You can never know what’s going to come out of David Scott Lewis’ mouth next — but you should know whatever he says will be insightful and you’d better pay attention. An ex-METAGroup analyst who — along with many others — lost his job when the company was acquired by Gartner Research, he had never been to China, aside from Hong Kong. But not so many months ago, he moved overseas because little was keeping him in the US and love beckoned from afar (and online). The last time we caught up with him — in July 2005, when he'd returned home for a brief visit to speak at the AlwaysOn conference at Stanford University — he was providing consulting services in China, primarily to US clients trying to locate just the right service providers to work with.In this interview, Mr. Lewis supplies a no-holds-barred rundown on the current state of the outsourcing business in China. Part 2 of a three-part article.
Let's bring this back to outsourcing. What American companies should be looking at [doing work in] China?
I think everybody should do it.
[I was just involved in a survey with the heads of enterprise software companies who said] they were afraid of intellectual property. I do warn people that if they come to China — if they are going to do software development, for instance — they pretty much have to assume it's going to be ripped off. You don't give them critical code. But most software development isn't critical. Only a small part of it is.
China is very good for the rest, and people should be looking at that.
I also recommend that they try to look more at where there is a little more R than D. But you don't want to do basic research there; you want to do research that might be three years to product. It keeps them far enough away from being able to really do anything from an intellectual property perspective but you can get well educated, cheap labor… and not pay for a lot of expensive PhDs. Give them a fun environment.
What's a fun environment?
For them it would be just the freedom to explore things. Microsoft gives them that. Microsoft Research Asia in Beijing is giving that opportunity. Now I think the research center is overblown. The headline on the cover of MIT's TechnologyReview was "The World's Hottest Computer Lab." I thought, "Who snowed these clowns in MIT?"
These guys are sharp. We have some good overseas Chinese — "sea turtles" we call them. They are returning to China.
The big problem I see is that the mid-tier outsourcing firms in the United States — I am sorry… systems integrators — they were always cheaper than the top tier… This guy that's number 100 could always be guaranteed to be cheaper than IBM Global Services, and now he is not. So it's like, oh-oh, what do I do? Well their immediate mentality is Air India. So they are all off to Bangalore. Bangalore is running into headcount problems. They can't hire enough people. So China is looked at in almost all DiamondCluster and other surveys as the next preferred location.
Malaysia and the Philippines are challengers, but they just don't have enough people. Even if you employ all of their engineers, there is still a lot more work for China. They are better at BPO anyway because China BPO is really Korean and Japanese. When they say they say they are going to do BPO for the United States, this isn't going to happen anytime soon at all.
Give them a coding problem.
Give them a custom application. Right now, virtually all the work is localization and testing. Next step on the food chain is custom application development. They would actually be pretty good. The Indians are building a consulting practice by verticals. They are getting pretty sophisticated, so their billables are also going up now too.
Let me give you an example. For Java programming, that person is charging $75 an hour, whether he is on contract or whether he is internal. You can pay that guy $3, $5 an hour max in China, and you can bill them at $15 to $20. So you are going to make at least $10 per employee per hour. So $10 per employee is $20,000.00 in the US a year. By US standards, gee, that is not much. But considering what you are paying him, you are getting a very good return. If you have 100 people like this, well, figure that out. You have 100 people, and if you drop that to a bottom line, that is pure profit.
What kinds of mistakes are you seeing American companies doing in the way that they look at possibly sending work to China?
The biggest problem is ignorance. They don't know where to begin… The most common questions that I get are about three things: economic, legal and political environment. The reason is just about everybody wants to go captive and they are scared to death.
What have they got to lose?
They are afraid they are going to get nothing out of it. They might do code development and they are going to get the code ripped off. That it could even be worse than zero. It could be negative return. They are very concerned about that whole aspect.
That is where the partnering comes in. I think for most US firms, partnering is a better strategy. Everybody realizes EDS created Infosys [in India]. I mean, they built up Infosys by giving them the financial services work. Everybody understands that is what is going to happen with China.
But what will happen is you will have US mid-tier integrators working with Chinese systems integrators; and sure enough, everybody knows in five years they are going to part ways. The Chinese firm will have a presence in the states but the US firm wanted to go captive anyway. After five years, they are going to know enough to go captive. They are both going to switch. And who knows? They might merge. It could be beneficial.
So people are going in with that kind of an attitude — as long as they realize this might be a short-term marriage. You are not going to die with your bride. This is five years, and maybe you shouldn't have any children…
What are the issues?
Beyond the technical legal issues of [intellectual property rights], what are the real legal issues you have to contend with? One: guanxi — when you are trying to get local contracts. Things like that. Getting work from government officials. It's not uncommon for firms that once you start making money, [he knocks his hand on the table]. And you have no way to predict this. That is the problems that all the westerners complain about. Once they start making money, the government officials — mostly local people — show up and they knock on your door and it's time for the payments.
How does this work?
