Quote:
Originally posted by 1967 R50/2
Even alot of the companies that were sending work to computer mills in India have now decided it's more economic to keep the work here.
|
Took them long enough to figure it out.
Let's say I offshore a simple coding change--something that would take me 25 hours to code, test, and implement. Roughly one work week.
First I have to have an hour meeting with and offshore coordinator. Let's say the meeting takes one hour. That meeting has burned 2 man-hours.
Now I have to spec out the change. It takes me one hour.
Offshore says the change will take 10 hours. (See--we saved money!) Only 13 hours burned--so far.
I get the code back a week later from offshore that has allegedly been coded and unit tested per my specs. It doesn't compile. I have another meeting with the offshore coordinator and some other offshore QA guy and re-explain the specs. I am promised the working code tomorrow morning. 3 more hours burned.
Tommorow morning, and no code. 2 days of pestering the offshore coordinator and I get that the code will be ready Monday of next week.
Monday of the next week, and I have the new code, and it compiles. *Just to make sure* I decide to run some simple unit test cases...and the code doesn't do what I speced out. Not wanting to have to wait *another* week, I go ahead and correct the code (which doens't meet coding standards). Correcting the code and running unit test cases takes me 15 hours. Additional levels of testing take another 5 hours.
So, I predicted it would take me 25 hours over one work week to make the change myself, and offshore promised me 10 hours of their time plus 5 hours of additional testing for a total of 15 over one work week, and one day, for offshoring. It really took 35 man-hours over two weeks, one day, to offshore.
But the offshore labor costs less, so despite the innefficiency, the company still saved money, right?
Well, lets say I make $30 an hour. If I made the change myself over 25 hours, that would cost the company $750.
Lets say the offshore programmers cost $10 an hour. The offshore coordinators also make $30 an hour. If they made the changes as promised, it would have cost 10 hours of offshore programming time at $10 an hour for $100. Add that to a 2 man-hour meeting at $30, an hour of spec writing, and 5 hours of additional testing for a SUBtotal of $240. That brings the total estimated cost for offshoring to $340.
Per my (often-repeated, real-life) scenario above, there was an additional 18 hours of $30 an hour time. That is an additional $540 of cost. That means it really cost $880. So what, it cost $130 more than it would have if I made the change?
The problem is that every single time I have been forced to send work offshore, this scenario has happened. Sometimes, I've had to send work back 3-4 times before I finally give in and correct it myself. The code NEVER meets standards.
Are Indians bad programmers? No. But they don't have the business knowledge of nor the repeated exposure to the systems I work on to make quick coding changes that meet business needs and coding standards. The offshore guys usually only work on a given system once before moving on to something else.
More an more, the business rules that a run a business are encoded in that company's computer systems. Is it really a good idea to farm out a company's proprietary secrets to the lowest bidder?