A border hall can be empty at one moment and hold an hour-long queue twenty minutes later. The cause is almost always the arrival pattern rather than the total number of travellers.
Arrivals bunch because flights bunch
Aircraft do not land evenly. Long-haul flights from a given region cluster because of departure times at the other end and the direction of prevailing winds across the route.
Several widebody aircraft can therefore disgorge passengers into the same hall within a few minutes of each other. The hall receives in one surge what it would comfortably process across an hour.
The queue that results is not a sign of understaffing on average. It is the visible consequence of a demand curve with sharp peaks and long troughs.
Desk capacity is fixed in the short term
Officers cannot be summoned instantly, and shift patterns are set well in advance around forecast arrivals. A schedule change or a diversion breaks the forecast.
Each traveller takes a variable amount of time. Most are processed in well under a minute, but a small number require questioning that takes many times longer, and one such case slows a whole lane.
Because service time varies so widely, queue length grows disproportionately once arrivals approach capacity. A small increase in demand produces a large increase in waiting.
Automated gates change the shape of the problem
Electronic gates read a passport chip and compare a facial image, processing eligible travellers quickly without an officer. Where eligibility is broad, they absorb most of the peak.
Their weakness is failure handling. A gate that rejects a passport sends that traveller to a staffed desk, so a bank of gates with a high rejection rate simply moves the queue.
Eligibility rules differ by country and by passport, and they change. A traveller who used the gates on a previous trip may find they are directed elsewhere on the next one.
Position in the aircraft determines position in the queue
Disembarkation is sequential, so passengers seated forward reach the hall first. On a full widebody the difference between the first and last rows can be substantial.
Walking distance then compounds it. At airports with long transit corridors or a shuttle between terminals, an early exit converts into a much better place in the line.
This is the practical reason many frequent travellers pay for a forward seat on long sectors even when the seat itself is identical to one further back.
The queue clears faster than it forms
Once the surge has passed, processing continues at full rate against a shrinking queue, so the hall empties quickly. The tail of a long line usually moves faster than the front did.
Travellers with flexibility can exploit this by not rushing. Time spent collecting a bag or using a restroom can cost nothing if the queue is draining at that moment.
Where a connection is tight, the opposite applies and every minute counts. Knowing which situation applies is worth more than any queue-choosing strategy inside the hall.