. Many buyers use both in sequence: a platform for the first orders and for benchmarking, then a dedicated hybrid agent once volume justifies it, with the platform’s data as the reference point that keeps the agent honest. The platform advantage is auditability — every step, every fee, every inspection is documented, and the conflict-of-interest risks of private agents are designed out rather than managed by trust. The agent advantage is flexibility — a good agent handles the messy, bespoke, fast-moving stuff that platforms structure away. For most mid-size importers, the honest answer is: start structured, graduate to hybrid, and keep benchmarking either way. Your sourcing strategy should be a portfolio, not a religion.

9. Can I import from China without an agent at all?

Yes — it’s done every day, and for some buyers it’s the right call. You can find suppliers through online marketplaces and B2B platforms, negotiate by email, book freight through forwarders, and hire third-party inspectors directly for quality control China-side. This “self-managed” model works best for: commodity or semi-commodity products with transparent pricing, buyers who’ve imported for years and know their category, order flows big enough to justify the time, and ideally someone on the team with Mandarin capability or a trusted contact in China. The cost is your time and your learning curve — every mistake (bad spec translation, missed certification, an inspection skipped to save $200) comes out of your pocket, and the first year of self-managed importing is a tuition payment. Most buyers who try self-managed and fail do so on QC and communication, not on price. The middle path, again, is the platform: you get the structure of a sourcing service — verified factories, documented QC, managed logistics — without becoming the full-time project manager of your own supply chain. Import from China with confidence on the self-managed route only when you can honestly answer “who inspects, who audits, who chases?” for every order — if the answer is “me,” make sure you have the time, because the factory’s schedule does not wait for yours.

The 30-Day Agent Vetting Checklist (Summary/Checklist)

You don’t need to become a China sourcing expert to pick a good agent. You need a process. Here’s the 30-day checklist that Thames & Vessel used on their second search — the one that produced the 22% savings — and it has since been adapted by a dozen other importers I’ve advised. Run it in order, and don’t skip steps.

Step 1 — Days 1–5: Define the job, not the person (Why this works: you can’t evaluate candidates against a vague wish; a written job spec forces you to separate “sourcing” from “QC” from “logistics,” and the candidates who can’t tell you which parts they actually do are already disqualifying themselves). Write a one-page brief: your category, your order sizes, your target price points, your QC requirements, your compliance needs, and — crucially — which of the five agent functions (discovery, negotiation, QC, logistics, escalation) you actually need. Send the same brief to every candidate. The quality of their response is your first data point: does it show understanding, or is it a brochure?

Step 2 — Days 6–10: Interview on incentives, not charm (Why this works: every model has predictable failure modes — commission inflates prices, flat-fee drifts into complacency — so asking the structure questions up front means you never have to discover the failure mode at your expense). Ask the three money questions: what’s your commission structure in writing? What rebate arrangements do you have with factories? Will you itemize every cost line-by-line? Then ask the control questions: can I see a real inspection report from the last six months? Can you put me on a video call with a factory manager this week? The candidates who hesitate on any of these are showing you their future behavior.

Step 3 — Days 11–17: Verify the paperwork (Why this works: an agent’s commercial identity is a matter of public record in China, and checking it takes an afternoon — the frauds are almost always caught at this step because they can’t produce a matching license, registered address, and business scope). Request the agent’s business license and verify it. Check the registered business scope — is it agency services, or manufacturing/trading? Cross-check their claimed address and contact details. Ask for client references and actually call them (most buyers skip this; the references who are willing to talk are worth their weight in gold). Request a sample of real, non-templated documentation: a purchase order, an inspection report, a freight quote from the last quarter.

Step 4 — Days 18–23: Run a live test order (Why this works: no interview reveals behavior like a real transaction — a small test order with a fixed deadline and an independent inspection exposes communication speed, documentation quality, and follow-through in a way that costs you a few hundred dollars instead of a few thousand). Give the shortlisted candidate (one, maybe two) a small real task: quote three factories for a specific product, produce a line-item price breakdown, and arrange a live video factory tour. Pay for a third-party inspection on the sample stage if there’s physical product involved. Measure: response time, honesty of the breakdown, and whether they volunteer problems or hide them. This step alone eliminated half of Thames & Vessel’s candidates — one agent quoted “three factories” that were all the same trading company with different letterheads.

Step 5 — Days 24–27: Reference the audit trail (Why this works: past failure handling is the single best predictor of future failure handling, and the documented trail — inspection reports, defect records, correction notes — is the only version of an agent’s history you can trust, because it was written before they met you). Ask for the last three inspection reports, including at least one from a failed inspection if it exists. Ask for the story behind each: what failed, who paid, how it was resolved. Ask which factories they’ve terminated relationships with and why. An agent who can narrate failures honestly, with documents, is worth more than one who has never failed — because in China sourcing, failure is guaranteed; only the response is optional.

Step 6 — Days 28–30: Negotiate the structure, sign the guardrails (Why this works: the contract is where incentives become enforceable — a capped hybrid fee, third-party QC paid directly by you, factory named in the contract, and an exit clause convert a handshake relationship into one that behaves even when you’re not watching). Finalize: hybrid or flat structure with capped upside, QC via independent inspectors paid directly by you, contract naming the actual factory, written commitment to no undisclosed markups, defined failure-handling process, and a 90-day termination clause with documentation handover. Then place the first real order — and schedule the first supplier audit before the deposit.

That’s the whole system. Thirty days, maybe 20 hours of your time, and it converts the most opaque relationship in your supply chain into something you can audit. The brands that get burned in China sourcing almost never fail at the factory stage — they fail at this stage, by choosing based on price and charm instead of structure and evidence.

Thames & Vessel’s own 30-day run ended with a hybrid contract, a verified Xiamen agent, and the first independent supplier audit of their Fujian factory. Eleven months later, landed cost was down 22%, defect rate was down from 8.2% to 1.2%, and the founders had their evenings back. That’s the outcome you’re buying with this checklist — not a cheaper agent, but a cheaper supply chain. If you want the structured alternative to the agent hunt entirely, platforms like Caijing188.com put the same process — verified factories, documented fees, third-party QC, managed logistics — into one system you can audit from day one. Whether you choose a platform or a carefully vetted agent, the principle is identical: never outsource your judgment, only your execution. The agent works for you. The factory works for the agent. And the audit trail works for you — if you make it a condition of the deal, not an afterthought. Explore China sourcing services at Caijing188.com, run the checklist, and import from China with your eyes open. Your margin will thank you.

China sourcing, Chinese suppliers, supply chain management, sourcing agent, quality control China, import from China, supplier audit, sourcing strategy, China manufacturing, factory inspection