Airlines and Big Data - Looking Within to Find the New Frontier

Much like The Cloud is marketing speak for "someone else's computer", Big Data is the latest marketing speak for "a database".  Simply put, it's nothing new.  As always been the case, though, information is valuable, especially if sought out and implemented properly.  Big Data is everywhere and the airline industry, with it's largely virgin data-rich ground, is a prime candidate to realize the benefits of exploiting the data that already exists.  Only that hasn't really happened.  Since Big Data at airlines has become synonymous with ancillary revenues, the true power of the information owned by an airline has gone largely untapped.  Ancillary revenues go up, airlines claim success, commission checks go out to the third party provider, and the Big Data project is put on autopilot and forgotten.

Where we inevitably end our search for applicable analytics at the airlines is precisely where the best airlines are beginning.  While most are stopping at customer behaviors, others are starting to use Big Data to better handle operations and partner relationships.  Some are taking the step beyond customer decision-making to improve the entire customer experience from considering a trip, to boarding the aircraft, to picking up their bags, to how that affects their consideration of their next trip.

The challenge isn't collecting the data.  The challenge is asking the right questions.

Yet, all of this information already exists at any airline, no matter how large or small.  There is no third party data required, only the will and ingenuity to find the information the airline already owns.  The challenge isn't collecting the data.  The challenge is asking the right questions.  How much more would a customer be willing to pay for an aircraft with a nicer interior?  Do your customers want more legroom or wider seats?  Better yet, which are they willing to pay for and how much?  Does a customer's buying habits changed after a cancelled flight or missed connection?  How do aircraft gate assignments affect the chance of a missed passenger connection?  Is there a correlation between how far back you sit in an aircraft and the chance to miss a connection? (hint:  I'll bet there is)

We are quietly starting to see real advancements with airlines beginning to utilize the information assets they already have.

Now consider the value of this information.  What if you could save just ten miss-connects a day?  That would be 3,650 people a year who avoided an interrupted trip being stuck in a city they never intended to be.   At a very conservative estimate of $100 per miss-connected passenger, the analytics group would have just paid for itself with $365k in benefit per year.  In reality, that single discovery would return several million dollars per year to the airline, and other discoveries and benefits would continue.

Information is important, and it's much deeper than the merchandising level, where most airlines stop.  We are quietly starting to see real advancements with airlines beginning to utilize the information assets they already have.  I have yet to meet an airline who does not see the value in information, so what is setting some airlines apart?  Why do some airlines utterly fail in understanding the invaluable information they already have, while others have pulled years ahead?

The driving force behind an analytics team should be discovery, not necessarily finding answers.

The "Go find something interesting" department

My first piece advice when listening to an airline discuss their new Big Data project is to stop calling it a Big Data project.  It is a clear symptom of taking too big of a bite, without knowing what it is they are biting into.  I use my four-year-old's advanced logic here:  if Dad asks you to close your eyes and taste something, you may get lucky, then again it may be broccoli.  The first time it was broccoli was the last time we played the game.  So often the first few times an analytics team brings bad news (or worse yet no news) it gets shut down.  While being focused on results is clearly a positive influence in our industry, it can easily be destructive here.  The driving force behind an analytics team should be discovery, not necessarily finding answers.  What you are searching for is correlation.  Find something interesting, and look everywhere.  Hypothesize, test, test some more, and just as importantly...

It's OK to be wrong

I'm wrong a lot.  I have come up with countless hypotheses for data that should correlate, only to find there is nothing.  Yet knowing there is no correlation is valuable in itself.  I've spent hours trying to show that passengers would rather fly in the aircraft my company builds than our competitor's.  Since there is a direct correlation between seat share and market share, my company's aircraft should be producing a higher market share per seat share.  People should be going out of their way to fly our amazing airplane... only they're not.  Hours down the drain looking for an answer that didn't exist.  But that information is still incredibly valuable to my organization.  Instead of resting our laurels on our prettier airplane (the value of which we could not prove), we used this information to focus on improving what we know an airline's passengers will appreciate:  legroom.  With the help of this information my company was able to focus it's engineering resources into a new interior that gave every passenger an extra inch of legroom over the competition.  Not bad for being wrong.

Here lies the majority of failed airline Big Data projects.

Death by "Why do you need this?"

The single greatest threat to any effective analytics team is the silo.  Each set of information valuable to an airline lies in a different part of the airline.  Take, for example, a team who is looking to quantify the value of a particular seat over another during supplier negotiations.  They would need survey data from marketing to know how a passenger rated their experience, reservations data to know which flights and dates the passenger flew, schedule data from operations to know which specific aircraft flew that passengers segment, refurbishment data from tech ops to know which seats were installed how long ago, and seat-specific information from supply chain to know the version or age of a seat, as well as how expensive the seat was to purchase.  That's five different groups, all in control of their own data who need to be willing to release that information to the analytics team.  Not coincidentally, that's also five different times the analytics team hears the question "Why do you need this?"  Here lies the majority of failed airline Big Data projects.  By far the greatest challenge for any airline to achieve truly meaningful results from Big Data is to simply allow itself access to it's own information, yet I've only seen complete success once.  The solution is simple:  this has to be a priority all the way to the top.  In any organization, there is always one person who represents all parties involved, and that person needs to create the culture of cooperation.  Whether that be the CEO or lower, their commitment to the project has to be clear.  As simple as it may sound, millions of dollars are being lost to the inability to read data the airline already owns because alignment across the organization was not achieved.

The new winners on the Big Data frontier are those willing to get out of their own way to see what's already there.

Big Data is much more than a third party selling their point of sale information.  As the airline industry begins to address this first layer of hired information, the true value remains with the data already within the airline.  Like being unable to see the forest from the trees, an airline is often too close to see the treasure already within their control.  The new winners on the Big Data frontier are those willing to get out of their own way to see what's already there.