Off-season windows: when to ship a heavy parcel

We treat a parcel as heavy once its chargeable weight lands above ten kilograms. That is not a round number chosen for convenience: it is roughly where the option set narrows, because weight ceilings start excluding services, per-parcel minimum charges stop dominating the total, and the per-kilogram band becomes the line that decides the bill. A heavy parcel is also slower to reroute, which is why timing matters more for it than for a two kilogram envelope.

The pattern itself is simple and is set by the origin calendar more than the destination one. Two congested blocks run from the November shopping event through the Christmas air cut-off, and through the four to six weeks before Chinese New Year, when factories push out final orders and then stop. Between them sit two quiet stretches: mid-March to mid-April, and mid-August to mid-September.

The two calendars are not aligned, and that misalignment is the part most people miss. A quiet week on the origin side can coincide with the destination building inventory for its own season, and a destination lull can sit on top of a Chinese holiday week. A trough is therefore a statement about the origin pipeline, not a promise about the door, and the window worth using is the one where both sides are calm. The misalignment shows up most sharply in late autumn, when a mid-November shopping event pulls a large share of orders into the origin pipeline in a single week while the destination is still two weeks away from its own busiest stretch. A parcel ordered in that week sits in the origin queue during the calm it was supposed to avoid, and then arrives exactly as the destination congestion begins. Reading the two calendars together, rather than either one alone, is what keeps a heavy parcel out of both queues.

A diagnostic separates the two sides in the time it takes to read two timestamps. If domestic dispatch is running at its usual latency while the parcel is arriving late at the door, the constraint sits at the destination and no origin-side timing will remove it. If dispatch itself has stretched, the parcel is still upstream and the decisions available to you are still cheap. Running that check before choosing a service prevents the most common wasted spend in this category, which is paying for speed to solve a problem that is not in the lane you upgraded.

We can measure stage durations for our own parcels and nothing else. Our log records five timestamps per parcel: payment, first domestic scan, inspection clearance, international freight payment, and line acceptance, plus the delivery event at the far end. With forty-one parcels in the log, that is enough to describe ranges and to notice a direction, and it is nowhere near enough to publish a seasonal average. Anyone who quotes you a monthly average from a sample that size is describing their sample.

What we cannot measure is where most of the seasonal effect actually lives. Airline space allocation is commercial and unpublished. The number of cut-offs a line runs in a given month is usually not published either. Customs batch sizes are invisible from outside. For all three we write Not verified and plan on the pessimistic end, which is the honest use of an unmeasurable variable.

The third kind of evidence is published notice. Carriers announce peak surcharge periods and holiday cut-offs, platforms announce warehouse pauses, and customs authorities publish holiday processing schedules. Those notices are the closest thing to a shared calendar, and we log the ones we can date in the quarterly snapshot series rather than paraphrasing them from memory. Where a notice stops being findable, the row stops being carried forward.

Building a personal sample beats reading anyone else's, including ours. Five timestamps per parcel is the entire method: payment, first domestic scan, inspection clearance, freight payment and line acceptance, plus the delivery event at the far end. After a dozen parcels across two seasons you hold ranges for your own route, your own goods and your own sellers, which is a narrower sample than ours and a more relevant one. The bias in a personal log is visible and knowable, and that is precisely its advantage over a general table whose sample you cannot inspect.

The first window is mid-March to mid-April. The holiday backlog has usually cleared by then, factory output has resumed, and the build for the June event has not started. For a heavy parcel this window has one specific advantage: intake benches are empty, so a large carton moves from receipt to shelving without waiting behind a stack of small ones. The risk is thinner consolidation schedules, and a heavy parcel that misses a build waits longer than a light one because fewer builds accept it.

The second window is mid-August to mid-September. Summer capacity is still available, the fourth-quarter build has not begun, and sellers are between seasonal pushes. The risk sits a fortnight after the window closes: National Day week produces a short, sharp intake pause at origin, and a parcel that enters the pipeline a week late can land inside it. We treat the last ten days of September as already inside the peak for planning purposes.

Using a window is arithmetic rather than intuition. Work backwards from the date you need, add the community-reported upper bound of the transit range for your line type and destination, add slack for the peak shoulder, and pay the international leg on the day inspection clears. Where the arithmetic says the parcel has to move outside a window, it has to move, and the useful response is to choose the line deliberately rather than to hope the date holds.

A third window opens less predictably: the days immediately after a holiday ends, before the backlog has formed. Sellers are back at full output, the intake queue is briefly empty, and the first flights after a pause still have space. It is short, it does not arrive on a fixed date, and it is the one window that rewards watching rather than planning. What closes a window early is worth knowing as well. A large shopping event pulls volume forward by a week or two, so a stretch that looked quiet on last year's calendar can be busy on this year's, which is why every note we keep on a window carries the date it was written.

The two ways of being wrong are not symmetric, and the asymmetry is the reason we bother writing about windows at all. Shipping late into a congested block costs a higher band on the per-kilogram line plus a longer queue, and for a heavy parcel both of those scale with weight. Shipping early into a quiet stretch costs a few days of waiting for a build to close, less choice of service on some weeks, and almost nothing in money terms.

Storage is the input people leave out of the comparison. Waiting in the warehouse is not free: beyond the published free window, storage is charged per day, and a heavy parcel occupies the kind of space that gets charged. So "wait for a better window" is a decision with a running cost attached, and the comparison is between that daily cost and the band difference you are waiting to avoid. We do not publish a storage rate here because it changes; the basis is recorded on the provider pages.

When the plan fails anyway, three fallbacks are worth keeping in mind. Split the parcel so the date-critical items travel on an express service while the dense remainder takes the slow lane. Switch to a service with a higher weight ceiling rather than paying a surcharge on a service that barely accepts the parcel. Or accept the delay and move the date you promised, which is frequently the cheapest of the three and the hardest to say out loud.

One habit turns a wrong call into something diagnosable rather than merely irritating: write down what you expected before you pay. The date, the service, the quote band and the stage you believe will bind, recorded in the same row as the parcel. When it arrives, the comparison says whether the error was in the calendar, in the service or in the parcel itself, and that is the only mechanism by which a seasonal judgement improves. Without the note, a late parcel teaches nothing except that late parcels happen, and the same choice gets made the same way twelve months later.