The Economist Gets It Wrong on Japanese Outsourcing

I must confess I don’t read the Economist as much as I used to. I suppose one factor was my belated recognition of how much to the right it had moved (it was evident to UK readers in the 1990s, but not being a close follower of British politics, I didn’t recognize it until the run-up to the Iraq War).

While it still produces special features that are often very good, its regular coverage has increasingly left me cool. And I’ve found justification for my reservations in a story this week on the failure of Indian outsourcers to get very far in Japan:

HCL he yokoso!” is how visitors are greeted at an office of HCL Technologies in Noida, a dirty suburb of Delhi. The phrase means “Welcome to HCL” in Japanese, and the neat line of young Indian employees bow deeply to underline the point: HCL is ready to do business with Japan, the world’s biggest computer-services market after America.

The welcome is falling on deaf ears, however. HCL, one of India’s biggest computer-services firms, has been trying hard to win Japanese outsourcing contracts for a decade. Its Japan Business Unit in Noida employs 25 fluent Japanese speakers who cater to visiting Japanese businessmen and supervise training in Japanese language and culture for the firm’s other employees. Around 250 software engineers are proficient enough in Japanese to converse with Japanese clients, albeit with laughably poor pronunciation. “They find it funny, but they appreciate our efforts,” says Vineet Nayar, HCL’s president…..

As a market for Indian outsourcing, Japan lagged behind the Netherlands and way behind America, which accounted for 67% of the total (see chart). Even these figures exaggerate Japanese companies’ enthusiasm. They include contracts with non-Japanese companies for work involving operations in Japan, which represent around half of the total value.

The article completely misses the real barrier to outsourcing in Japan: Japanese companies, including all their financial institutions, are still on mainframes. Japan still buys a lot of IBM big iron.

While distributed computing buyers have increasingly moved towards using off-the-shelf software as much as possible, mainframe computing was (and is) highly customized. And the documentation is written in Japanese. Being able to read Japanese requires a vastly higher level of skill than speaking it (contrary to popular perception, Japanese is a very easy language to speak at an entry level). So the only areas of the market that are realistically accessible to outsourcers are those that run on PCs, which is a fairly small portion.

If someone could figure out how to transition Japanese companies off mainframes, there is far more money in that proposition than in outsourcing (although outsourcing could follow from that). The fact that no one has cracked the underlying and more lucrative problem indicates how high the barriers to entry really are.