A sample on a supplier’s sales desk can be excellent while an ordinary customer order goes through a different warehouse, packing team, or carrier. A pilot must follow the same route your customers will use. It should also test what happens when an order does not follow the happy path. A vendor’s exception process is difficult to judge until someone asks it to resolve a real discrepancy.
Run a small, documented order pilot before sending a supplier routine customer volume. Measure the complete timeline, inspect the item against your offer, and exercise at least one support case. The goal is to find where the agreement and the actual process differ while the exposure is limited.
This is part four of the Dropship Supplier Guide. The earlier articles define requirements, verify supplier and product identity, and compare commercial and fulfillment terms. The AliExpress delivery test is a useful companion for marketplace sourcing.
Design the pilot before placing orders
Choose the SKUs, variants, delivery regions, and order scenarios that represent the business you plan to run. A single easy address near a warehouse may produce a flattering result and tell you little about your main market. Include the most common variant, a variant at risk of confusion, and an address in a normal target zone. If you plan cross-border sales, test that route separately; customs and last-mile handoffs may differ.
Write a pass threshold before results arrive. It can include correct variant, accurate packaging, carrier acceptance within your promised handling window, traceable delivery, no unauthorized invoice or marketing insert, and support response within your internal deadline. A small sample cannot prove a 99% service level. It can reveal clear failures and show whether the supplier’s systems produce the records you need.
Set a pilot budget and a stop rule. For example, if the first two orders arrive with the wrong variant or the supplier cannot identify which warehouse filled them, pause rather than ordering ten more in the hope of a better average. If a failure is isolated and explained, request a corrective action and repeat the affected scenario. Record the result as new evidence, not as an excuse to erase the first result.
Place orders through the ordinary channel
Use the same portal, API, file transfer, or email route you will use after launch. Do not ask a sales representative to hand-carry the test order unless you are specifically testing that manual process. Submit complete address and variant data, note the supplier order ID, and save the acknowledgement. Measure how long acceptance takes and whether the supplier can reject the order with a clear reason.
Track distinct timestamps: submission, acceptance, pick or pack confirmation, label creation, first carrier scan, delivery attempt, and confirmed delivery. A label without a carrier scan is not proof that the parcel moved. Capture tracking screenshots or event exports when they occur, because carrier histories can change or expire.
Inspect the buyer-facing notifications. Does the shipment email name the right product and carrier? Is tracking usable on a phone? Does the package contain a supplier invoice, another store’s branding, or a return address that contradicts your stated policy? These are not cosmetic questions. They affect customer trust and the number of support contacts the store will receive.
The FTC shipment rule guide emphasizes a reasonable basis for shipping representations and specific choices when promised shipment is delayed. Use the pilot’s actual handoff record to set a cautious promise. Delivery claims require evidence for the route you sell, not a supplier’s best-case screenshot.
Inspect each parcel systematically
Photograph the unopened parcel, shipping label, packing materials, item, included accessories, instructions, and any damage. Record package weight and dimensions if shipping charges or product identity depend on them. Compare the item with the specification and images your store intends to publish.
Use a fixed checklist. Did the exact variant arrive? Are material, color, size, and finish within the described range? Does the item work under a realistic test? Are labels and safety information present where required? Is it clean, sealed, and ready to give to a customer? A vague note such as “looks good” is hard to compare with the next lot.
If the product carries a measurable claim, test the relevant measurement or obtain appropriate independent evidence. Do not represent a kitchen item as heat-resistant, a textile as a particular fiber, or a device as compatible with a model solely because the supplier said so. The FTC advertising FAQs explain the need for substantiation of objective claims. A sample can confirm visible facts; it may not be enough to prove a technical or safety claim.
Check consistency across orders. Compare serial or model markings, packaging, component fit, and origin labels. A product that changes between two nominally identical orders needs an explanation. If the supplier uses multiple factories, make that variability part of the approval decision and any agreement about change notice.
Exercise an exception without manufacturing harm
Use a real problem when one occurs. If the parcel is damaged, report it with the evidence a customer could reasonably provide. If an item is wrong, ask for the documented replacement or credit process. Do not falsely claim a defect to force a test; instead ask the supplier to walk through a fictional scenario in writing and test the first genuine issue during the pilot.
Measure the path from first support contact to usable resolution. Record response time, case ID, evidence requested, who approved the remedy, whether the store had to pay again, when the replacement shipped, and when any credit appeared. A quick first reply that leads to ten days of silence is not a quick resolution.
Test escalation. If the ordinary support channel misses its stated response time, use the agreed backup contact. If no one can find the order by your ID, the integration is not ready for customer volume. The AliExpress order-exception guide explains why the store must still handle its buyer even while a supplier dispute is pending.
For a return, ask where it goes and who inspects it. Does the supplier issue a return authorization? Can the customer use a local return address? How are damaged returns documented? If the supplier requires a long international return for a low-value item, decide whether your store will absorb a refund without return and include that cost in the model.
Test the stock and data loop
Record the inventory value you saw before each order and whether the order was accepted. If a product is marked available and then rejected, ask when the stock changed and how the update was sent. Test a listing pause in your own storefront. Can you remove an unsafe or unavailable SKU before another buyer checks out?
Compare the supplier’s SKU, variant, price, and shipping status with your store’s order record. A wrong mapping can produce a perfect supplier shipment of the wrong item. Inspect how canceled orders and partial shipments appear in the export or API. If a supplier sends status only by email, decide who will reconcile it each day and what happens on weekends.
The Merchant Center data guide describes the need for page and feed accuracy. Use the pilot to test that chain from supplier stock to public availability. An item that is sold out but still shown as purchasable creates a buyer problem before it becomes an account diagnostic.
Score evidence, not impressions
Summarize the pilot in an order-level table. Keep raw timestamps and photos attached to the supplier and SKU record.
| Order field | What to record |
|---|---|
| Offer match | Exact variant, included parts, visible claims, packaging |
| Processing | Submission, acceptance, label, first carrier scan |
| Delivery | Carrier, route, arrival, tracking gaps, damage |
| Cost | Quoted and actual supplier charges, surcharge, remedy cost |
| Data | SKU match, stock status, status update timing |
| Support | Case response, evidence burden, credit or replacement time |
Do not compress a failure into a single average score. A 90% pass rate over ten orders may hide one serious safety or wrong-address problem. Distinguish critical failures, operational defects, and minor presentation issues. A critical failure blocks launch until corrected and retested. A minor issue may be documented with an owner and deadline.
Calculate actual contribution for each completed order, including sample freight and any exception cost, but separate one-time pilot expense from repeatable costs. A pilot order shipped by premium service at your request should not be used as the standard cost. Conversely, a free sample provided by the supplier should not conceal normal unit and freight charges.
Decide the next volume step
If the pilot passes, approve a narrow launch: named SKUs, variants, warehouses, regions, order volume, and a review date. Keep a daily order and exception review at first. Set a threshold that pauses new listings if stock accuracy, handoff, or defect rate deteriorates. The threshold should be observable from your records, not dependent on a monthly sales presentation.
If results are mixed, ask for a corrective-action plan with a specific owner and evidence. Repeat the failed route after the fix. If identity, product consistency, or critical safety evidence remains unresolved, reject the supplier for that product. A delay in launch is cheaper than promising an item you cannot identify or deliver.
The pilot produces a practical map of the relationship. The next article turns that map into written terms and backup options so the process does not depend on a helpful salesperson remembering what was promised.