How global logistics planning actually breaks down (a field note)

We keep a private list, updated quarterly, of the points at which the planning sheet stops matching the freight desk. It is not the list anyone in a planning role would publish. It's the list of moments where a perfectly reasonable forecast, built off purchase orders, lead times, and historical carrier performance, quietly drifts six to fourteen days from what's actually happening on the water. This note is a field-grade summary of what's on the list and why it ends up there.

Three things up front. We are talking about mid-market exporters with $5M–$80M revenue, shipping mostly ocean FCL with some LCL and the occasional air spot. We are talking about planning horizons of 30–120 days, which is the band where most of the broken decisions get made. And we are talking about real failure modes, the named, recurring patterns, not the abstract "supply-chain uncertainty" framing the textbooks like.

Failure mode one: the optimistic lead-time

Every planning spreadsheet contains a lead-time number. Origin port to destination port, by lane. The number is usually the median of the last 90 days. The number is wrong almost on principle, because the distribution it summarizes is not normal. It's a skewed distribution with a thin lower bound (you can't ship in zero days) and a long upper tail (you can absolutely sit on a vessel for an extra fortnight). The median understates expected lead time by 8–22% in our sample.

The fix is not to use the mean instead. The mean swings around on a single bad voyage. The fix is to plan against the 75th percentile for safety-stock decisions and the median for cash-flow modeling, and to keep the two distinct in the sheet. We find most operators conflate them. The result is an inventory plan built on a lead-time that, on a quarter with two missed voyages, simply will not hold. Our note on inventory positioning picks this thread up.

Failure mode two: the carrier schedule lie

Carrier schedules are not predictions. They are commercial offerings. When a carrier publishes a 28-day transit on a lane, that is the figure they would like to sell against. The figure that voyage actually clocks is set by blank sailings, port skips, slow-steaming decisions, and bunker economics, none of which are in the published schedule.

The Drewry schedule reliability index has been a useful anchor here for years. In tight markets, and we've spent most of 2024–2025 in one, global on-time reliability sits in the 50–65% band, with route-specific outliers as low as 35%. A planning sheet treating the published schedule as ground truth is, on a typical lane, asking for an embedded 30–40% error rate on arrival timing. The fix is to subtract a lane-specific reliability haircut at the planning step, not to discover it at the receiving dock.

Failure mode three: the inland leg as a footnote

Planning sheets handle the ocean leg in surprising detail and the inland legs as a single number. "Five days from port." That number was probably true once, on a specific lane, in a season that no longer exists. Drayage availability, chassis pools, intermodal rail capacity, and customs hold rates all move independently of the ocean curve and on different time scales. We've watched a vessel discharge cleanly and the cargo then spend nine days inside the terminal because a chassis shortage downstream had backed everything up.

This is the failure mode that connects most directly to our note on last-mile in cross-border B2B: the part of the journey that planners treat as a rounding error is structurally the most variable. A planning sheet that treats inland as a constant will mis-time the last 15–25% of journeys.

Failure mode four: the demand-side feedback that never closes

The fourth failure mode is the one nobody likes to discuss, because it's not a vendor or a port or a carrier. It's the buyer. The exporter's customer changes a forecast, or pulls forward an order, or sits on a release, and the new signal often arrives in the sales pipeline three weeks before it reaches the planning sheet. By then the vessel is booked, the container is loaded, and the inventory is moving against an outdated picture.

We have not found a clean technical fix for this. The closest thing we've seen work is a discipline: a weekly fifteen-minute call between sales and ops, with a fixed agenda and a written change log. It is unglamorous. It works. The version that doesn't work is the one with no change log, where verbal updates evaporate by the time the next P.O. lands.

Failure mode five: the schedule built off the wrong calendar

Finally, the calendar problem. Planning sheets are built on the Gregorian calendar of the planner. They ship cargo across the calendars of origin and destination, neither of which is the planner's. Chinese New Year is the obvious case and people now plan for it, more or less. The cases that still catch people are Eid in Middle Eastern transit ports, Ramadan working-hour effects in customs offices, Golden Week in Vietnam, and the surprisingly disruptive August window in parts of southern Europe. The planning effect of each of these is well-known to anyone who has worked the lane and invisible to anyone who hasn't.

The corrective is mechanical: maintain a calendar overlay that flags any vessel ETD or arrival falling within seven days of a known origin-or-destination holiday block, and book buffer accordingly. We've watched teams that don't do this absorb 4–8% of their annual demurrage line just from holiday-window misses.

What the field note is not

This is not an argument against planning sheets. The teams that don't plan move worse cargo on worse terms than the teams that do. It's an argument for being clear-eyed about where the sheet is structurally weak. The sheet is strong on what was promised. The sheet is weak on what was delivered. The freight desk works in the gap. Closing the gap is, in our experience, mostly about naming the failure modes rather than buying anything to replace them.

The instruments to monitor each of these are already in the public domain, from UNCTAD's logistics reporting at the macro level down to the lane-by-lane operational data. The work is in the integration, not the acquisition.