Shanghai International Logistics Co., Ltd.
International Logistics: A Practical Guide To Global Supply Chains
International logistics is the set of coordinated processes that carries a physical good from a supplier in one country to a customer, plant, or project site in another, and makes the paperwork, cash, and compliance records match. The movement is the visible part and often the least difficult. The difficulty is that a cross-border shipment is a chain of information handoffs, each performed by a party with a different view of the data, and any of them can stop the cargo.
Planning across borders starts from a different baseline than domestic planning, because the horizon is longer and less predictable. A replenishment cycle must cover production at origin, an export window that may be weekly, main carriage of several weeks, import clearance, and inland delivery, so orders are placed weeks before the goods exist. Two variables dominate: lead time variability, from congestion, roll decisions, customs examination, and equipment imbalance at origin; and demand variability at destination, partly inherited, since an upstream failure becomes a downstream surge.
The reorder point is not a fixed number inherited from last year. It follows from expected demand over the lead time, the variability of that lead time, and the service level intended. When lead time is highly variable, use a conservative percentile of transit time rather than an average, since planning on the average understates the buffer precisely when it is needed most. At origin, make the commitment explicit: a confirmed production date and a goods-ready date.
Service level has two components that are usually treated separately and should not be. The commercial promise is the delivery date given to the customer. The operational service level is the probability that the shipment arrives within that promise, given the transit, clearance, and exception profile of the lane. A carrier commitment describes the movement of cargo, not the delivery.
Routing should be decided on total cost to deliver the order rather than on freight rate, which covers freight, insurance, handling, brokerage, duty and import taxes, demurrage exposure, inventory carrying cost, and the internal cost of exceptions. A faster mode can be cheaper overall when it removes a week of pipeline inventory. A routing plan should also name a primary route, an alternative, and the trigger that authorises a switch.
Industrial goods are not a single profile. Small high-value spares behave differently from a low-value heavy project consignment. The table is a general reference; transit characteristics depend on the lane, carrier, and season.
| Mode | Typical industrial use | Cost profile | Transit characteristics |
|---|---|---|---|
| Sea, full container | Equipment, spares, volume goods | Lowest per unit | Weeks; varies by lane |
| Sea, LCL and breakbulk | Small lots, irregular pieces | Variable; handling-heavy | Weeks, extra handling and dwell |
| Air | Spares, samples, urgent parts | Highest per kilo | Fast, but clearance eats the gain |
| Road and intermodal rail | Regional distribution, feeder legs | Low on dense corridors | Reliable on short and mid distances |
| Project and heavy lift | Oversized, out-of-gauge cargo | Survey driven | Permits govern |
Documents are the executable form of a shipment. Most customs holds and freight claims are documentation problems rather than physical problems, because clearance decisions are made on paper. Three functional groups exist. The commercial layer establishes the transaction: proforma invoice, commercial invoice, packing list, and where required a certificate of origin. The transport layer moves the cargo: bill of lading or sea waybill, air waybill, consignment note, and the verified gross mass declaration. The regulatory layer authorises it: import and export declarations, licences for controlled goods, and party screening.
The most common cause of a hold is an inconsistency between documents rather than a missing one. Values, quantities, weights, country of origin, product description, and classification code must agree across invoice, packing list, transport document, and declaration. A description written loosely for readability is one an officer cannot classify, and an unclassifiable entry becomes a request for information and a delay measured in days. Document control is therefore a data-discipline problem: when the order record is the single source of truth, the mismatches that cause holds largely disappear. File before the cargo is needed, so errors surface while they can still be corrected.
| Document | Purpose | When it is required |
|---|---|---|
| Commercial invoice | What was sold, at what value, between whom | Almost always, for declaration and valuation |
| Packing list | Contents, counts, weights, dimensions | Most shipments; air, road, container cargo |
| Bill of lading or waybill | Carriage contract; bill of lading transfers title | Sea and multimodal moves |
| Certificate of origin | Establishes origin for tariff and preference | Where preferential tariff is claimed |
| Import and export declaration | Statutory entry allowing the border crossing | All commercial movements |
Clearance is a decision made by the customs authority in the destination country, based on the classification, valuation, origin, and admissibility of the goods and on the licence conditions attached to them. The importer of record lodges the entry, pays duty and taxes, keeps records, and answers post-clearance audit; the buyer is not automatically that party, and the difference decides who can control the process. Classification assigns goods to a heading in the Harmonized System, usually at six-digit level; the heading sets the duty rate and which licences apply. Valuation establishes the customs value, generally the price paid or payable. Origin determines preferential tariff treatment.
