Outsourcing to India
Discussion
Anyone here outsourced software development? If so I am interested in your opinion regarding:
- scale
- value
- quality
- do’s
- don’ts
I am experienced in outsourcing to UK companies so am more looking for a comparison rather than an idiot’s guide. I will be spending about 100k GBP on UK prices so it will be quite a large project.
I am also looking for any recommendations regarding who to use.
Any help appreciated but please, only comment if you have real world experience.
- scale
- value
- quality
- do’s
- don’ts
I am experienced in outsourcing to UK companies so am more looking for a comparison rather than an idiot’s guide. I will be spending about 100k GBP on UK prices so it will be quite a large project.
I am also looking for any recommendations regarding who to use.
Any help appreciated but please, only comment if you have real world experience.
Yes, done it - for a couple of very large blue-chips.
If it's piecemeal work - truly monotonous template creation/data entry/etc. fair enough. But, as a wholesale offshoring, there are only two core reasons:
- If you can't get the staff in the UK
- If you could get the staff, you have nowhere to sit them
It will cost about 30-35% more than doing the same work in the UK (in the long term). Hard to boil this down into a quick post, but the closest I can get here is:
- In the UK, you're not just paying for typists, you're paying for sanity checking. Specs are generally so imprecise, they need detail questioning at every stage. You won't get this with outsourcing to India - the culture is so different around asking questions of your "orders" you'll just get exactly what was asked for. Granted, it'll be about 50-60% of the cost in the UK, but it won't be what you wanted. So you'll end up having to have it largely rewritten.
- Whilst developer day-rates are cheap out there, you'll pay for a LOT more managers. And locally based ones, too - to get things done on time we've invariably ended up having to man-mark offshored managers with UK ones. (At UK rates)
- Difficulty in communication. Partially due to time-zones (5% impact), partially due to distance and inability to just "wander over" (10%), and partially due to nuances of understanding/context (5%)
- Difficulty in "touching" the product. Not so much a problem as a cost to bear in ensuring there's a slick demoing & feedback capability
If IT development is core to your business, I'd recommend against it, unless you're trying to restructure a company ready to carve up and sell on.
Most of those comments are based on something approximating a £1bn spend. At £100k, you have different issues - setup costs will be higher, for instance, but scope for HORRIFIC divergence from spec will be lower.
If you can structure the specification documents in a LOT more rigid fashion than normal, then you're on firmer footing. For instance, if you have bulleted requirements, with acceptance tests beneath each, along with screen mock-ups, details of the exact logic, the reports, etc. you'll be getting closer. (Basically, the acceptance criteria that would be needed for test-driven-development/automated acceptance testing)
Feel free to drop me a mail through my profile and we can go through this in much more detail - might well be able to help out here.
If it's piecemeal work - truly monotonous template creation/data entry/etc. fair enough. But, as a wholesale offshoring, there are only two core reasons:
- If you can't get the staff in the UK
- If you could get the staff, you have nowhere to sit them
It will cost about 30-35% more than doing the same work in the UK (in the long term). Hard to boil this down into a quick post, but the closest I can get here is:
- In the UK, you're not just paying for typists, you're paying for sanity checking. Specs are generally so imprecise, they need detail questioning at every stage. You won't get this with outsourcing to India - the culture is so different around asking questions of your "orders" you'll just get exactly what was asked for. Granted, it'll be about 50-60% of the cost in the UK, but it won't be what you wanted. So you'll end up having to have it largely rewritten.
- Whilst developer day-rates are cheap out there, you'll pay for a LOT more managers. And locally based ones, too - to get things done on time we've invariably ended up having to man-mark offshored managers with UK ones. (At UK rates)
- Difficulty in communication. Partially due to time-zones (5% impact), partially due to distance and inability to just "wander over" (10%), and partially due to nuances of understanding/context (5%)
- Difficulty in "touching" the product. Not so much a problem as a cost to bear in ensuring there's a slick demoing & feedback capability
If IT development is core to your business, I'd recommend against it, unless you're trying to restructure a company ready to carve up and sell on.
