Considering how much discussion we’ve had here on the blog about delays lately, I thought it might be a good time to dig in and review the latest Air Travel Consumer Report. This one covers January travel and it really wasn’t awful, surprisingly. January weather was bad, but February weather was awful. I’m not looking forward to seeing that one. Let’s get started.
On Time Performance
As usual, Hawaiian was the best, though it couldn’t crack 90 percent on time. It helps to do most of your flying around the mostly weather-free islands, so I always ignore their results. Alaska was next with 85.8 percent on time, probably showing that the Pacific Northwest has had a relatively mild winter this year. Then comes United with 83.7 percent, a great showing. But wait, we don’t know that for sure. Nearly half United’s flights are actually operated by another airline as United Express. (This doesn’t even count codesharing with alliance partners.) But those results all fall under that operating airline.
So SkyWest, which does a fair bit of flying for United, posted only a 74.6 percent rating, but that won’t show up in United’s numbers. That’s why I find these numbers hard to use and hard to trust. So how can we break these down and get more meaningful info? Well we can’t do it precisely, but we can start to approximate. Take a look at SkyWest again. They do most of their flying for either United or Delta and we don’t know how they did for each. How can we figure that out? They do break it down by arrival airport, so we can dig in that way.
In Denver, SkyWest had 78.4 percent of flights arrive on time (United had 88.6 percent). At LAX it was 79.6 percent (87.2 percent for United) and in Chicago it was 69.5 percent (84.2 percent for United). In San Francisco, SkyWest had a dreadful 53.2 percent while United had 76.9 percent. It’s a safe bet that the vast majority of those flights were operated for United.
On the other hand, SkyWest had an 81.9 percent on time arrival rate in Salt Lake compared to Delta’s even better 89.7 percent. Those were mostly operated for Delta. What does this tell us? It shows that no matter who you’re flying, regionals tend to operate fewer flights on time than the majors. The reason is that when the weather ramps up, smaller airplanes tend to get delayed first because fewer people are inconvenienced than if a big plane gets delayed. But it shows that the numbers you see are certainly deceiving. The DOT needs to change this.
Consistently Late Flights
Here’s a great reason why these numbers are again rough. Hawaiian may have had the best on time performance, but guess what? Of all the flights that were delayed at least 80 percent of the time in January, a full 40% were operated by Hawaiian. So they may have good on time performance in general, but don’t try to tell that to all the people stuck on those six flight numbers from the mainland.
But just because a plane is consistently late doesn’t mean it’s the worst thing around. For example, a whopping 87.1 percent of the time, Delta flight 6578 from Kansas City to JFK (operated by Comair) was delayed by an average of 34 minutes. That’s bad, but it’s not nearly as bad as Skywest 5942 between San Francisco and Sacramento. That flight was operated as United Express, and it was “only” late 83.87 percent of the time. But guess what? It was late an average of 92 minutes. You could drive there faster.
Flight Cancellations
When we look at flight cancellations, we can really see the plan to impact regional flights over mainline flights in place. Regional carriers filled the top 6 spots when it came to cancellation percentage with Pinnacle taking the cake with a 4.1 percent cancellation rate. US Airways was the top mainline carrier with 3 percent, but again, when you fill in the blanks with express carriers, it might change things entirely. We just don’t have that kind of visibility, especially since their two wholly-owned carriers (Piedmont and PSA) aren’t included in the data at all.
Ramp Delays Over 3 Hours
In January, there were only two flights with ramp delays over 4 hours, both Delta flights. The first was from Ft Myers to LaGuardia. That flight diverted to Harrisburg to wait out the weather. It waited for a long friggin’ time, but it did eventually go. With the new rule, this flight would simply have canceled in Harrisburg and some alternate mode of transportation would have had to be arranged, probably a bus.
The second was a flight from Charlotte to Atlanta. They were delayed coming out of Charlotte because of bad weather in Atlanta. In fact, there were a bunch of weather diversions into Charlotte so the gates were all filled and they couldn’t get back to the gate. They now have a procedure set up to get people off on the ramp.
