Technology is not an Island

airline flight schedules on flat screen televisions

The Wall Street Journal reported an interesting piece of news on the ongoing merger between United Airlines and Continental.

Consolidating technology systems is very common in mergers and it looks like from a technology perspective United and Continental had no issues.

‘the changeover from the old United Airlines system to the existing Continental Airlines system went as planned’

What is very common also is that although the technology itself had no issues the consequences on other parts of the business seem to have caught United Continental by surprise.

‘but the shift led to operational headaches over the weekend such as late flight departures and arrivals, missed connections, problems at check-in kiosks, long lines and extended wait times to reach reservation agents’

Not a good outcome for a business which depends on all these activities to work right in order to transport people. Given the competitive nature of the industry this kind of trouble can turn off customers who otherwise would be loyal to United Continental.

‘According to, a website that analyzes flight data, United subsidiary flights shower poor departure and arrival punctuality Saturday.’

One of the more important metrics that airlines are judged is the on time arrival/departure performance. United Continental may have had a successful ‘technology’ deployment but the risks to its operations have illustrated the importance of cross functional collaboration in mergers.

The technology was to put both airlines on a

unified platform for booking tickets, checking in and boarding passengers and managing the now-combined frequent-flier program.

Notice every activity above is a customer interaction that is critical.

Three things that may have helped in this situation are:

1- Very thorough real time simulation. I would be surprised if United Continental didn’t do this but what I do wonder about is if the simulation included the peripheral activities. A technology system generally has inputs, some calculation, and output. However before the input and after the output there are other elements such as the human element that can throw things apart. A real time simulation would also have to be as close to reality as absolutely possible and run long enough to capture performance issues.

2- A gradual roll out of the system. This is probably very complicated in airline systems but if doable it is a good way of figuring out the kinks before a major roll out. This is also why modular software systems are smart way to build systems. Simulation is suppose to help but nothing is like a system in production with all the bells and whistles that surround it.

3- A back up plan. It would be great to have a on-off switch to turn back to a working system but I suspect that would have been quite the challenge. However it would be a good capability to create.

Leave a Reply

Your email address will not be published.