Fog Creek Software
Discussion Board




Hidden costs of offshoring

http://www.cio.com/archive/090103/money.html

Preaching to the choir here, I assume, but its still a good rundown of many of the things said here.

Alex
Friday, September 19, 2003

Hmmm. Ok let's take this scenario.

Company X is a huge multinational. They BUY or CREATE a company in India that specializes in providing programmers for a specific industry (their own). They push out the current heads of technology and replace them with Indian heads of technology. The bring in the new Indian programmers, and start the layoffs once they prove that it can work.

No more vender selection. Salaries and benefits are for at India rates. Sure Company X provides housing for any programmers that come here, but they pay them Indian salaries, so they choose inexpensive housing, probably from a company they also own.

Most of the things the article speaks of are true for any outsourcing. Certain costs will go up simply because you're doing work in other countries, but that's a cost they're familiar with, being a multinational already. Any profits that go to the vendors really go to themselves. You have to hire middle managers in India to make sure the company is run smoothly, but you're also outsourcing to other companies...

Oh yeah, and you get the government to pass some legislation that makes it easier/cheaper to do business overseas...

Mark T A W .com
Friday, September 19, 2003

True, the few companies that are seeing *real* savings from offshoring generally aren't those who are outsourcing.  It's those doing offshore in-sourcing.

The trouble is, most non-technology companies don't have the know-how to start or run a software company.  Worse yet, starting or running a software company in another country they've never even visited.

Not only that, but it takes time, and the CxO's are predominantly interested in the next quarter's bottom line.  The companies like Citibank, GE, and Motorola who have thousands of employees in their Indian subsidiaries built up to that level over 10-15 years.  But the CxOs want to lay off 1000 people and have them replaced by offshore programmers within the next six months.

T. Norman
Friday, September 19, 2003

A Citi, or a GE can not only build up offshore companies early, they can influence legislation that makes it possible. Citi merged with Travellers *before* it was legal to do so, and were confident the legislation would pass. The ex Citi CEO is influencing foreign policy for asia...

Mark T A W .com
Friday, September 19, 2003

Quite a few people on this board, from the US point of view, are extremely 'off shore'..

i like i
Saturday, September 20, 2003

The truth is that the same way a developer jumps ship for 5$/hr more, so a company will outsource to save 5%, even more so 40%. It seems to me that even if the company saves less than 40%, the savings are pretty decent.

coresi
Saturday, September 20, 2003

Corporations exist to create profits, so that's OK if they're actually obtaining real NET savings of 5% or more.  But if they're saving 40% in direct costs, while having to add back 25% in increased costs for the added management effort and/or on-site consultants to coordinate the remote activities, plus 15% because of delays resulting from the reduced face-to-face communication and company-specific business knowledge, plus 10% from worsened employee morale and the drop in their productivity, plus the initial costs to transfer the work and knowledge over, then they really haven't saved anything.

Similarly, a programmer who moves from Miami to New York City to take a 10% pay raise will almost certainly find that he/she isn't making any more money on a net basis, after considering the higher taxes and cost of living in New York. If they're making the move for personal reasons, that's one thing, but if they're moving for a higher salary they would be sadly mistaken.

Many if not most companies will be sadly mistaken if they think offshoring will save them anything.  But it will continue to happen regardless of the real net savings, because it is a useful tool for creating the *illusion* of savings, since it's easy to focus on the reduction in direct costs and cover up (or don't keep track of) the hidden costs.

T. Norman
Saturday, September 20, 2003

There's a delightful irony in one of the likely consequences of offshoring.

A lot of the people who were happy to sack the "technical people" wil probably get replaced by those technical people, because remote working will greatly increase the importance of technical competence among middle managers and project managers.

Those people could once coast using the technical advice they received from their in-house teams. They won't be able to do that any more.

I would say give it 12 months for this trend to kick in. Boards will start to notice serious problems, and lots of technically trained people will be available for management jobs.

analyst
Saturday, September 20, 2003

Seems to me that there's a lot of conjecture here without much basis in fact... Doesn't seem like anyone here *really* knows what the net savings of offshoring is.

"Well a Munchkin's lifespan is 3 days, so when you offshore the manufacture of lollipops to Oz, there's an increadibly high turnover rate. In fact they have daily courses in how to make them for all of the newborn Munchkins who work there. Therefore, the net savings is 0 due to all of the training."