Incoterms 2020 defines where cost and risk transfer between seller and buyer, and who arranges carriage, insurance, and clearance. Under EXW, FCA, FOB, CFR, and CIF, the seller delivers the goods at the named place ready for export and the buyer arranges main carriage. Under CPT, CIP, DAP, DPU, and DDP, the seller arranges carriage to the named destination. The named place determines which party books the carriage, who files the export declaration, and where the risk of loss sits. Two failures recur: a delivered term used where the seller cannot lawfully act as importer of record, and a term allocated between parties who do not share the same understanding of it.
When lead time is long and variable, the natural response is to hold more stock. Buffering in the wrong place adds cost without adding resilience. Stock held at a foreign supplier ties up cash earlier, is exposed to duty, and sits behind the same customs and port bottlenecks as the shipment itself. Stock in transit is invisible to a system that tracks physical locations only, and it accrues carrying cost without appearing on any inventory report. Buffer sizing should follow reorder-point logic: the quantity at a node depends on demand variability, lead-time variability, and the service level there.
Consolidation is the mirror of buffering. Combining several small consignments into one movement lowers unit freight cost but adds a waiting period at origin and a deconsolidation step at destination. Whether that trade pays depends on value density, fixed cost per movement against variable cost per unit, and the risk of interruption during the extra days. For slow-moving, low-value goods, a longer consolidation window and a fuller load usually win.
A plan that only describes the happy path is not a plan. Every lane has a known failure catalogue, and the difference between a professional operation and an amateur one is not that exceptions do not occur, but that they are detected quickly, assigned to a named owner, and resolved against a defined escalation path. Exceptions fall into a few families. Documentation errors are the most frequent and cheapest to fix if detected early. Customs and regulatory holds include examination, missing licences, valuation queries, and trade-control screening, governed by the authority rather than the shipper. Capacity failures cover blanked sailings, rolled bookings, and container shortages. Infrastructure disruption covers congestion, weather, strikes, and border restrictions. Each family has a different response owner.
Two habits pay for themselves. Record the cause of every exception in a small controlled taxonomy, which turns anecdote into a trend line and makes the case for a process change evidence-based. And set a service-level expectation for exception response, separate from the delivery promise.
| Failure mode | Consequence | Correction |
|---|---|---|
| Vague product description | Classification query, re-entry, detention | Description library mapped to a code |
| Mixed document versions | Discrepancies across invoice and packing list | Single source of truth with change control |
| Booking before goods readiness | Rolled booking, storage at origin | Confirm readiness, then re-confirm |
Measurement determines whether any of this improves, and a cross-border operation needs two metric families with disciplined definitions, because ambiguous definitions produce numbers that cannot be compared and therefore cannot be acted on. The cost family should include total landed cost per order or shipment, decomposed into freight, insurance, brokerage, duty, import taxes, handling, demurrage and detention, and expedite premium. It should also include pipeline cost: inventory value in transit and at overseas nodes. Comparing actual landed cost against the cost modelled at order time exposes planning error.
The transit time family should measure end-to-end lead time from order release to delivery, split into dwell segments: production and allocation, export clearance, main carriage, transshipment dwell, import clearance, and final delivery. On-time performance should be measured against a stated promise date and reported as a distribution rather than an average, because an average is insensitive to the tail that damages customer relationships. Two rules keep the metrics honest: define each timestamp at the physical event rather than the data entry, and keep the metric set small enough to review monthly, with every measure tied to a decision.
Build the horizon from the components of the lane: confirmed production and goods-ready time, export handling and clearance, main carriage including transshipment, import clearance including any examination, and final delivery. Use a conservative percentile of each variable, not its average.
Air freight is justified when the value at risk exceeds the freight differential, or when delay would cost more than the carriage premium, for example a production line stopped while awaiting a replacement component. It is usually not justified for low-value, non-critical goods. Always compare total landed cost including clearance and handling, since part of the apparent time saving is consumed by expedited handling and the paperwork it triggers.
Plan on the tariff treatment that actually applies, and confirm classification and valuation basis with a customs specialist before the first commercial shipment, since an incorrect assumption is expensive to unwind later. Where rates depend on origin, verify that the origin criteria and evidence are met. For large or variable volumes, ask about deferred duty or bonded warehousing.
The importer of record normally owns the clearance obligation, but the party that prepared the documents and the party holding the customer relationship also need visibility of the event. In practice this means a single named owner per shipment for exceptions, a standing escalation path to the brokerage or carrier, and a customer communication plan agreed in advance for delay bands. Without a defined response deadline, ownership tends to produce silence.
Shanghai International Logistics Co., Ltd.