Arnie Weissmann
Bad weather is creating havoc at O'Hare, and a monitor shows that your flight to LaGuardia is delayed for more than an hour. But you also notice that a flight to Kennedy on the same airline is scheduled to take off on time, even though its scheduled departure is a few minutes after yours.
If you're a frequent flyer, you know there could be myriad reasons why. Maybe your equipment is delayed coming in. Or the crew on the other flight is in danger of timing out. Your flight may have a minor mechanical issue that needs attention. Perhaps your flight is being held for the benefit of connecting passengers whose inbound flights are delayed, or, conversely, the on-time, Kennedy-bound departure has several passengers who might otherwise miss overseas connections.
Yes, all of those factors can come into play, but thanks to recently deployed technology, the real reason might be that the collective loyalty status of flyers on the other flight is greater than the collective status on yours. You are delayed because, in one way or another, the passengers on your flight are not as important to the airline as people on the other flight.
Or perhaps those passengers are a combination of important, needy and whiny. The latest wrinkle in the way that Big Data intersects with the travel industry is the inclusion of a Flight Value Score that automates an airline's analysis of dynamic events that impact scheduling.
And for airlines that use it, that score determines who will fly and who will wait.
The formula takes crew "legality" and subsequent duties into account as well as the length of the trip and anticipated total passenger delay in minutes.
Individual passenger data and "engagement history" are scrutinized in detail. How many passengers on each potentially delayed or canceled flight have requested extra assistance? How many onboard hold elite status?
How many of the passengers on various flights vying for on-time departure respond well to the airline's promotions? How many chronic complainers are onboard? How many passengers have recently been on flights that were delayed 45 minutes or more?
Even the number of poor souls who have been frequently subjected to middle seats in the recent past are part of the calculus.
"Airlines have to serve all customers equally well, but there is a hierarchy within that," Arnab Gupta told me. Gupta is the CEO of Opera Solutions, a Big Data analytics company. Gupta likened Opera Solutions to a janitorial company that cleans up data, doing tedious and messy, but necessary, work.
His company's Flight Value Score has been deployed by two large airlines, although only one of them, British Airways, has authorized him to reveal its name. Opera Solutions works with a BA program called Know Me, which uses customer data to craft quick responses to service issues.
Gupta said that previously, when service was disrupted, the rule of thumb to resequence flights was "first come, first served." Delayed planes departed in the order they were originally scheduled to take off. Authorization for exceptions was decided on a case-by-case basis and only after review by operations personnel.
And the primary concerns of operations managers were airport-focused rather than customer-based, Gupta said.
So, what do we think of this?
One could argue that, traditionally, there have been only losers when disruptions occurred. Everyone was delayed equally.
Under the emerging system, there will be clear winners and losers. The system favors an airline's most important customers but also works to maximize the number of winners by factoring in a preference for those who have already endured recent flight hardships. Or whose onward trips would become nightmarish. Or who have already indicated they need special assistance.
And it reduces the burden on airline personnel who might otherwise be required to address potential problems later.
To some extent, however, these benefits are counterbalanced by those unlucky individuals who need extra assistance, who will miss connections they would have made under first come, first served and who will sit down in their middle seat even more grumpily because the people on their plane don't, collectively, have enough juice to achieve priority.
I would think that as the score criteria become more widely known, websites will crop up that will try to analyze and forecast what flights are likely to have the right balance of factors to possibly mitigate the risk of delays or cancellation.
Travel agents, online or off, can certainly take advantage of some assumptions. Since this is largely a numbers game -- airlines are trying to irritate as few people as possible -- it's a safe bet that the larger the plane, the more likely it is to have a simple majority of "important" people. Likewise, there will be higher concentrations of elite-status loyalty members on flights departing to and from an airline's hub airport.
It's hard to know whether to look at Opera Solutions' Flight Value Score as a program that primarily brings a level of personalization to a system that had previously been indifferent to individual circumstances or one that basically ignores individual needs in favor of the privileged few and automated decision-making.
The egalitarian in me is drawn to the inherent fairness of first come, first served, and I'm wary of how my personal data is used, but today's data mining can also be viewed as a de facto form of passive democracy: With the Flight Value Score, the majority always wins, and Gupta says customer satisfaction numbers show "an enormous uptick" where it's deployed.
If that's true, there's no turning back. It would appear that, in more than one sense, the janitors of data will be cleaning up for some time to come.