James Fallows (Free Flight) writes about DayJet’s on-demand taxis in the sky:
During the good times, [Ed Iacobucci's] wife, Nancy, had started an air-charter business called Wingedfoot. It was a “traditional” charter business — that is, multimillion-dollar aircraft hired for very high fees by corporations or tycoons. By that time, Iacobucci was following the same aviation-world chatter I was, about the coming of small jets. The difference between a writer and an entrepreneur is this: the only thing I could figure out to do was write a book about it; by January 2002, Iacobucci and his wife had raised several million dollars from friends, family, and tech-industry investors to start work on a new air-taxi company. (They raised some $16 million from investors, in two rounds, from 2001 through 2006; then another $50 million in early 2007; then $140 million in debt financing, to buy airplanes, about a year ago.) They code-named the company Jetson Systems, which they wisely changed to DayJet when they announced their plans in 2003.
Here are some of the essential reasons that the new business was not as outlandish as the original name might have made it seem.
First, the airplanes. Eclipse, founded by another software-industry veteran who had worked with Iacobucci named Vern Raburn, promised to deliver fast, small jets for about $1 million apiece — versus five to 10 times that much for Gulfstreams, Falcons, Learjets, etc. — and to build them with advanced, Toyota-style lean-manufacturing techniques that would make them dramatically more reliable than current versions. Their efficient engines would also make them cheaper to operate, bringing the overall cost per mile of jet travel to a small fraction of the private-jet level. The price of an Eclipse has risen to about $1.5 million, but it is still much cheaper than alternative jets.
Next, the airports. Although unnoticed by most travelers, the United States is studded with airports, at least 3,500 of which have runways large enough to allow small, Eclipse-style jets to take off and land. The real value of these airports is that most of them stand nearly empty and could handle many more airplanes carrying people to and from the suburbs, office developments, factories, or recreation sites nearby. Most were built during an aviation boom that ended 50 years ago, and have barely been modernized since then. Bruce Holmes’s efforts at NASA included exploring ways to make them all usable, even in bad weather — especially with safer, modern GPS-based landing systems.
And finally, the airlines. Despite the temporary dip in air travel after 9/11, compared with 2000, airlines now serve fewer cities, with less-frequent flights and often with smaller airplanes — but carry more people overall. And although a larger share of flights go through overcrowded hubs such as O’Hare, Dallas/Fort Worth, and Atlanta, those airports rarely add new runways. This has the same effect as trying to force more cars onto a given road: all of them slow down. Once in the sky, jets are as fast as ever, but cascading delays mean that the overall door-to-door speed of U.S. airline travel has been slowing down.
The “ant farmers” call it an “Internet for stuff”:
Jim Herriott and Bruce Sawhill, computer scientists in their 40s, are the ant farmers. [...] Their job has been to determine exactly how many people might pay to use an air taxi, and where they would want to go. Their answers have come through ant farming, which could less colorfully be called inductive reasoning.
For instance, to predict how many Floridians would pay to fly from Pensacola to Naples, they start not by gathering gross-travel or population figures but by trying to simulate the decisions that hundreds of thousands of individual travelers will make. Their computer models resemble a much more complex version of an “artificial life” computerized game like SimCity or SimLife — or, to explain the nickname they gave themselves, programs that simulate the paths a colony of ants will take across a floor as they discover and retrieve pieces of food. This process is also known as “agent-based modeling.” The ants, or agents, in DayJet’s model are the 500,000 people per day in the seven southeastern states who take business trips of 100 miles or more. Some 80 percent of these trips are now made by car. Commercial airlines account for most of the rest, with trains, buses, charter flights, etc. making up the remainder. (In the Northeast, commercial airlines represent less of the total, and trains more.)
[...]
The DayJet model factors in all relevant variables that could affect the traveler’s decision — something that is hard enough for a real person in real time. It contains up-to-date listings of all flights offered by all commercial airlines serving the region, and the prices for short-term bookings and seven- or 14-day advance-purchase fares. It has average highway-speed and congestion data for the routes people would drive between any two cities, and real-world travel time from different parts of a city to the nearest airport. It includes lodging and restaurant costs, if a driving trip means an overnight stay, and rental-car and gas rates.
