After posting Tuesday’s rundown of Etihad’s poor on-time performance on flights to the US, I received a lot of questions about how others stacked up. Now I have some answers for you. In short, Etihad is by far the worst performer over the Atlantic, but that doesn’t mean it’s the only airline running a bad operation.
I went back into the masFlight database and pulled up some summary tables for a bunch of airlines. I even looked into smaller carriers but decided to focus on the biggest Transatlantic players since the small players didn’t really add much to the discussion. Then what I did was look at on-time performance for flights departing each airline’s hub to go over the Atlantic. Yes, this means US carriers have eastbound flights while the other carriers have westbound flights. But I think it’s more important to compare how an airline does leaving its own hub since that’s where it has the most tools available to impact its on-time performance.
The end result is a somewhat clear stratification with Middle East carriers performing the worst (except Qatar), European carriers in the middle, and US carriers performing the best. Go figure. Let’s just lay out the raw data and let you pick at it, shall we? (You might need your glasses.)
A few things to note in this chart.
- The blue shading of each airline name is meant to show which ones are Middle East carriers, which ones are European carriers, and which ones are US carriers.
- You’ll notice that there’s some yellow shading as well. That’s merely meant to point out that data isn’t very complete for some airlines. Turkish lags a bit, but it’s really Etihad and Qatar that have the least comprehensive data. Please keep that in mind when you look at this.
- The current sort is using the arrivals within 14 minutes of schedule metric (far right column).
- To make it easier to see, the worst performer in each category is in red while the best is in green.
So what does this tell us? I do find it interesting to see how many airlines are bad at getting their airplanes out exactly on-time. It’s no surprise that US Airways (followed by Delta) are the clear winners in this area. Both have hammered home the importance of getting out on time. Hopefully US Airways can transfer this belief over to American. Meanwhile the Middle East carriers, minus Qatar which does alright, are the worst. And Lufthansa is pretty bad too.
What’s equally interesting, however, is the jump in on-time performance when you include the 15 minute grace period. Lufthansa goes from having one of the worst performances to one of the best. Clearly it can’t get out right on-time, but most of the departure delays are very minor. Interestingly, American has the opposite problem. It gets a good jump with flights going on right on time. But if they don’t get out on time, they’ll likely take a more significant delay.
Take a look at Emirates and Turkish as well. Both those carriers don’t do well at getting out right on time, but like Lufthansa (though not as good), they get a lot out within 15 minutes. The difference in performance between them and the rest of the pack shrinks dramatically when using that 15 minute mark. In fact, it’s Etihad that becomes even more of an outlier at that point, with the spread between its performance and that of the rest of the carriers widening tremendously.
Let’s flip this and look at arrivals. As a reminder, this shows arrivals in the US for European and Middle East carriers but arrivals on the other side of the Pond for US carriers. Though we don’t have data for all of Etihad’s flights, the ones we do have aren’t often on time.
Once again, the US carriers do well here, but take a look at United. It only gets 65.6 percent of flights out within 15 minutes of schedule yet 69 percent arrive on time. That looks like a clear case of schedule padding to me. I know there was a lot of talk in the comments on the last post about how departure time doesn’t matter because you can pad the schedules. I disagree. While departure time matters less than arrival time for sure, you can’t simply discard it completely.
A quick look at Etihad does indeed show some schedule padding, but if you can’t even get 10 percent of your flights out within 15 minutes of departure, then you’re just not going to recover unless you do some insane padding. United and Etihad look to be the worst offenders, though with padding, United is able to get to a respectable level of on-time arrivals. Etihad is not.
Looking at the last column we can see how arrivals stack on based on the DOT standard of “on time” as being within 14 minutes of schedule. Etihad is still the worst by far. Emirates and Turkish aren’t good, though leaps and bounds ahead of Etihad. Air France and BA have issues as well, but it only improves from there.
What does all this mean? Well it’s pretty clear that the US carriers are running the best operations over the Atlantic these days while the Middle East carriers (except for Qatar) are firmly at the bottom of the list. That being said, Etihad is in a league of its own here, and not in a good way.
21 comments on “Etihad Isn’t the Only Airline With an On-Time Performance Problem Over the Atlantic”
Seems like the airlines flying to the Indian subcontinent a lot are most delayed.
Varun – Well, yes, if you consider the Middle East carriers as the ones that fly to India most, then they are most delayed. But these flights have nothing to do with on time performance into India. It’s only looking at flights to the US. So they could (and in the case of Etihad, most likely do) perform better on Indian routes.