Moving down to 3 hour delays, there were 21 flights that were delayed over that limit for a .004% rate out of all flights. Delta had the most with 8.
Mishandled Bags
I really hate this number. I’ve written about this in the distant past, but let me refresh your memories. The mishandled bag number is the number of lost bags per 1,000 passengers. What a stupid number. So one month you could have 1,000 passengers that check no bags and the next they could check 3 bags each. The number needs to be the number of lost bags per 1,000 checked bags. I won’t even bother here.
Complaints
This is another rough measure, because it’s only complaints that reach the DOT. People don’t generally complain to the DOT unless the airline hasn’t helped them. So this is a better measure of what airlines aren’t helping their customers (or which airlines have the most irrational customers) than it is a measure of which airlines people have complaints with.
The metric shows the number of complaints per hundred thousand passengers, so for some of these airlines, the numbers are so small that year over year comparisons aren’t great. For example, in January 2008, Atlantic Southeast had .96 complaints per hundred thousand passengers, but this January it plunged to .30 complaints. Hooray! Oh wait. They had a total of 3 complaints for the entire month, down from 9 the year before. That’s such a small number that it’s hard to really draw any conclusions.
The other problem here is how complaints are allocated. If we look at the breakdown, there were 927 total complaints in January. Of those, 109 were related to reservations/ticketing/boarding. It’s going to be hard for Atlantic Southeast to get any complaints in that category since they don’t handle those functions. They have a bunch in baggage as well. What if it’s going from an express flight to a mainline flight? Whose fault is that? If the mainline carriers handles the ground handling for both, then it should be the mainline carrier, but if it’s tied to a flight number, it may fall to the express carrier.
Again, it’s the problem of operational carrier vs marketing carrier that clouds things here. We really need to have everyone show their numbers by marketing carrier to really make things fair.
18 comments on “A Closer Look at January Flight Delays, Lost Bags, and Complaints”
“”””We really need to have everyone show their numbers by marketing carrier to really make things fair.”””””
Agreed, but do you think that will ever happen? It would go against airline nature as then the couldn’t blame the ‘other guy’ like they like to do.
It’s not really the airlines but rather the DOT. They’d have to make the change.
Cranky writes:
“The reason is that when the weather ramps up, smaller airplanes tend to get delayed first because fewer people are inconvenienced than if a big plane gets delayed. But it shows that the numbers you see are certainly deceiving. The DOT needs to change this.”
Another systematic factor at play here is distance — weather forecasts beyond about two hours or so aren’t very precise. Once it’s known with reasonable certainty what kind of capacity reductions an airport has to deal with, a lot of times, the only thing to do is delay flights that haven’t taken off yet. These are generally flights within a short proximity of the affected airport — prime turf for the regional airlines.
Dan-
I thought this might be true also, but I ran some analysis and the results don’t prove it! If you chart % on-time against stage-length distance for delays, you get 0 correlation. Even if you limit to just flights into and out of NYC, you still get nothing even close to statistical (an Rsquared of 0.003, this was just for Oct 09 as a random month I chose to tackle first).
Any ideas of ways to prove this? I agree the relationship exists in some way, but it might be overwhelmed by so many other factors and it doesn’t, in and of itself, mean anything.
Evan
flightcaster.com
blog.flightcaster.com
Evan,
Two things:
1) I oversimplified when I said “distance.” In the real world, this stuff is done on an ARTCC by ARTCC basis. Because ARTCC boundaries are very weirdly shaped, I can see why straight distance might not work. (For instance, if there’s a ground stop at LGA, it may affect all neighboring ARTCC’s. The actual distances involved could very widely.) So a different way to try this might be a logistic regression for flights into NYC, using the originating ARTCC as the independent variable. Flights out of NYC get affected differently, so it’s probably best to treat them separately.
2) You picked October, which is a very quiet time for ATC delays. The types of delays where distance (or ARTCC proximity) becomes a bigger factor is during convective weather season, which we see very little in October. Something in the summer months should provide you with different results.