Most of those comments are based on something approximating a £1bn spend. At £100k, you have different issues - setup costs will be higher, for instance, but scope for HORRIFIC divergence from spec will be lower.
If you can structure the specification documents in a LOT more rigid fashion than normal, then you're on firmer footing. For instance, if you have bulleted requirements, with acceptance tests beneath each, along with screen mock-ups, details of the exact logic, the reports, etc. you'll be getting closer. (Basically, the acceptance criteria that would be needed for test-driven-development/automated acceptance testing)
Feel free to drop me a mail through my profile and we can go through this in much more detail - might well be able to help out here.
Edited by J_S_G on Monday 15th January 01:29
If you really want to go the Indian route, and from my experience I'd strongly advise you not to then you should split your budget £20K for Indian programmers and £80K for experienced onsite (in India) project management.
I don't really want to say too much, things might have changed now, my experience with outsourcing to India was nothing short of a nightmare.
I've PM'd my phone number to you, feel free to give me a call and I'll happily discuss my reasons.
Alternatively have you considered Romnian programmers instead? I hear very good reports about their skills and dedication.
best
Ex
I don't really want to say too much, things might have changed now, my experience with outsourcing to India was nothing short of a nightmare.
I've PM'd my phone number to you, feel free to give me a call and I'll happily discuss my reasons.
Alternatively have you considered Romnian programmers instead? I hear very good reports about their skills and dedication.
best
Ex
Been there got the T shirt. We outsourced an e-business web product. One of the main issues, was getting the Indian dev team to understand the product. Specs have to be spot on. I would agree with having a good project manager onsite, who the UK can talk to.
The standard of the work was very good though. We looked at TATA and a smaller company called ITC. In the end I went for ITC as they were a slightly smaller company and had more of a can do attitude.
The standard of the work was very good though. We looked at TATA and a smaller company called ITC. In the end I went for ITC as they were a slightly smaller company and had more of a can do attitude.
zumbruk said:
TheExcession said:
J_S_G said:
A very informative post
Have to agree with everything J_S_G has posted.
Me too.
And much as I hate to say "I told you so" to the people at work whose brilliant idea this was, "I told you so!"
And me. I've seen the results first hand, and this was with a project that actually shipped some of them over here for a while at the start of the project.
Don't not touch with a bhaji pole in my book.
picking up on the exact spec thing watch out for context/common forms - we had one field which was spec'd to allow a "Null Entry" as well as some other entry (e.g ABC123). What we got back was a field into which you could only type "Null" or "ABC123".
Duh!
Change Request here we go!
Miles
Duh!
Change Request here we go!
Miles
Picking up on this I have first hand knowledge of a couple of outsourcing "instances". In neither case were the end customers happy.
You simply cannot overvalue having a consistent working culture with your developers and the ability to get them in the same room as your acceptance test staff and end users.
We get approached *all the time* with offers to outsource our software development. The answer is always no way. We understand our business and our product. We'll keep it developed right here in the UK and continue to be able to offer same day fixes and patches to our customers. They pay for a level of support that an outsourced business simply couldn't match.
You simply cannot overvalue having a consistent working culture with your developers and the ability to get them in the same room as your acceptance test staff and end users.
We get approached *all the time* with offers to outsource our software development. The answer is always no way. We understand our business and our product. We'll keep it developed right here in the UK and continue to be able to offer same day fixes and patches to our customers. They pay for a level of support that an outsourced business simply couldn't match.
J_S_G said:
(Basically, the acceptance criteria that would be needed for test-driven-development/automated acceptance testing)
That's just crazy talk, nobody does that...
I'd pretty much agree with JS's appraisal. It'll cost more, be more stressful to work with and everyone will hate you, apart from that, it's a go-er.
trooperiziz said:
J_S_G said:
(Basically, the acceptance criteria that would be needed for test-driven-development/automated acceptance testing)
That's just crazy talk, nobody does that...
