We have been developing our On-Time database and wanted to share an insight with our readers. Our source data is the US DoT On-Time data. The charts below are for the three big airlines at JFK in New York. We show taxi out time because that is where the real delays seem to be.
Let’s start with a terminal map of JFK. 
American is located in terminal 8, jetBlue is located in terminal 6. Delta is located in terminals 2, 3 and 4.
The layout of JFK shows how old it is. Its a sort of patchwork and any recent visitor to the airport can testify to how old many terminals look.
But we want to address the airside of the terminals for the big three airlines at the airport.
The next image shows the runway layouts at JFK. Knowing where each airline departs from provides a reader an idea of the amount of taxi time involved. It appears that none of the three big airlines at JFK is particularly disadvantaged in terms of access to runways.
As you can see Delta Air Lines is really significantly slower in taxi time compared to its competitors. Its 2009 taxi times were considerably slower than American or jetBlue.
In 2010 Delta seemed to show some improvement while the other two show much improved taxi times. American and jetBlue flights seem to work their way around the airport much faster. 
Early on in 2011 Delta did much better than in previous years but remained behind its main competitors.
Indeed when we look at Delta’s taxi times over the period to see how much they have improved; Delta is doing better but not much. 
We looked at Delta’s taxi times before. All three big JFK airlines had a tough 2009. Since then American and jetBlue made good progress. Comparing performance improvements at JFK for American and jetBlue look like this.
Delta seems to find JFK a tough place to taxi – the sooner it completes its terminal rebuild the better. The following chart shows the average taxis time for these airlines at JFK.
In 2009 Delta was 18.1% slower than the combined average. In 2010 Delta was 21.2% slower than the combined average. To date (through August) Delta is once again 18.2% behind the combined average. For winter 2011 jetBlue has 145 daily departures from JFK, American has 90. Delta averages 154 flights departing JFK every day – and each of these flights averages about five minutes more taxi time than its competitors which translates into 12.8 “lost” operational hours every day.
Running an airline is complex and the way to stay operationally “on-target” is to stick to schedule . In an 2010 an interview with USAirway’s COO Robert Isom (Isom_04_28_2010) we learned how crucial sticking to schedule can be. Anything that allows an airline to shave off schedule minutes means savings go straight to the bottom line – given revenues are set by the market, running an airline successfully is mainly about relentlessly cutting costs. At JFK Delta is incurring operational costs it should cut, and its peers seem to be better at taxi times (and the consequent costs).
Good data mining of DOT information.
If you break this out by orientation of runways in operation, you’ll find the 31L departing scenario is a significant challenge for Delta due to the present T-2/3 ramp locations, not so much for the others. Using the 4-22 combination is pretty much even for all.
Great insight – thanks
This issue has nothing to do with airport orientation because there is one major missing piece of the puzzle: time of day.
JFK has the worst delays during the evening “international” push because desirable departure times are in the same range for both Europe and Asia due to geography. Take a look at JFK at about 5:30pm or so and you will see an international flight preparing for departure at nearly every gate while most of those gates are empty mid-day.
Guess when most of the Delta flights depart because they are international departures? Meanwhile, JetBlue was only granted slots during the off-peak hours when JetBlue started flying back in ’99 and later acquired some evening slots. Those mid-day zero delay departures are averaged with JetBlue’s evening departures and thus the appearance of shorter taxi times.
And no, I’m not a consultant. 😉
You have some good data and responses. You are missing several significant pieces of the puzzle. For examples, ATC playbook for dominate and sub-optimal scenarios, relative fleet mix (weights of aircraft including turning ability and engine/s shut-down for taxi, company procedurtes for push-back and block time recording, sticky surface and required trust, etc,). In total these factors can collectively add minutes to one’s taxi times.
Suggest you do the analysis on the playbook and fleet mixes and start there looking for variance.
Thanks,
Bud Bensel