Your folks estimated it would take our people 1500 hours, beginning to end. The indians estimated 2000 hours. They have now burned thru the entire budget, and have not yet had a design doc approved. They're asking for 1500 more hours to finish the 'requirements gathering'.
Your folks would have done it -- front to finish -- in 1500 hours at 75,000/annum or $36/hr, at a total of $54,086.
The Indians work cheaper, at 15,000/annum, or $7.21/hr. They have already asked for 3500, JUST ON REQUIREMENTS GATHERING. Call coding another 3000 and support/deployment/bug fixes another 3500. Thats 10000 at $7.21, or $72,100.
Let's see: Americans get it done at $54,086. Indians get it done at $72,100.
Oh, yes, we must select the Indians.
And it's worse than that.
Even if they did save a little money . . . it would have taken us about 4 to 6 months, beginning to end.
It looks like it's going to take them a year and a half or so.
Time-to-market in the software world is critical. If a product or tool takes an additional *year* using them, even if you do save a little money (and I don't think you do), is that really a 'success'?
I've an interesting update on this story -- today was judgement day for Satyam. Because of this out-of-control project, we are cancelling our contract with them.
And the VP in question's solution for this project?
He's hired another Indian subcontractor, and is going to start the project all over again, with them.
Failure didn't teach him one single thing.
But it's a step. They've finally admitted that bunch of Indians failed. It's one first step.