I'd pretty much agree with JS's appraisal. It'll cost more, be more stressful to work with and everyone will hate you, apart from that, it's a go-er.
My experience is that systems I've been involved with have been outsourced, the results have been less than successful, then I've picked up work getting the clients back to where they were.
I've run across the "exact spec" problem in various places too. Another common problem is the "overpoliteness" one, where you will get agreement and apparent understanding of an issue, and later on in the project you'll find out that they didn't have a clue what you were talking about, but would rather agree than argue (and therefore develop their understanding). I know it's wrong to generalise, but I have seen this on more than one project, so it's not an isolated problem.
Edited by victormeldrew on Monday 15th January 14:33
victormeldrew said:
trooperiziz said:
J_S_G said:
(Basically, the acceptance criteria that would be needed for test-driven-development/automated acceptance testing)
That's just crazy talk, nobody does that...
I'd pretty much agree with JS's appraisal. It'll cost more, be more stressful to work with and everyone will hate you, apart from that, it's a go-er.
My experience is that systems I've been involved with have been outsourced, the results have been less than successful, then I've picked up work getting the clients back to where they were.
I've run across the "exact spec" problem in various places too. Another common problem is the "overpoliteness" one, where you will get agreement and apparent understanding of an issue, and later on in the project you'll find out that they didn't have a clue what you were talking about, but would rather agree than argue (and therefore develop their understanding). I know it's wrong to generalise, but I have seen this on more than one project, so it's not an isolated problem.
Edited by victormeldrew on Monday 15th January 14:33
We found that an easy way to piss off your UK workforce was to ignore them when they said don't outsource, when it comes back in a none working state that doesn't do anything in the spec.... Go back to the developers that said don't do it and tell them to fix it before the go live deadline.... Result, a bodged bit of code that works, but only just and a UK workforce that is stressed and pissed off cause they have just had to pull some major overtime to get the product working, but because it went in on time, Outsourcing is seen as the idea way to go cause it costs less.
Go figure....
And no I hate it for all of the reasons mentioned above.
ProPlus said:
victormeldrew said:
trooperiziz said:
J_S_G said:
(Basically, the acceptance criteria that would be needed for test-driven-development/automated acceptance testing)
That's just crazy talk, nobody does that...
I'd pretty much agree with JS's appraisal. It'll cost more, be more stressful to work with and everyone will hate you, apart from that, it's a go-er.
My experience is that systems I've been involved with have been outsourced, the results have been less than successful, then I've picked up work getting the clients back to where they were.
I've run across the "exact spec" problem in various places too. Another common problem is the "overpoliteness" one, where you will get agreement and apparent understanding of an issue, and later on in the project you'll find out that they didn't have a clue what you were talking about, but would rather agree than argue (and therefore develop their understanding). I know it's wrong to generalise, but I have seen this on more than one project, so it's not an isolated problem.
Edited by victormeldrew on Monday 15th January 14:33
We found that an easy way to piss off your UK workforce was to ignore them when they said don't outsource, when it comes back in a none working state that doesn't do anything in the spec.... Go back to the developers that said don't do it and tell them to fix it before the go live deadline.... Result, a bodged bit of code that works, but only just and a UK workforce that is stressed and pissed off cause they have just had to pull some major overtime to get the product working, but because it went in on time, Outsourcing is seen as the idea way to go cause it costs less.
Go figure....
And no I hate it for all of the reasons mentioned above.
Yup, we had a similar situation at a company I used to work for, business managers thought it looked great on paper until we saw the results.
The UK engineers had to spend a significant amount of time reviewing the work performed by the counterparts in India. In the end many walked out (including myself) after the fact that if we had done the interesting part of the work here we wouldn't have spent 3 times as long on the project and of course management give us grief for not having it completed on time, when it was their f**k up in the first place! RANT OVER.
annodomini2 said:
ProPlus said:
victormeldrew said:
trooperiziz said:
J_S_G said:
(Basically, the acceptance criteria that would be needed for test-driven-development/automated acceptance testing)
That's just crazy talk, nobody does that...