Thanks for the great analysis… it really goes to show that those numbers carry little weight because there are so many variables, and any carriers can spin those numbers to their advantages (like UA a few weeks go). I am a little curious as to why Virgin America wasn’t included in the list.
Also, I was reading the news about the DL/US slot swap. I would be very interested to hear your views/analysis on the deal, as it sounds quite confusing, and It would be nice to know who benefited the most/least with this deal.
Thanks,
I’m writing about the slot swap tomorrow
Note to Gary Kelly: don’t expect an invite to any airline functions hosted by Dougie and Richard in the near future. ;-)
“With the new rule, this flight would simply have canceled in Harrisburg and some alternate mode of transportation would have had to be arranged, probably a bus.”
…
“In fact, there were a bunch of weather diversions into Charlotte so the gates were all filled and they couldn’t get back to the gate. They now have a procedure set up to get people off on the ramp.”
“[A] procedure set up to get people off on the ramp” is EXACTLY the sort of thing the new rule is setup to force the airlines to implement! (And it is what the airlines had refused to consistently implement before.)
If Delta can get pax off on the ramp in Charlotte now, why not Harrisburg? That fulfills the new rule requirements just fine, with no cancellation necessary.
Hey Cranky, I know it’s a small nit-pick, but could you please link to the site you got the information when possible?
It is linked. Everything came from the Air Travel Consumer Report that is linked at the top of the post.
I’m sorry. You are right. It was just hard to see.
I’m not sure I believe much of what the DOT stats seem to purport regarding operations. The stats showing the number of cancellations and diversions are interesting, but why did the cancellations and diversions occur? No reasons given. Regarding HA, their numbers here seem awfully suspect…very, very low!
On delays, it’s the reasons I’m interested in and who came up with them. There are 5 categories: 1. air carrier delays (maintenance, crew, bag handling, cleaning, etc.), 2. extreme weather (hurricanes, blizzards, etc.), 3. national aviation system (non-extreme weather, such as wind and fog, airport ops, heavy traffic, etc., 4. security issues, and 5. late arriving aircraft, which doesn’t really explain the underlying cause at all.
But, who decides what to call the delay? The airline? Like UA and its web site showing flight status, and those reasons for delay, which I find laughable! And then, HA. Sure seems to have a high percentage of delays related to “air carrier.” Why?
Yes, I know you mentioned Flightstats. OK, “gate change,” “runway,” etc. But, what does this tell me?
GAO reported on some of the DOT data, and delay categories, in its July 15, 2008, study (actually, it’s a written testimony GAO-08-934T. (See http://www.gao.gov/new.items/d08934t.pdf.)
In the end, the delay stats, to me, are not much more than fodder for academics to make generalizations, but not particularly useful for customer fight selection purposes. But, I guess, “the data are interesting/my tax dollars at work!”
Thanks for bringing this to the forefront, Brett – and also Evan. I noticed this was a problem a few years back and think some letter-writing might be in order. I’d like to see something that lists the average performance stats for the brand followed by mainline vs. regional flights, and even operating carrier if possible.
For example:
United (Mainline and regional): 79 percent
– Mainline: 86 percent
– United Express: 72 percent
– SkyWest: 80 percent
– Mesa: 79 percent
– Republic: 81 percent
– GoJet: 82 percent
The DOT stats would be more meaningful that way, and would allow passengers to make better informed decisions.
score another collasal delay by the incompetent managers of American-US Air. Guests standing on a bus for hours waiting to board a DCA U.S. Air flight to Cincinnati (US 3940) with no explanation.
Alitalia has to be worst airline out there, Besides lost luggage being a total nightmare to recover and connecting flights are for magicians since the time span doesn’t give ample time to get to the gate, It’s a total ordeal to talk to customer service in which the call is a pay by the minute call (takes quite a few minutes to hear the recording and you just get to a dead end ) The latest experience is the cherry on the cake, Attempting to do an on-line ck -in and pay for additional luggage resulted the site had didn’t work properly in which every attempt of ck-in came back with an error but at every attempt i was being charged for the luggage ( credit card acc. showed multiple charges for the same thing ) Finally calling customer service the operator didn’t speak no English or Italian.