Sort of like the underpants gnomes in reverse (inverse? converse... I always get these confused):

Step 1. Hire someone to manage the vendors.
Step 2.
Step 3. No Profit!

Mark T A W .com
Sunday, September 21, 2003

Whatever the real net savings are, probably nobody knows because so many costs are hidden, but the savings are certainly less than what the gung-ho offshore proponents say.  The worst ones are those who promote offshoring by stating the low Indian programmer salaries, when their salaries are a minor part of the equation, given that the real estate and infrastructure costs add up to more than the typical programmer salary (another thread here several weeks ago has a link to a forbes.com page showing that office real estate is more expensive in Bombay and New Delhi than in Boston).

The articles about offshoring at cio.com and computerworld.com do provide some factual basis, and acknowledge that if you don't do things right all your savings will be eaten up.

T. Norman
Sunday, September 21, 2003

I spoke to someone who vacationed in India and said that we'd be surprised at the work conditions there. Computers there are terminals with the bare minimum tools they need to get the job done, probably squeezed into any corner available, without even an internet connection.

There are two shifts, to coincide with an American schedule, so their 9-5 and after that our 9-5, their 6-2:30 or so. Factory style scheduling, and twice the bang for your real-estate buck.

Turnover and burnout is high, which means they're driving the workers pretty hard. Each programmer has a quota of items to complete, and if they don't meet it, they get a bad report that week. I don't think you have to guess what happens if they don't meet their quota often enough.

Just saying real estate costs there are on par with ours doesn't tell the whole story. I don't think any of us have the whole story, and while it's fun to piece it together, unless we have all the pieces we can't know - one way or the other.

Why don't we write an article on, "The hidden costs of in-house development," or on the similarities between the assassinations of JFK and Lincoln? I'm sure articles like this exist in "Vendor Magazine."

Mark T A W .com
Sunday, September 21, 2003

Whatever the precise savings are, nobody knows.  But whatever they are, they aren't as enormously compelling as the proponents would like us to believe. Citing real estate costs doesn't nearly tell the whole story, but it is enough to debunk statements from the bandwagon drivers like "I can have ten Indian programmers for the price of one American".

T. Norman
Sunday, September 21, 2003

I believe we just try to pamper ourselves here. Evidently costs in India or East Europe are not as high as Manhattan. In my very rough estimations, the project I’m working on can be done 40% cheaper in East Europe (with which I’m more familiar), including all hidden costs.

And what makes you think there are not hidden costs for American developers as well?

coresi
Sunday, September 21, 2003

Mark,

I really like your simple 3 step plan!

Dennis Atkins
Sunday, September 21, 2003

coresee,

American firms know what the costs of American workers are. So it's not hidden by definition.

Dennis Atkins
Sunday, September 21, 2003

"Turnover and burnout is high, which means they're driving the workers pretty hard."

So what happens when these programmers start to get a little entrepeneurial spirit, and go out on their own outside the big consultancies?  The only coders left at the "sweat shops" will be the ones not competent enough to work for themselves or find a job elsewhere.  If the best developers do this, wages will start to go up and find some equilibrium between offshoring/outsourcing and in-house (or at least on site) development.

Developed world programmers share common cause with developing world programmers; we all want wages to go up.  It's the PHBs, whether American, Indian, French, or Chinese that we strive against.  They want our wages to go down.  We should be organizing with Indian, Phillipino, Chinese, etc. programmers to find ways to make ALL our wages go up.

So how can we encourage bright programmers to get out of the sweat shops and find their true worth in the marketplace?

Jim Rankin
Sunday, September 21, 2003

When you go to a vendor they give you a quote and the project finishes +/- a certain percentange, what hidden costs are there? Since this is outsourcing, one would assume the firm knows all of it's own "hidden costs," so real estate and benefits can't be thrown in as extra costs American's don't know about.

Vendor management is vendor mangement, whether you're talking about outsourcing to a firm down the street or one halfway across the world. How many more warm bodies do you think a firm will throw at a project simply because the vendor is farther away?

