Peak season: the six weeks when every line slows down

A peak-season delay does not begin at the warehouse. It begins at the seller, on the day order volume rises past what their dispatch bench can pack and hand to a domestic carrier. Nothing in the agent dashboard shows this phase, because the agent has not received anything yet. The first thing that moves is the gap between the moment you pay and the moment the domestic tracking number produces its first scan.

We started recording that gap because it is the one part of the pipeline we can measure without asking anyone for data. In our own order log, parcels placed inside the November to December block took between five and eight days to produce a first domestic scan, against a two to three day band for the same sellers in the March to April window. The sample is twenty parcels in total, small enough that we read the direction and not the size.

The consequence of week one is that it cannot be repaired later. Four days lost at dispatch stay lost; no amount of chasing at the consolidation warehouse returns them. If a fixed date matters to you, the date is decided in the week when you have the least visibility, which is the argument for treating the first scan as a monitored event rather than a formality.

Intake is a receiving process, and receiving processes degrade in a predictable way when arrivals exceed bench capacity. Parcels get signed for in bulk, stacked, and shelved afterwards. The visible symptom is a widening gap between the domestic carrier marking the parcel as delivered to the warehouse address and the agent marking it as received and weighed.

Two things stretch together during intake congestion: the wait for the first inspection photo, and the response time on any message you send. The storage clock is a separate matter. It is counted from intake rather than from arrival, so a slow intake does not consume your free days. A slow decision after intake does, and that is the part under your control.

What we do in week two is pre-commit. Anything already known to be wrong, a size read from the wrong column, a duplicated line, an item that will need a second photograph, gets flagged to the desk before the photos land, because during congestion every message competes with a longer queue. The second week-two habit is to pay the international leg on the day inspection clears rather than the day you finish deciding. The queue is at the front of the process, and parcels that enter it earlier are loaded earlier.

The status that matters here is not "shipped". It is the moment the line accepts the parcel into its own network, which is a different event from the warehouse printing a label and marking the parcel as dispatched. Between those two events sits a handover, and during a peak the handover is where days accumulate quietly because the tracking number often shows nothing at all.

The mechanism is capacity. Consolidated air freight moves when a build is full, and a build fills faster in December than in April. That sounds helpful and is not, because the space allocation behind the build is smaller at the same time. Freighter capacity is fixed and belly capacity competes with passenger schedules, so the queue lengthens even though each individual build closes sooner.

The question we ask of any line before relying on it in a peak month is how many cut-offs a week it is running, and whether that number has changed since the summer. A line that ran a daily cut-off in April may run three a week in December, and that single fact can separate a five day wait from a twelve day wait at the same stage. Where a line does not publish it, we mark the answer as unverified and plan on the pessimistic end of the range.

Customs brokers process manifests in batches, not parcels. Your parcel dwells inside a batch until the batch is released, and the batch is released as a unit. This is why clearance time is not proportional to anything about your parcel: a small parcel behind a mis-declared consignment waits exactly as long as the consignment.

Three entry-level facts decide most of what happens in week four. Whether the entry is filed electronically before arrival or prepared on arrival. Whether the destination requires a tax identifier, and at which point it is requested, since our Q2 snapshot records one destination moving that request to before departure. And where the consignment value sits against the published threshold, which our region pages carry with the date of the version we reviewed.

A clearance hold is usually a data problem, and data problems are the one category of delay you can remove entirely. In our own log the holds we have seen concentrate in parcels whose declared description was generic, whose declared value did not match the invoice inside the box, or which carried several unrelated categories under one line. Fixing that costs an evening of care and removes a variable that no amount of shipping budget can buy back.

When a hold does arrive, it almost always takes the form of a request for one document rather than a refusal, and the request carries a deadline. Missing that deadline is the expensive part, because a missed response generally sends the parcel back into the queue instead of producing a decision. The preparation that removes the risk is boring and cheap: keep the invoice, the item list and the declared value in one place before the parcel moves, so that answering the question takes minutes on the day it is asked. The parcels that have spent longest at this stage in our own record were rarely the complicated ones; they were the ones whose paperwork lived in three different places when the question arrived.

After arrival come the steps nobody photographs: depot intake at the destination carrier, sortation, a delivery attempt, and either a signature or a card through the door. Holiday staffing affects all of them at once. A destination carrier running reduced routes will hold a parcel for a day rather than send a van down an uneconomic road, and the tracking will keep saying the same thing while it happens.

This is the stage where our estimates deserve the least confidence. The windows we publish per destination are community-reported ranges rather than service levels, and during a peak the upper bound is the bound that carries the information. An express air courier window of four to nine days is a statement about a distribution, not about your parcel, and a single parcel never settles which part of the distribution it came from.

What you control in weeks five and six is small but real: an address that a driver can find without calling, a phone number that receives messages, a delivery point that is open when the van arrives, and a redirect decision made on the first failed attempt rather than the third. Two of the four parcels in our log that ran past the upper bound of their published range did so after a first failed delivery attempt rather than inside a depot queue.

Shipping earlier means paying the international leg earlier, not ordering earlier. That distinction carries the whole section, because the queue you are avoiding starts when the warehouse hands over the parcel, and the warehouse cannot hand over a parcel whose freight is unpaid. Five triggers are worth acting on: a destination date inside six weeks; a cut-off count that has dropped; a seller whose dispatch latency on your last order passed four days; a battery or a liquid anywhere in the parcel, since acceptance narrows first when volume rises; and a free storage window that ends before the peak passes.

There is a cost to moving early, and it is honest to state it. Once the international leg is paid the parcel is gone, which means you lose the option to add one more item and consolidate it at no extra base charge. In a quiet month that option is worth something. In a peak month it is worth less than the days it costs, because a parcel that misses a peak cut-off waits for the next build rather than for the next day.

If you are already late, the useful move is not to pay more for the same line but to re-cut the parcel. Dense goods with no date attached can travel on a slower line while the time-critical items travel on an express service, and splitting is worth its second minimum charge only when one part of the parcel has a hard date and the rest does not. Where nothing has a hard date, the correct decision is usually to accept the delay, plan for the upper end of the range, and stop watching the tracking page daily.

One further trigger belongs before ordering rather than after it, because it is the only one you can still act on cheaply. Where a haul contains one item with a fixed date and the rest without one, order the dated item first and on its own. A single-seller parcel can be shipped the day it clears inspection, while a three-seller parcel waits for its slowest domestic leg, and during a peak that wait is regularly longer than the international transit it precedes. Splitting at the ordering stage costs an extra minimum charge and buys back days, which is a trade worth making while the calendar is still open.