Also, crucially, it tries to place some value on people’s time.
Fascinating, but I can see such a bottom-up simulation going very, very wrong.
Anyway, DayJet noticed that a shift in airline policy was helping it:
As airlines “upgraded” the equipment they used for small cities — sleek-looking 50-seat regional jets in place of clunky 13-seat turboprops — many of them had to downgrade their service. This shift was bad news for small cities in many ways. For instance, Lakeland, where I went on my DayJet flight, had with federal and state funding built a new $6.8 million terminal in 2002. The investment was based on the assumption that a feeder airline would add Lakeland to its schedule — probably Comair, a Delta subsidiary, which had served Lakeland with small planes in the 1980s. In fact, no airlines came, and the expensive terminal stood empty — until the first Dayjet flight arrived last October. Cities that could keep small turboprops full might generate too little traffic for larger regional jets, but plenty to sustain an air-taxi business.
The ant-farmers are just one of the two tech teams. The other team is composed of Russian mathematicians who have devised a complex real-time scheduling and pricing system:
On the Web site, you say where you’d like to go — to Naples, from Tallahassee — and when. Then comes the crucial part: specifying how flexible you are about your travel plans. If the flight itself takes just under two hours versus seven hours of driving (the site tells you how long the flight will take), and you have exactly two hours in which you’d like to travel, you say: “Can’t leave till 2 p.m., must arrive at 4 p.m.” After only a few seconds, the system gives you a quote, in this case DayJet’s highest rate: $1,296. But if you are free to travel any time that day as long as you get there by dinnertime, you enter: “Could leave as early as 11 a.m. but must arrive by 6 p.m.” In that case, the system comes back with a quote about one-quarter as high: $346. Airline fares for this route on Orbitz or Travelocity are usually higher, and the trip always takes longer, because there are no nonstops.
If you accept, the trip is booked — and the night before your trip, you get an e-mail specifying your exact departure time, meaning you won’t have to devote your entire “travel window” to traveling. If the e-mail says to get to the airport by 3 p.m., the plane will be there waiting for you. All you do at the airport is show your ID at a counter and walk onto the plane. If you have specified a wide-enough travel window to get a lower price, there may be at most one intermediate stop to drop off or pick up another passenger. Combining trips this way, in a familiar SuperShuttle model, is the key to DayJet’s per-seat, on-demand service, which keeps prices well below what they would be for chartering an entire plane.
In the few seconds it takes DayJet to price your trip, a system called RTR (for “real-time routing”) is figuring out how your request will affect the placement of planes, pilots, and passengers for all other flights that day, and exactly how much the company must charge to make a profit on your flight. The mechanics of making all of this work are what have made the Russians famous within the company — along with a vast computing system called ASTRO, which runs round the clock, constantly looking for more efficient ways to combine planes, pilots, and the time windows requested by passengers.
Between 6 and 7 p.m. each evening, the computers “gelatinize” the assignments for the next day — make them firm enough to tell passengers exactly when to show up at the airports, but still pliable enough that pilots and planes might be reassigned if last-minute requests come in. From the passenger’s point of view, everything is truly set by this point. (Passengers can change or cancel their flights until 6 p.m. the previous night, for a $100 fee. Between 6 p.m. and two hours before flight time, the cancellation fee is 50 percent of the fare. Within two hours of flight time, passengers who cancel must forfeit the full fare.) But until an hour or two before each flight, the company may not be sure which airplane, with which team of pilots from which other destination, will be making the trip. These last-minute assignments are left to ASTRO, which keeps track of the unbelievably intricate technical, legal, and human variables required to meet the promised schedule. (For instance: not simply which pilots are due for rest time but also which pilots weigh how much, so that two fat ones won’t lead to an overloaded flight.)
I told Brad Noe, the vice president for software development, that the only place I had seen something similar was in a logistics center tracking the flow of electronic products from factories in China, through U.S. air couriers, to delivery in the United States. “This is a logistics company,” he replied, “that happens to be moving people.”
(I’ve discussed DayJet before.)
China Stocks, Once Frothy, Fall by Half In Six Months — after increasing six-fold over the last couple years: