The chronic susceptibility of businesses to IT solutions that overpromise and underdeliver was illustrated by Qantas today, when the Amadeus system that it was using to check in passengers wrecked its schedules for more than 12 hours.
The Amadeus suite of airline, travel agent and individual on-line booking applications are in fact widely regarded as world leading. But when they fail all hell breaks loose, and such failures have paralysed operations in recent times abroad as well as in Australia.
The flight delays will cost Qantas dearly, but is this part self-inflicted damage, reflecting the willingness of businesses in general to submerge common sense evaluation of the claims of IT providers to the mentality of uncritical acceptance of overambitious claims and the lowest-priced deal that can be taken to a board for approval?
And do airlines in general try to do things with their IT solutions that push the limits of practicality. Or, put more bluntly, do they suffer geeks gladly, and pay for it later.
Past bad experiences with airline booking sites near and far has encouraged this traveller to record, in graphic file format all airline bookings using a screen-grabber. It takes a few seconds to make one and drop it into a folder-or at the least, just send the booking code (a short set of upper-case letters and numbers) to your smart phone so you can put it in front of any check-in person who claims that you or it don’t exist.
In an unrelated IT failure over the holiday weekend thousands of retailers using Bank of Queensland EFTPOS terminals lost serious bucks when the system insisted it was January 2016, thus rejecting as invalid almost every credit or debit card presented because they were only valid up to expiry dates no later than 2011.
Maybe these events point to a need for companies to select IT services on the basis of how little they promise rather than how much. They could come out well ahead.
4 Comments
I’ve been in commercial IT for over 24 years now and your comments re: airlines & their IT are right on. The same bullshit management decisions that lead to failing IT systems are also reflected in the entire 787 program. Some idiot gets the idea that they can do the impossible and no one has the balls to tell them they’re wrong.
Look at the Telstra Transformation project – any 1st year Comp Sci kid could have foreseen that it was a disaster waiting to happen.
It’s a product of people in high positions pinning their bonuses on impossible goals and throwing their weight around while surrounded by sycophantic toadies and/or “scared of my mortgage” managers who hope to still be employed when the shit hits the fan. Throw in consultants & outsourcers who stand to make a packet and voila, you’re on a path to problems.
I don’t see any easy solutions to the problem as it takes effort, diligence and a healthy application of cynicism to the grandiose plans & predictions from on high.
Cheers,
Grant
The solution is simple….redundancy of infrastructure. Critical systems should have redundancy in depth so that one server failure should not result in outages of this kind. Costs money but is cheaper than paying the support guys overtime/on-call rates when the bloody things die over the holidays.
I agree with Grant, but have a slightly different take. I’ve worked on a number of large software projects.
There seems to be a phase in each project (usually 80% complete) where the Architect leaves for greener pastures without having any of their structure truly tested (cynically I think they choose technology based on how it will look on their resume). Without the Architect parrying and blocking the “this doesn’t work” message from below, it takes about a month to reach a stage where, there is a “How can we save project/ever finish?” meeting and a real plan put together.
The plan works if two things happen:
1) There is enough money left; and
2) The project management team stays.
When hiring an architect, always call their previous employer and ask whether they would use them again.