Usually people tell them to drop dead, then take it the provincial level or to the central level. You have to be careful about what you do. There are ways of handling that. I am familiar with the ways to get around that problem, but it can happen.
You don't have to worry about physical harm or arson?
No. You don't have any of that.
People suddenly not showing up for work?
No. But you might have your electricity shut down.
You would lose your electricity. Your net access gets cut off. Things like that. No physical violence. They are not that cruel. They are not physical, but they will do the other stuff.
What are the other mistakes that people make wanting to go to China?
They think that there are only two cities in China.
The problem is that a lot of firms run higher cost structures in Shanghai and Beijing than in India, so that becomes a real problem. You can get better talent on a comparable basis in India, and why would you go to Beijing or Shanghai then? It doesn't really make sense. There is that problem in those two cities. It's not cheap to operate a business in Beijing or Shanghai.
So you really need to look at second-, third-, fourth- or fifth-tier cities depending on your risk level. Second-tier cities are probably OK. And third tier cities are the up and coming ones, but nobody speaks English in the third-tier cities. Third-tier cities can have your best payback, but it's a tough management thing because nobody speaks English.
Xi'an is where the Terra-cotta soldiers are. They are going to rename it IBM Global Services. IBM is going in there and they are basically going to own the city. I'm not really at liberty to say where else they are going, but Xi'an they have mentioned. They are going to own both of their cities.
Talk about management.
The smartest thing China can do is to allow as many Taiwanese to come over there and run the country. There are over one million Taiwanese [in the People's Republic of China]. I have heard that there are as many as two million. I really don't know what the number is. They are all the senior managers, and they are the bright guys that are running the country.
The problem is, like in the IT industry, if you are over 35 and you are a mainlander, that is impossible. So you don't have people with a lot of experience. The Taiwanese come with experience. They are the adult supervision that makes this thing kind of work.
The Taiwanese are not really hot in software. They are better in IC manufacturing and all that. The mainland is lacking that kind of adult supervision. They need it.
So if you are looking for services in China, say, software development services, how do you structure and evaluate the management that the service provider has to know whether they can actually carry out your project?
Doing your due diligence is pretty much exactly the same way you would do it if you were looking for a partner here. There should be no difference except all the financial stuff is totally fudged. You cannot rely on any financial data at all.
First of all, they are all private companies for the most part. Actually in the IT sector, anything in software is going to be private. The companies are notorious for running multiple books. It's even a joke. If you go in, they will say, "Well, do you want to see a profit or a loss?" Then if you say, "I want to see the real books," they will say, "We have got them here somewhere. We will take a look and will get back to you."
They actually run into problems that way. If it's CMM certification, you have to show your books and you have to kind of show a profit. Well, the problem is that a lot of these companies are showing losses, so they don't have to pay [taxes]. Although in China, I think you pay taxes even if you lose money. It's a very strange system. Everything is strange there. You pay less money if you lose money.
They may not want to show any profit. They may want to show some deep losses, but then they have to go in front of CMM and they have to show how they are making money. So it's like, "Well, now what do we do? We are showing losses." So they have to rectify their books.
…You really have to sit down and do more of a thorough technical evaluation. You have to make sure they have people who are trained in America [who] are PMs. You have to talk to their references, although that is shaky as well, but you need to at least try and feel them out.
Take a look at their code. You can't be shy about taking a look at actual code they are putting together. Take a look at their budget management capabilities.
One argument that I have for the Chinese is that the whole CMM stuff is kind of overstated for them. At CMM, [Level] 3 is adequate for Chinese firms because no one is going to give them any work that requires anything beyond the CMM [Level] 3. Let's face it: You don't need to know how to manage the Space Shuttle if you are making a glass cup or something.
I tell [Chinese companies] they need to emphasize vender certification because they are weak on that. US systems integrators are usually based on two things: platform and vertical, and that is it. If you look at most systems integrators in the states, it's platform, vertical. You choose. And the Chinese don't do that. And I always get on their case for that. I always tell them, "Look. You knuckleheads need to figure out what your platform is. Don't have 50 guys doing Java and 50 guys doing .NET. That is the wrong strategy. You have to choose a religion.
"Then you have to choose a vertical." One of the problems they all have is they all want to do financial servicing because the Indians do it. The financial services market is hypercompetitive. Another problem is the financial services domestic market has nothing to do with the overseas market…[Particular Chinese companies] are great in the domestic market. They try to bring this to the United States. All the systems are radically different. Our systems and the Chinese banking systems have nothing in common, so they can't leverage it. Maybe your knowledge is actually a liability because you might be thinking along those lines when our systems don't work that way. It has been very difficult for them to leverage that.
Out of the top 100 solution providers in China, only one has manufacturing as its key vertical and that company is probably going to die… But manufacturing is a logical place for all these companies to be. They don't do it. They just don't think strategically at all. They do what they want to do because it sounds fun, because some guy's brother has a connection at Citicorp — "So that is our vertical." That is not strategy. That might be an opportunistic contract, but don't do that as strategy.
Analyst David Scott Lewis Shares His
The Current State of