Short answer. The fastest method is designed around every passenger walking on in one exact order, which is hard to arrange at a crowded gate. Boarding groups also do other jobs: they give priority to people who need assistance, reward frequent flyers and premium fares, and keep families together. In this simulator, the fast methods retain much of their advantage even when many passengers ignore the order, so the obstacle is largely practical rather than mathematical.
What the fastest method asks of passengers
The Steffen method boards alternate rows, one side of the aisle at a time, with window seats first. For a 30-row aircraft, the first wave follows a sequence like 30A, 28A, 26A and so on towards the front, then 30F, 28F, 26F on the other side, followed by the window seats in the odd-numbered rows on each side. Middle seats follow in the same pattern, then aisle seats.
Simulator result: On the reference flight used across these articles (easyJet A320, 175 passengers, typical cabin luggage, forward door only, median of seeds 20260801 to 20260808), it took 8:21. Back-to-front zones took 19:15, and conventional airline groups took 19:26.
To get that result, the gate has to line people up in something close to that sequence. That means either calling many small groups one after another or assigning numbered positions at the gate, with passengers ready to board when called.
The reasons airlines give other priorities
Priority boarding is something airlines sell. Premium cabins, frequent-flyer status and paid priority all come with early boarding. A sequence built only for speed would mix those passengers in with everyone else.
Families and groups travel together. Parents do not want to board ten minutes apart from their children, and they should not have to. A sequence that splits a row by seat letter can split families.
Passengers arrive at different times. Some are still in the lounge, the queue for coffee or the bathroom when their group is called. A method that relies on exact order can lose part of its advantage when someone is late.
Gate space. Lining up 180 people in a precise sequence needs room and staff time that many gates do not have.
What happens when passengers do not follow the order
The simulator has a setting for the share of travel parties who board one or two groups away from the group they were called in. Someone travelling alone counts as a party of one, and on the reference flight everybody travels alone, so there the share of parties is also the share of passengers. This is a practical stress test for an optimised order.
Simulator result: Reference A320 flight, share of passengers boarding out of turn:
| Out of turn | Steffen method | WILMA | Back-to-front | Random queue |
|---|---|---|---|---|
| 0% | 8:21 | 12:31 | 19:15 | 14:50 |
| 10% | 8:56 | 12:30 | 18:44 | 14:50 |
| 20% | 9:39 | 12:44 | 17:55 | 14:50 |
| 40% | 10:38 | 13:40 | 16:19 | 14:50 |
Simulator result: With 40% of passengers out of turn, the Steffen method was about 27% slower than with perfect discipline, but still faster than a random queue. Back-to-front got faster because the drift mixed its zones and spread passengers out. The random queue did not change because it had no order to break.
Families have a similar effect. Simulator result: With half of passengers travelling in parties who board together, Steffen went from 8:21 to 10:07, WILMA from 12:31 to 13:47, and back-to-front from 19:15 to 20:45.
What airlines do instead
Many use compromises that capture part of the gain without an exact sequence. Window-middle-aisle groups for economy are the simplest: they need only three calls and they sharply reduce passengers squeezing past seated neighbours, although families who board together bring some of that back. Simulator result: On the reference flight, WILMA took 12:31, about 35% less than back-to-front. Other approaches include boarding the rear and front of the cabin through two doors at once where the airport allows it.
Research finding: The Steffen method was derived by optimisation for a single-aisle aircraft with one door, and a later physical experiment with volunteers supported the advantage of parallelised methods over block boarding. Neither the research nor this simulator accounts for the commercial reasons above, which is why the question is ultimately one for airlines, not for the model.
Frequently asked questions
Has any airline used the Steffen method?
The method is mainly used in research and demonstrations. The ordering it needs is difficult to run at a real gate, and airlines generally use simpler groups that capture part of its benefit.
Would the Steffen method still work if people board in the wrong order?
Simulator result: Largely yes in this model. With 40% of passengers out of turn, it slowed from 8:21 to 10:38 on the reference flight, still faster than the other methods tested there.
Why do airlines board by zones if random is faster?
Zones organise the gate, give staff a clear order to announce and let priority groups go first. In this simulator, conventional back-to-front zones were slower than a random queue on the reference A320 and on the Ryanair 737-800.
Updated 2026-09-25. Figures labelled simulator result are outputs of this site's model at the settings stated, not airline measurements. What the model leaves out.