
"The commission said its new system will incorporate Oracle's Siebel customer relationship management system, Tata's BaNCS banking system and products from Experian, Genysys, IBM and Oracle"
COTS over custom built is no longer the panacea. It has taken two decades, but technology illiterate business types have started to discover that COTS projects where more than 20% (or even less) of the system has to be "customized" do not represent any savings from custom built apps.
The reasons are simple. In one hand, those 20% of customizations are usually done by fighting to bend some part of the product that was not intended for the purpose. The cost of working around the package is simply higher than to write from scratch.
In the other hand, modern app platforms provide already a good set of frameworks. DIY no longer needs to bear with the limitations of the development staff in the knowledge of a lot of different standards and technologies, they can leverage a lot of already built components that provide out of the box a lot of functionality and at the same time, produce an exact fit for purpose.
And let's not even start talking about maintenance costs. Those that believe that COTS is cheaper to support or maintain have forgotten that the 20% that is "customized", not being part of the standard package, has the same resource requirements as if it were a custom built component. And it's going to be that part that is going to eat the 80% of the support cost.
Oh, yes, I see a continuous profit stream for Tata just for the cost of interfacing all those things together and for continually keep changing the standard packages. But I'd like to see the business logic behind the decision.