Preface: I work in the IT business.
Transitioning between 3 completely different proprietary systems is a massive undertaking. Given the importance and critical nature of the system i think two years is actually pretty impressive.
As mentioned in the previous post by Adrian, its not just a website and booking system they're changing. It will effect the entire operations of essentially two complete airlines, not to mention the additional systems (load control, air craft movement etc) as well as partner airline systems. DJ has gone through a somewhat typical growth stage with their IT infrastructure, similar to most businesses. Starting with something small and barebones that does the essentials - they had a tight budget, and i'm sure some things were just easier to manage on paper than having some complicated IT system. They've then grown dramatically and wanted to offer new features, however i would suggest that typical of most businesses, IT had the minimal funding possible to create the best system in the world. Consequently alot of things were done on the cheap. Proprietary systems were developed and specific programs were created to do one or two simple things.
To explain what i mean, i'm translating a typical IT project example into an airline context.
Someone would have gone gee its getting difficult to estimate how many drinks we should be loading onto a particular flight. They would have gone to the IT department and said "IT Nerd fix this problem and you have a budget of $2.50 to do so". Now the simplest thing would be too go to the vendor of the current software package to see if this is available as an add on, however in typical software fashion the "drinks management" package is part of the advanced airline management package which is just a 10 squillion dollar upgrade and offers everything from carpet cleaning management through to drinking straw management.. so this would mean spending 10 squillion for a system that your only going to use 2% of. So boy wonder the IT guy pulls some late nights and with lots of head scratching develops a simple little program that pulls some information from their existing system, and with some fancy maths prints out a simple drinks usage estimate for each flight. Problem solvered! the IT nerd is an absolute genius - he's found a solution to the problem, and it only cost them a few hours of his time. Now 6 months down the track, some other smart guy notes that for every can of diet coke they sell, they use 2 cups, 1 serviette and 1 straw and thinks if only we had a way to take the information from the drinks management system and we can easily know how many cups, serviettes and straws we need. Same process is repeated and after a few late nights there's another seperate program that talks to the drinks system and now produces a report for stock ordering. Rinse and repeat this process many times over and you can begin to understand how these things develop.
Now as time goes on these programs have become ingrained in the way things are done. Times gone past, staff have moved on and now not many actually no why certain things are done this way, but its just the way things are done. The systems not exactly elegant and crashes regularly - but for the most part the reports are produced and business carries on (lets ignore the fact that IT now spends majority of it's time putting out fires - aka fixing things that randomly break for no reason other than someone typed three letters instead of two in a particular line). Oh sure every now and again someone looks at the drink management package and goes this is silly why doesn't it do this, or why isn't it part of our main system. There's some yelling and screaming, some gnashing of teeth and crunching of numbers, money gets spent on doing "feasibility studies", knowledgeable consultants with long beards are brought in to design solutions, and then its put on the way too hard pile and forgotten about.
So this has gone on for a few years, finally they have decided to spend some money as part of their new business model to move to a single platform that does everything nicely. They've had the company sales reps in who've shown pretty powerpoint presentations and promised silver spoons for everyones mouths.
Documents are signed and the project begins... (this is where the fun starts!)
The new wonder software's drink management system is brilliant, but it's pretty much useless without the historical information from the existing system. However the existing system was written years ago in the equivalent language of paintings on the cave wall. To make things worse it was written by the first settlers who didn't actually understand the caveman wall painting language. Add to this the many generations of people who haven't understood this language, yet have had to make changes to the system or fix particular problems that have crept up and the whole things pretty much a mess. 3 months go past with software monkeys typing away trying to develop a system that will talk in caveman to the old system and retrieve information in a way that can be easily put into the new system. Finally after many revisions and lots of head shakes they finally have something cobled together that brings most of the information out - however some will still need to be manually entered. (To save getting into too much detail, we'll skip over the period where a change is made in the old drinks system which is still being used daily at this point that completely breaks the system to export the data..)
In the blink of an eye 12 months has gone past, progress has been made, things have been cobbled together to allow information to be exported out of the existing programs and imported into the new program. Clear well defined scopes of who's responsible for doing what have been written, the projects been given a cool code name complete with fancy logo and custom t-shirts for the project team.
Now its time for the presentation of the system to a small test group. Usually a small number of staff and they have say 4 weeks to play with the new system.
So the testing period closes and feedback forms are sent out to the test group. Everyone except the IT department (who actually understand the importance of testing!) fails to send back the forms. After a few phone calls it's discovered the testing group has been so busy managing their day to day positions they've had next to no time to even login to the new system, let alone have a look at it. "some gnashing of teeth and yelling/screaming occurs" and finally another 4 week test period is set with strict instructions that the testing staff must actually test it and give feedback.
Another 4 weeks passes and sure enough the feedback begins to come through - "it doesn't look like the old system, i clicked button A in the new program but it does something different to button A in the old program - must be broken..." Hopefully however there's been some constructive feedback received and there's another period of development followed by another round of testing.
(for those keeping track - there's another 3-6 months gone!)
Finally at long last we're almost at the point where we can start rolling out the new system. It's usually around here that some massive critical bug or issue is discovered. You know the one along the lines of "what do you mean we can't connect our existing website to the new system" or "what do you mean the existing computers we have in every terminal and office worldwide won't run the new program..". In typical fashion, someone's head - usually the person who pointed these issues out in the initial meetings gets cut off and someone new who has no knowledge of whats been done so far is brought in to fix all the shortcomings of the previous person.
We're now 20 months into the project and it's here the real fun begins. We start developing training modules for the staff to complete and sending out training teams to offices around the world training everyone on how the new system will work. There's lots of resentment as staff realise just how much everything's going to change, and how much 'harder' their jobs will become given they will have to use four letter acronyms for the airports rather than the three letter ones they've always used, and that in the drinks management system a can of coke will have the item code "CanOfCoke" instead of "283749293cddsc" meaning they have to relearn all the new codes.
It's generally about here that the projects 18 trillion dollars over budget, can't run on the existing IT systems and the staff hate it that it's cancelled completely and business continues to run on the old systems for ever and a day.
Assuming you've made it this far.. fingers crossed DJ complete their migration to sabre with as few problems as possible. I expect once they complete the move we will see some great improvements to how their websites interact together and everything will be a great improvement. That said - i personally don't have any major issues with either the velocity site or the main virgin site!
