A small reselling business does not begin with finding the cheapest product in China. It begins with a specific customer, a problem worth solving and a unit-economics model that still works after shipping, advertising, returns and quality control.
The basic model is straightforward: identify demand, choose where you will sell, test the product, source a controlled first order and learn from real customers. The difficult part is protecting the assumptions between those steps. A promising sample can change during mass production. A cheap quotation can become expensive after DDP freight and platform fees. A product with strong views can still have weak purchase intent.
1. Validate real demand—do not fall in love with your idea
Useful products solve a functional problem, an emotional problem or both. A squeeze toy may offer a simple form of stress relief. A portable power station can provide backup electricity where access is unreliable. Clothing, food, household goods and transportation businesses all exist because someone repeatedly solves a recognizable need.
The danger is confusing personal enthusiasm with market demand. “I would use this” is not the same as “a reachable group of customers will pay enough for this after comparing alternatives.”
Write a demand statement before contacting suppliers
Complete one sentence:
For [specific customer], this product solves [specific problem] better than [current alternative] because [clear reason to switch].
If the customer, problem or alternative remains vague, supplier quotations will not fix the business model. Validate the assumption through customer conversations, search behavior, competitor reviews, waitlists, community posts or a small advertising test before buying significant inventory.
- Customers already spend money on an imperfect alternative
- Competitor reviews repeat the same unresolved complaint
- A clearly defined group asks where or when it can buy
- The problem occurs often enough to support repeat demand or referrals
- Friends say the idea is “interesting” but will not preorder
- Views or likes arrive without clicks, email signups or purchases
- The product needs a long explanation before the benefit is understood
- The idea depends on having no competitors rather than serving customers better
Demand validation deserves its own guide. For a first sourcing project, the practical rule is simple: spend a small amount to test customer behavior before spending a large amount on tooling or inventory.
2. Choose your sales channel and understand your local moat
Decide where and to whom you will sell before requesting a final production quote. The channel changes acceptable packaging, labeling, fulfillment, advertising and return costs.
| Channel | Potential advantage | Cost or constraint to model |
|---|---|---|
| Local Facebook groups or communities | Direct feedback, trust and low-cost demand testing | Manual selling, inconsistent scale and local delivery |
| eBay or similar marketplace | Existing buyer traffic and quick listing tests | Marketplace fees, price competition, returns and account rules |
| Amazon FBA | Large demand pool and outsourced fulfillment | Referral and fulfillment fees, labeling, storage, returns and inventory performance |
| TikTok Shop or social commerce | Demonstration-led discovery and creator distribution | Creative volume, commissions, advertising and volatile demand |
| Shopify or an independent store | Brand control, customer data and flexible merchandising | You must create traffic, trust, fulfillment and customer support |
Can a Chinese factory simply sell directly and remove you?
Some manufacturers and trading companies do sell directly into overseas markets. That does not eliminate the role of a capable local reseller. Manufacturing and local retail are different operating systems. A factory may understand materials and production but lack the customer knowledge, content, community trust and after-sales capability needed in your market.
Your defensible advantage can come from four places:
- Local interpretation: you understand how customers describe the problem, compare options and use the product.
- Fast customer service: buyers can reach someone in their language, time zone and consumer environment.
- Flexible offer design: bundles, instructions, packaging and warranties reflect a local niche rather than a generic export catalog.
- Community and brand trust: customers buy from a source that consistently answers questions and resolves problems.
Do not try to beat a factory only on the unit price. Build the layer of value that a distant production line is poorly positioned to deliver.
3. Master the math: unit economics and total landed cost
Start with the expected selling price, then work backwards. The maximum affordable landed cost depends on the channel fee, customer-acquisition cost, fulfillment, returns and the contribution margin required to keep operating.
A factory’s EXW or FOB quotation is not the final number. A usable model should include:
- Product price and tooling
- Samples and sample freight
- Domestic transport and consolidation
- Inspection, laboratory testing and corrective work
- Export packing, labels and marketplace preparation
- International freight, insurance, duty and tax
- Marketplace, payment and fulfillment fees
- Advertising customer-acquisition cost
- Expected returns, replacements and damaged inventory
DDP can give a beginner a convenient door-to-door quotation, but the quote still needs a written description of destination, chargeable weight or volume, customs assumptions, duties and taxes, delivery conditions and excluded surcharges. Compare like with like rather than treating “DDP” as a complete price specification.
Off-the-shelf product or custom mold?
For an unproven first product, an existing design or public mold usually reduces tooling exposure and shortens the test cycle. A private mold can create meaningful differentiation, but it adds design validation, tooling ownership, trial runs, revision risk and a larger capital commitment. Customize packaging, color, bundle or instructions before assuming that custom tooling is necessary.
The “off-the-grid” factory advantage: why the best fit may not rank on Alibaba
Alibaba, Global Sources and similar platforms are useful discovery tools, but their rankings reflect marketing activity as well as manufacturing capability. A polished listing may belong to an excellent factory, an export-focused trading company or a manufacturer carrying the overhead of multilingual sales teams and paid platform promotion.
