I’ve been thinking—wondering, really—about airfares after we all start flying again. Until ten or fifteen years ago, caprice is the word that came to mind when I pondered how airlines set airfares. They didn’t always make sense. Airline capacity modeling on which fares were based still then involved a goodly amount of human judgment to tweak rules in the calculating software.
That was prior to today’s fare calculations using complex algorithms. A.I.-based airfare formulas now routinely optimize profit margins based on highly sensitive demand and supply rules and models free of most human override that take into account every variable imaginable, including city-pair routing, time of year, day of week, hour of day, airplane type and model, fuel and other variable cost factors, airport reliability factors, competitive factors, direct versus connection options, same-flight booking history, proximity to flight date, customer stratification as booked real-time, weather and wind patterns, and on and on and on.
Setting airfares became a highly specialized and refined software science thanks to big advances in sophisticated computing capabilities coupled to statistically predictable variables. That predictive accuracy relied on the law of large numbers fed by data derived from the histories of thousands of daily flights.
Before the COVID-19 shutdown, the number of planes worldwide in the air at any given time averaged close to 10,000, carrying 1.3 million people. In late January, 2020, the number of daily commercial flights was about 112,000 according to flightradar24 stats, with load factors hovering around 80%. Lots of large numbers there to grease the wheels of the airfare software.
However, worldwide daily commercial flights dropped to around 30,000 by the end of April with near zero load factors. That’s a 65% reduction in flights, pretty dramatic, but it’s the load factor drop of 90+% that throws a monkey wrench into accurate statistics. The law of large numbers doesn’t work without large numbers, after all.
Sans the certainty of sharply accurate predictive modeling, pegging fares is like juggling bowling balls in the dark. Which makes me think that airfares after the coronavirus crisis will be in flux until demand returns to January, 2020 levels. And of course no one knows when, or even if, that will happen.
Out of curiosity, I made a cursory search of current pandemic-era fares in a few markets out of RDU. For instance, on Kayak and Orbitz for RDU/SEA in May, June, July, August, and October, every day I tried came up the same pre-pandemic fare levels of $330-370 or so in coach round trip in that market except in May when it was twice or three times as much.
Why such a premium for near-date flights when all load factors are virtually nil (I write this in late April)? I can only assume it’s because some airline airfare algorithm insensitive to reality is jacking up prices for any flight that is soon, a software decision based on the long-held airline conviction that if I have to fly tomorrow, I’m going to by-god pay a huge premium. Even if the flight would otherwise go empty without my butt in a seat.
All domestic test fares came up with similar results. No need to look at international fares since most overseas flights, and even to Canada, are now forbidden.
As things begin to morph back to normal, will airfares be higher or lower? I wish I knew. No one does yet. So we wait.
Nothing I can do about it right now, so I baked salmon and asparagus for the family.