What are you gonna do where you're outsourced?

[Paul C. Anagnostopoulos]

I'm beginning to see friends and associates losing programming work to outsourcing. It's going to India. I've been out of the mainstream software business for seven years, so it's not affecting me.

However, I'm also seeing book production moseying off to India, Taiwan, and other countries. It's taking a little longer because of the language barrier, but that barrier will fall. I figure there will be no book production in the US in ten years.

Is this affecting you? What are you going to do? Do you care? Shall we have a huge retirement party at TAM2?

~~ Paul

[epepke]

Quote:

Originally posted by Paul C. Anagnostopoulos
I'm beginning to see friends and associates losing programming work to outsourcing. It's going to India. I've been out of the mainstream software business for seven years, so it's not affecting me.

Well, I'm out of work, not because of oursourcing, but outsourcing is probably one of the reasons I haven't found work.

First, I think it's obvious that outsourcing is a really stupid idea for certain classes of tasks, and software development is one of them. But the results don't come back until two or three years down the road. I'm starting to see some results from people who have been bitten by outsourcing.

I'm predicting that by 2005, outsourcing will be widely recognized as causing more problems than it's worth. Then I think there will be quite the revolution in thought. For me, at least, the main problem is staying alive and keeping my spirits up until then. But I think I won't have any ethical qualms left over about fleecing those who now think that outsourcing is a great idea for all they're worth.

[Paul C. Anagnostopoulos]

But when you're paying $20/hour for programming instead of $150/hour, you can stand getting bit in the butt. Also, all those programmers in other countries aren't stupid; they will learn how to write better software.

In the book business, I charge perhaps $10--14 per page to typeset a book. The compositors in India charge $4. No one even bothers to ask me if I want to compete with that price.

~~ Paul

[epepke]

Quote:

Originally posted by Paul C. Anagnostopoulos
But when you're paying $20/hour for programming instead of $150/hour, you can stand getting bit in the butt.

Not when it's $20 an hour for something that you have to throw away, especially if getting bit in the butt means losing several million dollars, which is not uncommon for even moderately sized companies. Besides, $150/hour is an exaggeration or an abberation of the Internet bubble. It's more like $80/hour for a high-quality senior developer, including overhead. Consultants charge more, because they're available immediately and don't work all the time.

Quote:

Also, all those programmers in other countries aren't stupid; they will learn how to write better software.

Why does everyone assume that the idea that saying outsourcing is bad means that people in other countries are stupid?

Look, when I write software, here's what I do. I go talk to the people who are going to use it. I get them to tell me what they want and write it all down, and then I largely ignore it except for a first step for figuring out what they need. I watch them work and look at their body language. I try to get them to forget that I'm a software developer and talk about the business problems. Then I put together some requirements, knowing full well that at least half of them will change. Then I put together some stuff in prototypes and try it out. Usually, it isn't what they wanted, and the more it is like what they said they wanted, the less it is like what they wanted. Then I do lateral thinking and try out some new things.

I don't care how smart someone is, there is just no way that this is possible if you are on another continent.

What people get when they outsource software is either something that either a) follows the specs so closely that it is useless, or b) fills in the business logic with assumptions from a completely different business culture operating under a different set of laws.

This is in fact what people are starting to report as the results of some outsourced projects.

This doesn't even get into the niceties of user interface design. A text message that is fine in one business context may seem brusque and rude in another, and if you change it for the latter context, it may become unsuitable for the former.
Screen layout may be influenced by where the users are used to looking on the screen for that kind of information, maybe even on the layout of the paper forms that they use. Overall organization of the program is influenced by the workflow. Even for something as seemingly simple as data entry, it depends heavily on whether the head accountant or whoever likes to enter the information as the papers come in or save them up and do them as a batch.

I could go on and on (and probably have), but you get the point.

[reprise]

Why should anyone give more of a damn about high-paying jobs being exported offshore than they did when low-paying work was being sent offshore? People seemed to think it was a great idea then because it cut the cost of production which meant that consumers could buy stuff cheaper, companies could make greater profits, and shareholders got bigger returns on their investment.

The low-paid workers who lost their jobs to offshore labour were told to stop whining and reskill (and hurry up about it because it wasn't the state's responsibility to support them).

Welcome to the realities of a global economy, instant data transmission, and transnational corporations.

[Zep]

I happen to think that epepke has hit the nail on the head. While there are indeed SOME jobs that CAN be successfully outsourced, others simply cannot if you look at just one side of the equation.