Mark T A W .com
Sunday, September 21, 2003

> So what happens when these programmers start to get a little entrepeneurial spirit, and go out on their own outside the big consultancies?  <

America may move away from India when this starts to happen. From what I'm told, programmers there are already beginning to ask for more and more money.

Perhaps you want to add the hidden cost of "jumping ship" every few years when the costs get too high in one region.

Mark T A W .com
Sunday, September 21, 2003

"When you go to a vendor they give you a quote and the project finishes +/- a certain percentange, what hidden costs are there?"

Hidden costs include:
1. The extra efforts spent on the US side to manage things.  Those are not included in the vendor's quote or final invoice, except for the on-site consultants from the vendor.

2. Significantly increased documentation requirements (although this isn't always a bad thing, if they were doing little to no documentation before).

3. Not getting what you really want because the outsourcers don't understand your business like your employees do (the completed project may meet the letter of the contract, but not satisfy the business goal).

4. The adverse effect on employee morale and productivity as you lay off people.

5. Increased risk to security and intellectual property, which will make you either spend money to mitigate it or bear the cost when it bites you.

6. High internal and external turnover within the vendor's employee population reduces the ability to have enhanced productivity in future projects as a result of business and system knowledge increasing with time.

7. Changes in requirements will cost more.  A change to an in-house project could make the project more costly to complete, but sometimes there is no effect on the schedule or cost because you already designed that flexibility into the system.  But a change involving a vendor often escalates to a project renegotiation, and they will charge you for it whether it costs them extra or not.

8. The vendors have very little long-term interest in the maintainability, scalability, or security of the code they develop for you.

True, vendor management is vendor management, but companies are using offshoring as a way of replacing their own employees.  By doing that they add vendor management costs and risks that didn't exist before, which could be high enough to negate whatever savings they were going after.

T. Norman
Sunday, September 21, 2003

"America may move away from India when this starts to happen. From what I'm told, programmers there are already beginning to ask for more and more money."

Yes they are.  Turnover of 30% is typical, and salaries are increasing at 15% a year.

T. Norman
Sunday, September 21, 2003

> True, vendor management is vendor management, but companies are using offshoring as a way of replacing their own employees. <

T. Norman. Good points.

All of those points relate to vender management in general, not offshoring in particular. Sounds like any time someone fires their internal staff in favor of a vendor they incur these costs.

To paraphrase Dennis Atkins, these costs are known to American companies.

> companies are using offshoring as a way of replacing their own employees <

Ah this is interesting. Offshoring as a way to replace in house development teams.

Essentially, you're saying that an in-house IT department will understand the business THAT much more, and will actually be more flexible, understand the requirements better than a vendor, etc. The increased cost then comes from this delta.

Let's talk about your hidden costs.

1. Invoicing, etc. - I would say this is comparable to the HR effort required to manage a large group of programmers.

2. Documentation - In house documentation needs to be less than out of house documentation... simply because of the proximity of the programmers to the requestors? Most large firms will have IT departments that serve several segments of the company, and will have their own requirements process. I'm not sure how big the delta is here between in-house and vendor. I guess your argument boils down to "water cooler chats" and things like this?

3. Understanding your Business - Many IT firms specialize in a particular industry, or in a particular type of software. For example I know of a small firm in NYC that specializes in sweepstakes websites. Are you saying it would be easier for your in-house team to create a sweepstakes website than for a company that specializes in this to create a website for your company?

4. Employee Morale - Assuming, again, that you have an IT department that serves multiple segments of the corporation, firing 20 programmers and hiring a vendor effects the IT department more than it effects the rest of the firm. Also the remaining employees will be under a "rule of terror" and fear the next decimation. They'll probably start being much more accountable.

5. Security risks - This argument can go either way. Accordin to #3, vendors don't understand your core business model as well as in-house guys, so the amount of damage they can do is inherently less than a few internal disgruntled programmers.

6. Turnover rate @ the vendor - This is less about turnover rate at the vendor, which isn't really quantifiable, and a through vetting process might look at the vendor turnover rate. The real argument here is that you're paying full price for updates, where in-house you may be able to sort of fly it under the radar at the water cooler. I'll accept that.

7. Change requests cost more - As with all vendors, change requests usually come with a price tag. Internally change requests come and the man-hours can, again, be slipped under the radar.