Other capable manufacturers operate primarily through domestic supply chains and referrals. In Guangdong—and in emerging manufacturing regions such as Vietnam—some are located in industrial towns well outside city centers. Lower land and operating costs can support competitive production, but distance alone is not evidence of quality or a low price. Remote factories may also bring higher transport, communication and oversight costs.
What an in-person meeting can change
Chinese supplier relationships often develop more effectively face to face than through a stream of anonymous online inquiries. Walking the production floor and discussing the order with an owner or production decision-maker can clarify whether a small batch is commercially serious, which existing materials can reduce the MOQ and where the factory has spare capacity.
A conversation over tea is not a magic discount mechanism, and a local representative should never invent leverage or promise priority. Its practical value is access: questions can move beyond a salesperson’s standard quotation to tooling, line scheduling, payment timing, defect responsibility and the conditions for a paid trial order.
The catch is execution. A remote workshop is only an advantage when someone can verify the legal entity, confirm the actual manufacturing site and return during production if necessary. Otherwise the travel, logistics and communication friction can consume the price difference. This is where supplier verification, an on-site factory audit and local sourcing support through Chinese domestic channels serve different purposes.
4. Vet samples and lock the “golden sample”
A good sample proves that one acceptable unit can be made. It does not prove that hundreds or thousands of units will match it. Sample rooms may use more experienced workers, selected materials or slower processes than a normal production line.
Approve a physical golden sample and convert its important characteristics into a bilingual specification sheet. Depending on the product, record:
- Material, grade, composition and finish
- Dimensions, weight and tolerances
- Color reference and acceptable variation
- Functions, test method and pass criteria
- Logo, artwork, barcode and labeling position
- Individual packaging, carton quantity and shipping marks
- Critical, major and minor defect examples
Keep one approved sample under buyer control and agree which counterpart the factory will retain as part of your broader quality control system. Photographs are useful records, but they do not replace measurable specifications.
Confirm which safety, labeling, testing, registration or documentation rules apply in the destination market. Requirements vary by product and country. Do not rely on a supplier’s statement that an item is “export quality” or that a certificate from a similar model covers yours.
5. Control production, quality and fulfillment
Negotiate a trial order around risk—not emotion
A factory is more likely to accept a smaller first batch when the request is operationally credible. Use an existing material, color or production setup; accept a rational unit-price difference; pay for necessary setup; and show the reorder condition if the trial sells. “Please lower the MOQ because I am a startup” is weaker than a defined paid trial with an approval schedule.
Inspect before the balance payment and export
Mass production can diverge from the approved sample through material substitution, tolerance drift, inconsistent assembly, rushed packing or subcontracting. A pre-shipment inspection should take place when production is sufficiently complete and packed goods are available.
AQL 2.5 may be used as a major-defect threshold in some general inspection plans, but it is not a universal quality promise. The sampling level, AQLs, critical-defect rule and product-specific tests must be agreed before inspection. The purpose is to produce evidence for a ship, hold or rework decision while the goods are still in China—not to discover defects through overseas customer returns.
Choose the fulfillment model deliberately
Your first shipment might use air express, consolidated air freight, less-than-container-load ocean freight or another route suited to the value, size and urgency. Some factories offer one-piece dropshipping, but confirm whose branding appears on the parcel, how inventory is synchronized, who handles returns and whether the factory can meet the delivery expectations of your platform.
For sellers using an Amazon FBA sourcing solution, define carton, barcode, labeling and appointment responsibilities upfront. For several suppliers, receiving and sample or shipment consolidation can reduce fragmented parcels and create one controlled handoff. Freight forwarders should quote against a clear cargo description, origin, destination, Incoterm and service level.
6. Launch, support customers and iterate
A product does not create its own distribution. Organic content can reduce paid acquisition over time, but it still requires useful demonstrations, consistent publishing and an offer that converts. Paid advertising can create faster feedback, but the customer-acquisition cost must be included in the unit model from the beginning.
Start with a controlled quantity and treat the first batch as a learning system:
- Acquire seed customers from the most relevant channel rather than every channel.
- Track why people buy, hesitate, return or request support.
- Separate product defects from unclear instructions, positioning or fulfillment problems.
- Feed verified customer complaints into the next bilingual specification and inspection checklist.
- Reorder only after the updated landed-cost and demand assumptions still work.
Customer service is not only a cost. It is the information advantage that helps a local reseller improve the product faster than a remote factory selling a generic version to everyone.
7. Put local execution behind your first order
A sourcing agent should not choose the business for you. The buyer owns demand validation, sales-channel strategy and the final commercial decision. A capable China-side team can reduce the execution gap after you have a credible product brief: compare supplier quotations, verify company identity, coordinate samples, visit a production site, inspect finished goods and consolidate an approved shipment.