The Analysis of Hot Air

It’s 9:10am on a Saturday morning. As I sit and wait for my LA flight to board, an announcement comes over the loud speaker. “Ladies and Gentlemen. Unfortunately there will be a delay in boarding as the engineers have reported the plane is too hot. We don’t know how long this will take to fix, and apologise for the inconvenience.” Panic ensues: What time will we leave? Will we make our connecting flights? Should we be flying on a faulty plane? Why don’t they just open a window?

With IBM Business Analytics, the story could have been very different.


6:00am: Before many passengers are even out of bed, the A380 lands at the airport and engineers start the system check, entering in test results and documenting observations as they go.

6:04am: The engineer types in a comment about that the auxiliary power unit is generating a lot of heat.

6:05am: A reading from the auxiliary unit travels outside of “normal” operating limits.

6:06am: An automated decision stream is triggered, running analysis of the auxiliary unit hardware. Based on historical patterns and trends with similar hardware performance, combined with the textual information that was entered by the engineer, the system identifies that this unit is 83% likely to fail in the next 24 hours.

6:07am: An automated alert is sent to the engineer on the tarmac as well as the operational manager, informing them of the finding. An optimisation process is also started to evaluate the best location to fix the unit based on scheduled maintenance, airport facilities, and parts inventory.

6:08am: The operational manager is informed that replacing the unit in LA would provide maximum availability of the aircraft and minimise cost. The system also recommends he get external fans to the aircraft before 6:30am – in the event the unit fails, it will take three hours to cool the interior of the plane to flight temperature.

6:15am: The external fans arrive. Passengers are on their way to the airport.

7:30am: The auxiliary unit fails. The external fans are deployed.

9:00am: The plane is cool enough to fly, but still uncomfortably warm. The ground host checks her passenger analytics report on her iPad and can identify 15 passengers who are “high value” customers, as well as 5 additional passengers with strong social networks – some of which are connected to high value customers. She selects these 20 passengers for a complimentary upgrade and apologises for the warm conditions.

9:10am: “Ladies and Gentlemen. Welcome to flight QF93, please board at your leisure.


Okay, so clearly I’m not an aircraft engineer, but hopefully you get my point. Gone are the days where airlines have a 14 hour flight to win back your loyalty. Customers have access to immediate outlets in the form of social media – in the space of a 20 minute delay, thousands of tweets and status updates can be blasted with negative consequences for a brand. Not to mention the downstream impact and cost of flights missing their scheduled departure.

Utilising the SPSS platform, airlines can better understand their customers and automate business decisions to serve them better, whilst still achieving their performance goals.

Wouldn’t life be easier if everyone employed IBM Business Analytics to automate their business decisions?