I actually think a departure within 15 minutes of the scheduled time can realistically be called an on time departure, especially as all the flights of the non US airlines are long haul and 15 minutes is a very short time compared to the flight length. What are dismal are the European airlines arrival stats. The Germans have a reputation of a very time conscious nation. Only getting 2 out of every 5 flights in on time isn’t anything to brag about.
How much of this is trying to get passengers through the gauntlet of inbound US flight security? Does this skew the data a bit? Its easy to get out of the US, and lousy to come home with the most remote gates, extra security, and the “20 questions” song and dance that happens on the return trip. I wonder what percentage of the delays is just due to that.
I think you are missing the point, this has nothing to due with what the passenger has to do after arrival to the USA, but if the arriving aircraft is on time. Each airport has dedicated gates for arriving International passengers to allow migration to ICE without leaving a secure area, so there shouldn’t be delays from landing to arriving at the gate. Pre-expansion DTW used to be a problem as they only had 4 gates at a segregated Customs area, so if more than 4 flights arrived at the same time, a delay occurred until a plane could get cleared, unloaded, and towed away from the International gate.
I think what Jim is getting at are foreign airports with CBP Pre-Clearance facilities (i.e. AUH), and whether the song and dance required for that process is contributing to delays. For example, going from HYD to DFW, EY shows a connecting time of 2 hours 20 minutes at AUH. Ordinarily that might be alright, but if doing so requires that you go through a security check, walk to the farthest end of the terminal, go through a secondary security check, and pre-clear customs, would that contribute to delays? I’m guessing EY isn’t just going to allow half their connecting pax to miss their flight just because the three-ring circus takes so long…
How much of this is trying to get passengers through the gauntlet of inbound US flight security? Does this skew the data a bit? Its easy to get out of the US, and lousy to come home with the most remote gates, extra security, and the “20 questions” song and dance that happens on the return trip. I wonder what percentage of the delays is just due to that.
Arguably arrival schedules should incorporate some measure of average “delays,” so that people can realistically make onward connections. You can call it padding, but you can also it realism about operating into busy airports.
It seems like United’s departure delays which are then recovered to a greater degree inflight mean that more passengers will make their booked connections at the other end. If you have a schedule that reflects best case scenario you’re going to be rebooking people more frequently.
I’ve regularly used the data on Flightradar24 to calculate the amount of time that should be added on to the scheduled arrival time to allow for a bit of confidence in making a connecting flight. I noticed that many airlines seem to use average or median times, almost certainly to maximize staff and fleet productivity and to come up at the top whenever someone sorts by flight duration. Others do seem to pad a bit: I remember one exercise in which the median (or “typical”) Delta-branded flight could be expected to arrive 10 minutes early.
But medians, though great for point-to-point operations, makes life far more stressful for passengers and employees alike at hub airports. Hubs are probably the most complex systems in which 50-percent confidence is considered good enough.
The ideal hub would be designed so that any given passenger could expect, with 90 or 95 percent confidence, to disembark, walk at a normal 3 m.p.h. pace to the next gate, and board the connecting flight with 15 minutes to spare. This would require airlines to add 30-45 minutes to their scheduled arrival times at the larger hubs; maybe five minutes less at smaller airports. It would also raise MCTs to about 60 minutes if no additional security screening, border formalities or long hikes to other terminals or concourses are required; and longer if they are.
A final note about scheduling realism: As a Canadian who lives closer to the MSP and ORD hubs than to YYZ, YUL or YVR, one thing that stops me from using those hubs, and U.S. carriers for some or all of my trip, on the way to and from Europe is not knowing whether clearing U.S. Customs and Border Protection on the return trip will take one hour or four. Despite the uncertainty, United persists in selling 80-minute International to Canada connections via ORD, which seems absurd. Delta seems a bit better, with two-to-three hour connections on the return trip via MSP, but even then I’ve heard complaints of that being just barely enough time. Too bad, as some of the prices and schedules via U.S. airports would otherwise be very competitive.
Just guessing, but it almost looks like on time performance is inversely proportional to the reliance on connecting traffic. There’s a careful balancing act between ground times and managing the volume of connect pax and bags. The U.S. Carriers and the major Alliance players tend to do that well. The Gulf carriers haven’t found that sweet spot yet. It also seems to get worse when you look at long haul to long haul reliance vs a more traditional short/mid haul to long haul. Since that data gets held closer to the vest, proving that gets difficult.
How many flights a day are coming out of the hubs you have listed and how many are missing due to not including alternate hubs like MIA and DFW for American, DTW for Delta, Houston for United, and Charlotte for pre-merger US Airlines? Living in Miami, I see British Airways, Air France, and Lufthansa are usually on time, with Air France being the major culprit of late flights.
