EUROCONTROL’s May 2026 snapshot records first-wave departure delay falling from 7.5 minutes in 2024 to 6.6 in 2025. Full-day delay fell from 17.5 to 14.7 minutes, and reactionary delay declined by 1.5 minutes per flight.[1]
The figures do not establish a causal multiplier. Dividing the full-day improvement by the first-wave improvement would produce an attractive number and an unreliable conclusion.[1]
The snapshot’s first-wave window runs from 03:00 to 08:59 UTC, using weekday observations from January through October in the two years. That definition matters. It is a specified slice of the European operation, not a universal measure of every airline’s first aircraft movement.[1]
The operational question is therefore more specific than whether mornings matter. How much of the later improvement can be attributed to departures beginning closer to plan, and how much reflects a network that was easier to operate throughout the day? The available aggregate comparison cannot separate those effects. A plausible mechanism is not the same as an estimated causal contribution.
A better summer had several authors
EUROCONTROL’s Network Operations Report records a substantial improvement during summer 2025: average departure delay fell from 23.4 to 18.2 minutes, and average arrival delay from 19.7 to 14.6. Its decomposition shows ground delay down 6 percent, departure delay associated with en-route regulation down 35 percent, and reactionary delay down 24 percent. Arrival-regulation delay remained at about the previous year’s level.[3]
Those summer figures use a different reporting window from the first-wave snapshot. They should not be spliced into one time series. Their value here is explanatory: several sources of delay improved together. A reduction in en-route constraints can benefit an aircraft that departed on time and one that did not. Fewer late inbound aircraft can then improve ground performance at the next station, making the direction of influence difficult to untangle from network averages alone.
The separate Performance Review Report points to a combination of operational initiatives and reduced convective activity. It also describes a network in which the five most delay-generating area control centers accounted for 47.1 percent of en-route ATFM delay despite handling 13.2 percent of controlled flight hours. Better punctuality did not eliminate concentrated capacity problems.[2]
The review’s annual arrival-punctuality measure improved to 76.1 percent in 2025 from 72.5 percent in 2024, while remaining below 2019. That is a meaningful recovery with a qualified baseline. Comparing only with a difficult previous year can make improvement look like the resolution of a problem that still constrains the operation.[2]
Where the first minute becomes valuable
Imagine two otherwise similar rotations. One has four sectors remaining and little ground recovery time; the other is returning to a maintenance base for an extended stop. An early delay of the same length does not carry the same downstream exposure. In the first case, removing it can protect several subsequent events. In the second, the scheduled pause may absorb it. This is an illustrative comparison, not a reconstruction of flights in the EUROCONTROL dataset.
The distinction suggests that the valuable unit of analysis is the rotation, with its planned recovery opportunities, rather than a departure considered alone. A morning improvement that appears small in a network average could be concentrated in the rotations most likely to propagate disruption. Equally, a widely distributed improvement could be absorbed before producing much visible benefit later. The average does not reveal which pattern occurred.
A useful study would compare similar aircraft sequences by airport, day of week, departure bank, scheduled ground time and disruption exposure. It would follow the aircraft far enough to see where delay was absorbed, transferred or compounded. It would also account for changes to the schedule itself. Adding recovery time can improve punctuality while changing utilization and the economics of the service.
The passenger can move in the opposite direction
The first-wave snapshot names earlier boarding and gate closure among the measures used to protect departures. It also describes punctuality being prioritized over passenger connections. This is the most consequential qualification in an otherwise encouraging operational picture: an improvement for the aircraft sequence can impose a larger delay on people left behind.[1]
That does not establish that holding for a connection is always preferable. It establishes that the decision has two outcomes worth measuring. Aircraft arrival punctuality counts the performance of operated legs; passenger arrival delay follows people to their intended destination. A small group missing an infrequent onward service could experience a large cost that barely registers in the flight-based average.
An operator evaluating a departure-protection policy would therefore want to know how many connections were preserved, how many were broken, and what recovery options existed. Holding five minutes at one bank might have a very different effect from holding five at another. The case depends on spare capacity, the next departure, remaining aircraft duties and the alternatives available to the affected passengers.
There is also a distribution question. Which stations, flights and passengers received the benefit of the improved operation? A system-wide average can improve while a persistent bottleneck becomes more disruptive for a particular part of the network. Reporting the median, the severe-delay tail and repeated problem rotations would make the improvement more legible than a single average alone.
A result worth following through
The first-wave result is useful because it offers a concrete starting point for examining interventions that occur before most of the day’s flying. It does not supply a universal multiplier or a reason to treat an on-time push as the only objective. The stronger interpretation is that early readiness belongs inside a wider account of schedule resilience.
For the next comparison, the most revealing evidence would connect interventions to complete outcomes: what changed on the ground, how the same rotations performed later, whether the gain survived difficult weather, and where passengers ultimately arrived. That would turn an attractive network statistic into a clearer guide to which minute is worth recovering, at which station, and at whose cost.
Sources & further reading
Original reporting and research behind this article.