I'd pretty much agree with JS's appraisal. It'll cost more, be more stressful to work with and everyone will hate you, apart from that, it's a go-er.
My experience is that systems I've been involved with have been outsourced, the results have been less than successful, then I've picked up work getting the clients back to where they were.
I've run across the "exact spec" problem in various places too. Another common problem is the "overpoliteness" one, where you will get agreement and apparent understanding of an issue, and later on in the project you'll find out that they didn't have a clue what you were talking about, but would rather agree than argue (and therefore develop their understanding). I know it's wrong to generalise, but I have seen this on more than one project, so it's not an isolated problem.
Edited by victormeldrew on Monday 15th January 14:33
We found that an easy way to piss off your UK workforce was to ignore them when they said don't outsource, when it comes back in a none working state that doesn't do anything in the spec.... Go back to the developers that said don't do it and tell them to fix it before the go live deadline.... Result, a bodged bit of code that works, but only just and a UK workforce that is stressed and pissed off cause they have just had to pull some major overtime to get the product working, but because it went in on time, Outsourcing is seen as the idea way to go cause it costs less.
Go figure....
And no I hate it for all of the reasons mentioned above.
Yup, we had a similar situation at a company I used to work for, business managers thought it looked great on paper until we saw the results.
The UK engineers had to spend a significant amount of time reviewing the work performed by the counterparts in India. In the end many walked out (including myself) after the fact that if we had done the interesting part of the work here we wouldn't have spent 3 times as long on the project and of course management give us grief for not having it completed on time, when it was their f**k up in the first place! RANT OVER.
I found a solution to that.... move into an industry that is in realtime, people say jump and you do, there and then!!!
They couldn't afford for it to
up so no outsourcing is done or even thought about (too much money at risk and we find it complicated enough), except for answering the telephones and we get all the fun stuff. Fantastic stuff.
Oft heard saying:
If you send a job off to India, give them the rough outline of what you want and let them loose. The project will be completed on-time, under or on budget and the communications with the people working will be good. You will, however be left with a box when you wanted a 3 bed semi.
If you send a job off to Russia, give them the rough outline of what you want and let them loose. The project will be late, over budget and they'll utterly ignore you're attempts to find out what is happening. You will, however, be left with a hand built castle that has hot and cold running hookers, when you wanted a 3 bed semi.
What the others have said about the advantages / disadvantages is pretty much true. My, perpetual NDA with IBM stops me from quoting chapter and verse about the budget increase moving USF2 support to Bangalore, however... keep in mind that if you absolutely must outsource (neither local talent or imports will suffice) then their are other countries in the world that might better suit your requirements. India just has a very good PR department.
If you send a job off to India, give them the rough outline of what you want and let them loose. The project will be completed on-time, under or on budget and the communications with the people working will be good. You will, however be left with a box when you wanted a 3 bed semi.
If you send a job off to Russia, give them the rough outline of what you want and let them loose. The project will be late, over budget and they'll utterly ignore you're attempts to find out what is happening. You will, however, be left with a hand built castle that has hot and cold running hookers, when you wanted a 3 bed semi.
What the others have said about the advantages / disadvantages is pretty much true. My, perpetual NDA with IBM stops me from quoting chapter and verse about the budget increase moving USF2 support to Bangalore, however... keep in mind that if you absolutely must outsource (neither local talent or imports will suffice) then their are other countries in the world that might better suit your requirements. India just has a very good PR department.
ProPlus said:
I found a solution to that.... move into an industry that is in realtime, people say jump and you do, there and then!!!
They couldn't afford for it to
up so no outsourcing is done or even thought about (too much money at risk and we find it complicated enough), except for answering the telephones and we get all the fun stuff. Fantastic stuff.
I assume you mean Embedded Real-Time?
It was! Automotive infact!
Gassing Station | Computers, Gadgets & Stuff | Top of Page | What's New | My Stuff