8. Long term interest in the code base - Again, this isn't really quantifiable. I know of a company whose programmer turnover rate is pretty high, and they didn't outsource anything. Even simple scripts that could be bought off the shelf were written from scratch. Do you think these 3 year turnover employees had an interest in the long term maintainability of this code? I hear they have code that nobody understands, but they leave it in for legacy purposes.... Just in case.


It seems your assumption is that a small firm where everybody knows everybody, all employees are essentially lifers and has an IT department that lunches with the front office has eliminated the IT department overnight in favor of a nice looking vendor in Bangalor.

Mark T A W .com
Monday, September 22, 2003

Let me reply to your responses:

1. Managing an offshore or outsourced project isn't just about invoicing, unless it is the type of project where you can throw specs over a wall and expect a finished product six months later. Project managers and technical leads will spend extra time coordinating things during the project.  Talk to anyone who has managed an offshore project.

2.  Documentation has to be more detailed because the vendor doesn't understand your business as well as you do (I'm also referring to things specific to your business, not just industry knowledge), so they won't understand what you want well enough or they will exploit ambiguous requirements to their advantage.  Whereas in-house programmers would either know what you really want or ask.

3.  Most offshore firms don't specialize.

4.  The morale of the IT department is what I am talking about in particular. "Rule of terror" doesn't make programmers more productive, they just become more apathetic, especially when the company's actions lead them to believe that they will be replaced anyway no matter how good they do their job.

5.  Sending things offshore generally involves exposing your existing code base and other documents to them, including revealing trade secrets and data, and guarding against back doors.  It is more of a problem with the offshore vendors since you are much less likely to be able to detect when it happens or successfully litigate against them.

6.  Programmers being rotated within the vendor and job hopping reduces their ability to be more productive on future projects of a similar nature, unlike in-house programmers who are with the company for 5-20 years.  Of course, if most of your in-house programmers don't last more than 2 or 3 years this probably won't make a difference.  But if your turnover is that high, you have other problems to think about.

7.  I am referring not to man-hours that get slipped under the radar, but man-hours that don't exist because the requirement can be accomodated with something as simple as changing one constant, or the part of the system that the changed requirement affects hasn't been built yet and the new requirement is only a different way of doing something with equal effort.  If a change can be absorbed without extra effort, the vendor will still find a way to charge you even if it doesn't cost them anything extra to implement it.  They also don't have much incentive to build flexibility into the system, when they can charge you to make the changes either now or in another project.

8.  About long-term interest -- like with #6, if your in-house programmers don't last long anyway, there won't be much difference with the vendor.

Most of what I was saying apply to firms large or small where the average employee stays for 5 or more years.  That is the case where I work.  But if your programmers don't stick around very long, most of the above points don't apply.

T. Norman
Monday, September 22, 2003

I was also assuming that most of the in-house programmers are very competent.  But I realize that often isn't the case, because many non-IT companies suck donkey balls when it comes to hiring technical staff. If your in-house teams have a significant percentage of clods, then there are fewer disadvantages involved with sending the work to a vendor, whether that vendor is local or offshore.

T. Norman
Monday, September 22, 2003

Mark, I think to understand the hidden costs we first need to understand that there are significant hidden benefits in local development. It is those unrecognised benefits that will be destroyed.

For example, all projects undergo improvement during development. They improve in functionality, usability, robustness and business relevance. Some of those improvements are worth an enormous amount to the company.

They arise from all sorts of unrecognised capabilities in having local developers, from their ability to understand cultural factors that might be important in the software, or important in understanding internal communication, and so on.

A lot of the managers who might once have been corrected along the way by their developers will no longer receive that correction. After three or six months, they will get back exactly what they ordered. That's when the fun will start.

Security is a secondary issue, but is probably more important than is being recognised. Organised crime or worse wanting to discover important customer data would find it ridiculously easy to implant a sympathiser into, say, an Indian centre than a New York one.

analyst
Monday, September 22, 2003

I don't think sweat shopping is a good way to develop software no matter where you are.  No matter how cheap labour is, treating people badly doesn't make for good products.  I think everyone from Gates to Buffet to Packard would agree on that.

christopher baus
Monday, September 22, 2003

*  Recent Topics

*  Fog Creek Home