However AA, has a major on time departure issue with their flights to South America. I wonder what the numbers look like since they went to their hard banks versus the rolling banks they used to operate prior to the merger.
TC – The number of flights tracked is listed in the chart. I didn’t bother including every hub with a Transatlantic departure but instead focused on home hubs to get the point across.
Cranky, can you do the same type of analysis for trans-Pacific airlines and
Bob – I could, but I’m heading out on vacation tomorrow and won’t be able to touch this for awhile. Might be a fun follow up in a month or two.
Feel free to ping me as a reminder.
So let’s say UAL (or any airline) knowingly pads their schedule. There are a number of reasons they might do that but to get to the point why do we care? You buy the ticket based on the schedule that is published and if your in early because it was padded why is that bad? There are reasons it’s bad for the airline but why is it bad for the customer?
121Pilot – From a customer perspective the biggest issue with padding is if the flight actually leaves on time. That means you might get to your destination really early, and that does sound good in some ways but it can be a real pain for many who are counting on the arrival time. Now, is that better than arriving late? Heck yeah. When I was at America West, we ran an awful operation (most people didn’t know it because United was bigger and far worse back in 2000). So we started padding the heck out of the schedule in order to get the operation back on track. It worked until we could get reliability up and then we were able to pull back on it.
Really, schedule padding is a sign of an unreliable operation. It says that you know you can’t fly on time so you need extra buffer. That adds a lot of cost and shareholders really don’t like that. It’s great as a tool to use when you are trying to fix a problem, but it’s not something to rely on.
Another problem with schedule padding is that it can render illegal connections that you would otherwise be reliably able to make. So, if Delta has a minimum int’l-domestic connection time of 1:45 at JFK, and has padded its schedule by half an hour, I wouldn’t be able to legally book a connecting flight that leaves 1:30 after my international flight is scheduled to arrive, even though the flight regularly arrives earlier than scheduled and I have Global Entry to sail through Customs. Instead, I am stuck waiting hours for the next flight to my destination, or paying the standby fee to hopefully get on the flight I would have initially wanted to book in the first place.
So, basically you are comparing performances on flights originating from different parts of the world to almost identical destinations, but not on the same routes, while taking some very dubious benchmark, while you fail to explain why and how it is important, and you come and claim that you have something to tell us. Well, sorry to disappoint you but while it is interesting, it is not telling us anything. How do you say in English – aaaaa, what it was——–aaaa that it was – comparing apple to oranges. D-, but you still get the class credits for your effort.
Unless you start comparing performances on identical city pairs, preferably in similar time slots your musings are useless. Besides, on-time performances, is just a part of what customers take into in account when they book let’s say a 13 flight. Flying New York – Washington, it is of utmost importance, not so much when you fly Doha-Dallas, especially when majority of passengers on the latter flight are not originating in Doha and most probably 15 or 40 minutes in 25 hour itinerary is not so important.
Uh no. That’s not how this works. Airlines in their entirety are compared on all sorts of metrics, because finding one route where several compete is a) rare (near non existant in the ME3’s case) and b) completely useless unless you’re on that one route. The ME3’s on time performance is atrocious and has been for a while, and they (and most EU carriers) have a much more simple operation than the US carriers, with fewer fleet types, and *one* hub, with fewer banks, but a much stonger necessity to ensure connections. There’s really no excuse when it’s such an integral part of your network.
Good point. This comparison gives too good impression of US airlines punctuality although the difference is probably quite small. For instance United’s B757s can easily fly from New York to Scandinavia & Berlin during the winter but fuel stop in Gander/Bangor is often necessary during the return flight.
What I’m curious about with these stats is trans-Atlantic performance on non-US routes, particularly Canada and the Caribbean. British Airways, EgyptAir, Air Canada, Aer Lingus, Jet Airways, Emirates, TAP, Austrian, Lufthansa, Air France, and Air Transat all have regular trans-Atlantic flights to and from Toronto that could be tracked, and a large number of European airlines (TUIfly, British Airways, etc.) have flights to and from Caribbean and Central American destinations. I think Jim M has a point…even in non-pre-cleared situations, the airlines still have a ton of extra hoops to jump through for US-bound departures (extra documentation, extra security, etc.).
Also, with KLM, are we taking into account Amsterdam’s layout of performing security checks at the gate? I have flown through Amsterdam many times (admittedly to Canada), and I always see delays out of the non-Schengen zone because of problems with the at-gate security checkpoint (too many people arrive at the gate too late).
I fully believe the data, but have to wonder what other factors play into this that could be acknowledged with a slightly wider look at the problem.