Take epepke's software development task, for example. Not only all the issues described there, but when there ARE problems at the user end, the company loses income while users are not doing the job they are being paid to do. And they are still being paid to be there to do nothing while they wait, so the company loses twice when that happens. Now multiply that effect by how many people are idle, and then again by how many hours they are not working, then AGAIN by their hourly pay rate...the dollars add up MIGHTY quick! And that has a direct and disasterous effect on the almighty bottom-line, for which the performance of same may cause some people to be VERY answerable...

But these issues were thrashed out decades ago (I know, I was around the first time!) so it's hardly a new problem, nor are the solutions new either (insource vital developments). It just seems too obvious that some corporate barons can only see the upside cost benefits, yet it seems absurd to me that they would remain unaware of the drastic downside costs. Perhaps they plan on making the bucks, taking the bonuses, and bugging out before the shower-of-sh*t starts!

[reprise]

Zep, one reason why I dislike the trend towards share ownership in Australia is because so many of those who are now demanding (and indeed depending on) the companies in which they have invested return ever-increasing amounts to shareholders are utterly unable to foresee the inevitable consequences of those demands until those consequences affect them as either consumers or employees.

[Zep]

I agree wholeheartedly.

I made one of those "when I'm grown up and rich" decisions a long time ago: Any company I owned would NEVER be floated on the stockmarket. I don't care about the prospects of more investment, I just don't like to gamble with employee's livelihoods. They MAKE the company work better.

Of course, as to whether I would ever be in the position to enact this, well ... time will tell!

[epepke]

Quote:

Originally posted by reprise
Why should anyone give more of a damn about high-paying jobs being exported offshore than they did when low-paying work was being sent offshore? People seemed to think it was a great idea then because it cut the cost of production which meant that consumers could buy stuff cheaper, companies could make greater profits, and shareholders got bigger returns on their investment.

Not true. I drive a US-made car for just this reason. (And, yes, I did check out to see what percentage of the parts of the car were made in the US.) I also own one of the only make of VCR made in the US.

The trouble is, during the 1980's and 1990's, buying US-made-goods was an effective way of being branded some ultraconservative right-wing lunatic. I could never understand the logic behind this, but there you have it; all of my liberal friends were tooling around in their Toyotas (some of which had "Save the whales! Boycott Japanese products!" bumper stickers on them).

[reprise]

I always think shareholders in banks and telcos have no right to complain when the fees increase and the level of service declines - after all, it's being done in the name of improving the return on their investment.

[epepke]

Quote:

Originally posted by Zep
It just seems too obvious that some corporate barons can only see the upside cost benefits, yet it seems absurd to me that they would remain unaware of the drastic downside costs.

I think one major problem is the fungible career CEO. When I worked at one place for a little more than two years, there were three separate CEOs. But that, I think, like the people who called themselves programmers but weren't and who caused a great deal of damage to the craft, was a particular disease of the bubble of the 90's. It's going to take a while, but I think that in time there will be a shakeup at that level, too.

[shanek]

Quote:

Originally posted by epepke
Look, when I write software, here's what I do. I go talk to the people who are going to use it. I get them to tell me what they want and write it all down, and then I largely ignore it except for a first step for figuring out what they need. I watch them work and look at their body language. I try to get them to forget that I'm a software developer and talk about the business problems. Then I put together some requirements, knowing full well that at least half of them will change. Then I put together some stuff in prototypes and try it out. Usually, it isn't what they wanted, and the more it is like what they said they wanted, the less it is like what they wanted. Then I do lateral thinking and try out some new things.

I don't care how smart someone is, there is just no way that this is possible if you are on another continent.

I'm going to agree with this. Our development process very much involves the client. We are constantly submitting questions for them to answer and every so often, usually a week, we show them what we have done so far and, if it's consistent with what they want, they sign off on it and we bill them for the work done up to that point, then continue on. Our clients really like that model.

[Paul C. Anagnostopoulos]

So you have two highly-paid systems guys to chat with the customers, watch their body language, and mutter "the customer is always right." Then these two guys explain to the overseas programmers what to do.

I think you guys are good software engineers, which means that you think management gives a damn about the quality of the software. Well, they do, but only up to a point that is complexly correlated with the bottom line.

Epepke, no one is assuming that overseas programmers are stupid. But that doesn't mean they won't get even smarter with time.

I hope you're right and it does blow up. We shall see.

~~ Paul

[epepke]

Quote:

Originally posted by Paul C. Anagnostopoulos So you have two highly-paid systems guys to chat with the customers, watch their body language, and mutter "the customer is always right." Then these two guys explain to the overseas programmers what to do.

