Gemini makes schedule reliability a network-design question

The launch reorganized a substantial part of East–West shipping

Maersk's launch release described a shared network with Hapag-Lloyd covering major East–West trades. It said vessels would phase in from 1 February through late May, with June expected to be the first fully phased month. That transition detail was as important as the headline scale. A new network does not begin everywhere at once; old agreements wind down while vessels and schedules move into the new pattern. Cargo owners booking during any network change should confirm which schedule applies to their specific sailing. Historical transit performance from an expiring service may not represent the replacement. Keep the booking confirmation, planned port sequence and cut-off together so procurement and operations compare the service actually being sold.

The hub-and-shuttle architecture is the product

Gemini's published design combines 29 mainline services with 28 regional shuttle services. The intended logic is to keep the long-haul network controlled through selected hubs while connecting a wider port range with dedicated regional loops. For a shipper, this architecture can create two different experiences. Cargo at a mainline port may gain from a simpler long-haul pattern. Cargo at a shuttle port depends on a well-executed connection. Neither is inherently better or worse. The deciding factors are the lane's frequency, connection window, terminal performance and recovery rule. Ask the provider to show the route as a sequence of operational legs. A network name alone does not reveal how many handoffs sit between the supplier and consignee.

A reliability percentage needs a definition and a lane

Maersk called the above-90% figure an ambition for the fully phased network. Customers should avoid treating it as a contractual guarantee or a result already achieved at launch. They should also ask what is measured: vessel arrival within which time window, at which ports, over what period? Even a credible network-wide figure may not describe a particular shuttle connection. Build an internal measure from cargo-ready date to cargo-available date, then record the variation around it. If the carrier supplies schedule-performance data, compare definitions before combining the two. The point is not to dismiss the ambition. It is to make reliability relevant to purchase orders, factory plans, inventory and customer promises.

Connections carry a cost even when they work

A planned transshipment can be efficient, but it introduces handling, dwell and dependency on the next vessel. Those factors should be visible in the commercial comparison. Ask whether the quoted transit includes the expected hub dwell and whether the onward service runs often enough to recover from a missed connection without a long delay. Sensitive cargo may require additional controls: reefer monitoring, dangerous-goods acceptance, secure high-value handling or protection from repeated moves. For standard dry freight with reasonable inventory, the network may offer a strong balance. For urgent project material, the cost of one missed connection can dominate a small rate saving. Route selection should reflect consequence, not a blanket preference for direct or transshipment services.

Shared vessel capacity does not remove allocation decisions

The launch described a fleet of around 340 vessels in the shared network. That scale provides broad capacity, but it does not mean every sailing, equipment type and origin is open. Each carrier continues its own commercial relationship with customers, and local demand can make space or empty equipment tight. A useful booking process confirms both: room on the intended voyage and the box needed to reach it. Submit accurate volume and weight early, especially for special equipment or a peak week. If cargo readiness is uncertain, discuss alternatives before reserving a plan that the supplier cannot meet. Network capacity is a system resource; the shipper experiences it as a specific confirmed booking at a specific depot and cut-off.

Judge a new network over a sequence, not one shipment

Early performance can be affected by phase-in work, seasonal demand and unrelated disruptions. One smooth shipment proves little, and one rolled box does not by itself disprove the network design. Track a representative group of movements by lane and cargo type. Record planned and actual departure, hub connection, final arrival, cargo availability and exception communication. Separate delays caused by supplier readiness, customs or delivery from carrier-network events. After enough shipments, compare the result with the previous service on the same definition. This creates a fair procurement record and identifies where inventory or booking behavior should change. Gemini's launch gives the industry a serious reliability proposition; disciplined lane-level evidence is how a cargo owner decides whether it delivers.