I guess I shouldn't laugh too hard, because, after all, you've been away from software development for a number of years. But this approach to software design is largely a thing of the past. Some shrinkwrap companies still apply it, but the majority of programming in the US in in-house.

The lower-level programmers that can be told what to do on a project generally don't exist. There are sometimes lower-level programmers. They get easier, but still self-contained tasks requiring just as much direct contact with the stakeholders. When multiple people do work on the same project, it's usually organized more like sous chefs in a French kitchen. You've got the guy who really knows how to make a killer Bechamel sauce, and so when the order comes in for something with Bechamel sauce, he's the guy.

There are good reasons for this, which I tried to explain but obviously didn't succed. (Ironically, this is a self-referential lesson.) Some horrible percentage of software projects, like 80%, fail under the old model. One reason is that the design specs, no matter how carefully thought out, planned, and reviewed, are always wrong in significant ways. But these ways often do not become apparent until such time that, according to the traditional waterfall model, it is too late to change them. So, they have largely been replaced by agile software development.

This has had auxiliary benefits. Computer scientists have long been pushing for appropriately modular systems. However, this largely did not happen until the needs of software development and the adoption of agile techniques forced them to happen. The results of this, such as the organization of large enterprise systems, have been mostly to the good.

Outsourcing forces people back to the ways that don't work.

Quote:

I think you guys are good software engineers, which means that you think management gives a damn about the quality of the software. Well, they do, but only up to a point that is complexly correlated with the bottom line.

No, quite the opposite. I'm very cynical about it. I don't have any illusions that management considers quality software per se to be a goal. I would quibble with the point that they necessarily care so much about the bottom line; in my experience, what they mostly care about is monkey games and their own careers.

But again, this applies mostly to shrinkwrap software development. There are other managerial problems with in-house development. But when the goal of the business is to do business, not just write software and charge people to make calls to your understaffed (and probably outsourced) help desk, the effectiveness of the software to get the job done does at some point become visible to management.

Quote:

Epepke, no one is assuming that overseas programmers are stupid.

You were the one admonishing me that they weren't.

Quote:

I hope you're right and it does blow up. We shall see.

Indeed. It's a prediction, and like all predictions, it will either come to pass or it won't.

[CFLarsen]

History repeats itself.

This was also the craze in the late 80's, too (at least in my whereabouts). Jobs were deemed to move to India, oh, my, what to do, panic, panic.

I am sure that Indian programmers are just as good as European ones, but building large computer systems is not like assembling a Ford T. It is far more "organic", with lots of interaction between the people who build it. There are so many obstacles, two of them being different timezones and lack of face-time within the project.

We do not communicate well by email. When we need to explain a problem, the far most effective way is to sit in front of each other, doodling on paper or blackboards (whiteboards seem the rage, but hey...). Ideas are created far more efficiently this way.

I can explain a problem a lot better, if I am able to draw two boxes with an arrow going from one to the other. Then, people understand me. If I write an email (and have to wait for HOURS, sometimes until the next DAY!), the chance of misunderstanding is much higher.

Meetings are impossible, if the participants are timezones away. Video conferencing? Don't make me laugh....

It's a fluke. It will go away again, when enough companies have wasted enough billions on enough failed projects. We see so many of these management fads, and it all reverts to the ol' proven way of developing systems. I have had so many bosses, but only a few who understood this. We produced way above the possible, when those guys were "in charge". The most efficient bosses describe the problem in detail, then go to their offices and play Solitaire, leaving the work to us. They only emerge if there is a problem that absolutely needs their decision. Not very often, it turns out.

[shanek]

Quote:

Originally posted by Paul C. Anagnostopoulos
So you have two highly-paid systems guys to chat with the customers, watch their body language, and mutter "the customer is always right." Then these two guys explain to the overseas programmers what to do.

I think on large projects that would work, but with smaller projects it just wouldn't be cost-effective.

[Paul C. Anagnostopoulos]

Epepke said:

Quote:

I guess I shouldn't laugh too hard, because, after all, you've been away from software development for a number of years. But this approach to software design is largely a thing of the past. Some shrinkwrap companies still apply it, but the majority of programming in the US in in-house.

So you're telling me that some large percentage of the people working on custom applications or in new start-ups are talking to the customers? How do they get any work done?

The venture capitalists must be out of touch with this. Ravi Chiruvolu (Charter Venture Capital) recently said "Right when you think about Employee 11, you should think about India. My view is you should not start a company from scratch in the US ever again." I'll ask my Dad; he's a VC guy.

Quote:

You were the one admonishing me that they weren't [stupid].

Saying "so-and-so isn't stupid" is just a figure of speech.

~~ Paul

[epepke]

Quote:

Originally posted by Paul C. Anagnostopoulos
Epepke said:
So you're telling me that some large percentage of the people working on custom applications or in new start-ups are talking to the customers? How do they get any work done?

Let's separate the two. Working on custom applications, yes. I'm not sure where the question "how do they get any work done" comes from, but I'm guessing that it involves some assumptions based on the old waterfall model. The only answer I can give is that you've answered your own question. It gets done, and it gets done incredibly fast, because there is close communication with the "customers."

Maybe you could state the reasons why you ask that question, and I could come up with a better answer.

As for the other I think that start-ups are primarily in the business of starting up, no? If it's a start-up in the shrinkwrap software business, well, I've already talked about that. If it's a start-up in another business, then they don't generally require software development services at all. Most startups do just fine outsourcing their accounting and payroll, but this is usually local.

Quote:

The venture capitalists must be out of touch with this. Ravi Chiruvolu (Charter Venture Capital) recently said "Right when you think about Employee 11, you should think about India. My view is you should not start a company from scratch in the US ever again."

Besides the good questions of whether VC's are still in touch after the bubble burst, don't you think this falls in the "news flash, film at 11" category?

[Paul C. Anagnostopoulos]

Epepke said:

Quote:

Let's separate the two. Working on custom applications, yes. I'm not sure where the question "how do they get any work done" comes from, but I'm guessing that it involves some assumptions based on the old waterfall model. The only answer I can give is that you've answered your own question. It gets done, and it gets done incredibly fast, because there is close communication with the "customers."

It was a sarcastic question. In my career, I found that the more I talked to customers, the less I got done, because they didn't know what they wanted, either. Of course, my experience was in operating systems and start-ups.

Quote:

Besides the good questions of whether VC's are still in touch after the bubble burst, don't you think this falls in the "news flash, film at 11" category?

I don't know, that's why I started this thread. We better hope that this outsourcing idea turns out to be obviously flawed, because if it doesn't, that's the way it's going to go. The thing is, there is a vested interest on the part of the managers of new companies to make the outsourcing thing work.

~~ Paul

[epepke]

Quote:

Originally posted by Paul C. Anagnostopoulos
Epepke said:
It was a sarcastic question. In my career, I found that the more I talked to customers, the less I got done, because they didn't know what they wanted, either. Of course, my experience was in operating systems and start-ups.

I saw the sarcasm, and I'm countering by being striaghtforward. But thanks for a better clue on what you mean.

Of course the customers don't know what they want! It's unfair to expect them to know what they want in the first place. How can they possibly know what they want until it exists? That's my job. If they knew what they wanted, and it were the same thing as what they needed, then the waterfall methodology wouldn't fail so often.

If you notice in my original reply, I said to write down everything that the stakeholders say they want, and then largely disregard it except as a guide for finding out what they need.

Because they do like to say what they want, possibly because the whole culture works that way. Besides sometimes giving good information, it's a necessary social game. There's often hostility at first, because people have been burned, but once that's passed, then one can establish trust. Once that happens, the customers become allies and come up with all sorts of incredible stuff. But you have to get past the initial stages.

I have a couple of tricks that I use to try to accelerate the process. I may ask a customer one simple question: "What keeps you up at night, worrying?" The direct approach sometimes works. Or I may say, "I'm pretty smart, but probably only about computers. I'd like to watch you work for a while, so I can learn from your skills. You see, I don't understand <insert job> very well, and if I do understand it, it's based on another person in another company. I want to learn how you do it." Yeah, it's shuck 'n' jive, but it works, too. The biggest problem is when the person's supervisor inisists on sitting in on the observation. Then I have to use a more complex approach which also works. (Fortunately, I've studied anthropology a lot.)

What I'm pretty sure of, though, is however one gets through that barrier of trust, once having punched through, all the effort is rewarded manyfold.

Your exasperation, ironically, buttresses my point.

Quote:

I don't know, that's why I started this thread. We better hope that this outsourcing idea turns out to be obviously flawed, because if it doesn't, that's the way it's going to go. The thing is, there is a vested interest on the part of the managers of new companies to make the outsourcing thing work.

The only thing that most managers can do is make it so that they can pretend that something works. That works rather well in an era of prosperity, when all of the money is phony anyway. It doesn't work in an era of depression, which is what we have now.

[Zep]

Having seen it happen before, my own view is that outsourcing company-vital developments offshore is attempting to ram a really big square peg in a solid round hole. It will go in, but only if you bust it up real good so it's not a square peg any more!

[michaellee]

epepke, et al..

Very interesting topic. I have read all the posts and find all points of view valid in their own way. However, all of the discussion revolves around companies, be it start ups or old hats. My input comes from the other side not yet discussed, which is small business programming, or individual programming, or specific application oriented programming. You say HUH?

I spend about half of my working hours programming; I develop custom programs for specific customers- some of these apps take just a couple of hours, others might take longer than a year.

The outsourcing quotient rarely comes into play in my market(POS retail software specific to only one particular system) as retailers who use this POS software and want custom Windows apps crafted to data mine their inventory, sales, etc. can rarely find someone in this country to do the work, let alone outsource to India. Flying solo, I work hands on with the customer throughout the development process, knowing damn too well that what the customer thinks he wants and what I know he will need are two different things.

Depending on the expected development time (keep in mind that I work completely alone and am responsible for creating, developing, packaging, marketing if necessary, support, etc..) I charge an average of $50 an hour. This includes travel time, meetings and programming and telephone time. This rate is a bargain for the type of customized programs my customers receive.

Well, it then finally hit. The POS software that I program around, with and to, has been in the retail market for 20 years, with over 5,000 retailers using the system worldwide. Two years ago, the local programmers who built this system in their own language, decided that it was time to join the Windows world, and change the system into a full fledged Windows application. Not a small task by any means, but they managed to complete the task and the new software has and is being installed worldwide.

A few bugs arose in the new system, and the company and the programmers did not want to nickle and dime the fixes by hiring local programmers to do the work. They instead opted to overhaul the entire package, and to do so their solution was to.... outsource the entire job to.... India.

The cost here in the U.S. would have been around $50-60 per hour; the clincher for India was that they agreed to complete the task at $12 per hour, and at a maximum number of billable hours.
The initial time frame for the job was 9 months to a beta version, 1 year for first release. When 9 months had passed, the local programmers and owners were on a plane to... you guessed it... India. Although the Indian programmers were adept programmers, it was impossible for them to complete the tasks deemed the easiest by US programmers.-- they just don't have the experience in the POS retail world to know enough detail.

So the time frame was moved back 6 more months, and lo and behold, another 6 months on top of that! So now two years have passed, and not even an alpha or beta version to test!

As I said before, the India programmers are eating the costs now, but that still doesn't mean the local company is jumping for joy either. They are in to deep to back out, so they continue to work with the India guys, praying for something to come out by January of next year. LOL.

The moral of this story? The local company's sales have been somewhat stagnant the last two years; but enough to keep afloat. The India people are taking a deep hit, having to spend 4-5 times the hours originally estimated to complete the project without being paid. How am I affected? Not at all, as I have more than enough requests for programming work, more than I could ever finish in three lifetimes.

I suppose being the small guy really does have its advantages, especially in the topsy-turvy world of computers, programmers, the US, and India.

[Paul C. Anagnostopoulos]

I mentioned this conversation to a good friend of mine who has been in the embedded systems business for about two decades. He had two remarks.

First, he thinks that the problem is not that programming will move overseas, but that the entire software development process will move. He's not so much worried about India as Norway, Belgium, and the UK.

Second, he has seen his business drop to about half of what it was five years ago, and he claims that almost all of those projects have gone overseas.

~~ Paul

[epepke]

Quote:

Originally posted by Paul C. Anagnostopoulos
First, he thinks that the problem is not that programming will move overseas, but that the entire software development process will move. He's not so much worried about India as Norway, Belgium, and the UK.

For embedded systems, he may well be right. Europe is currently a much more interesting market for e.g. cell phones than the US at the present, especially since Motorola didn't really take off.

[UnrepentantSinner]

I'm not in telecom, but I work at a telecom equipment manufacturer.

In 1998, the American owned company was baught out by a French company and until the dot.com bust was humming along eliminating redundant personnel from the merge, but still planning on growing and increasing product line and buildings occupied on the coportate campus.

Well, 2000 stated a change and a slow hemorrhage of jobs to India but more so to Canada. When the merge first happened, the US division was the North American top dog. Now there's more action in Ottawa than in the Plano. First they shaved off manufacturing. Then test and repair. Call center. Technical support. Help Desk. It's all eliminated, outsourced or sent to Canada.

Our security company signed a new 3 year contract in Septermber, but I wonder if there'll still be a campus here for us to watch over in that time.

[epepke]

Looks like I'm going to turn out to be right:

http://www.businessweek.com/smallbiz...31231_3576.htm