29th Jan, 2010

What Does It Really Mean for Google to Pull Out of China

The recent Google bombshell has certainly been heard around the world. Tons of coverage both in the US and in China — lots of headline echoes, some sensational nonsense, and very little substantial and knowledgeable analysis.

How will this movie play out? Let me spoil the plot here. My executive summary – not insider info from Google, just my own back-of-the-envelope analysis — is that (a) google.cn the service will move offshore, back to where it was. But the move has complicating factors that I will elaborate below; (b) Google will continue to sell their services in mainland China and earn revenue, but this may be negatively impacted if the pullout is not done carefully; (c) Google will keep its R&D center open, but its utility and importance is clearly diminished due to the pullout of the service offerings and the recent hacking incidence.

Now to the nitty-gritty. To understand what it might mean for Google to “pull out of China”, one has to understand what Google has inside China. It has, broadly speaking, three parts. It has the search service, Google.cn; a business team (including a sales force); and a R&D center. What Google the company has put a strong stake in the ground is with regard to its Google.cn service, and given the official statement contained in the bombshell, it does not appear to have any way out of this tangle, except for the service to move offshore, to HK for example. (Or, equivalently in search terms, just offer Google.com straight. The differences between Google.cn and Google.com lie not only in terms of content filtering but also in the other services that Google currently offer only in China – such as music search. (More on these other services later.)

How much will Google lose out by moving Google.cn offshore? You have to understand why Google brought it onshore in the first place. Back at that time, Google.com was reachable from inside China, but subject to sporadic blocking. Google figured that it needed a sufficiently pervasive presence to the Chinese audience. Otherwise, it may become irrelevant in China. Smart idea — look what happened subsequently to Facebook and Twitter – they are irrelevant in China because they are blocked early and thoroughly. Another less well known reason is that many places, especially university campuses, charged (and some still charge) students extra for international internet access. In order to reach these users, Google had to separately negotiate passageways with those gatekeepers (and often pay access fees) to make Google.com reachable from behind those walled gardens. Moreover, a local service would offer a platform to launch China focused services such as music search and download – a feature that has been often credited with cementing Baidu’s early dominance.

Moving the local services offshore now will incur less damage due to the improvement in environments. On the one hand, Google has established itself sufficiently in China such that blocking it completely will be a very testy proposition. For example, lots of university students voiced their strong displeasure during recent blackouts. On the other hand, the access environment has also improved, and internal demand will put enough pressure on the pipe owners to keep Google data flowing.

However, there is potentially serious collateral damage if this move is not done carefully, and this is perhaps why there is even the need for a negotiation in the first place and why it has taken so long to conclude. I am not talking about issues of the company’s social reputation or political standing in China. I am talking about their local services. Consider the music service, which is visible on its Chinese site but not on the US site. An educated guess says that the song rights Google has licensed are probably for mainland use only. Can they still offer these from HK or elsewhere? Ditto about map and other similarly locally acquired services. It is an option, technically, to keep these non-search services inside mainland China; however, they would be decoupled from the core search functions with much diminished user experience. It is hard to imagine that the Google brass in the US would permit such a world’s first exception for the China market. After all, these other services are loss leaders, and without the money making machine called Search, they do not have a reason to exist, corporately speaking.

More critically, Google has search partnerships with the likes of Sina.com, one of the biggest portals in China, and large mobile operators. Now, if the search service is provided from offshore and may be unavailable from time to time, would Sina (as a Nasdaq listed company) continue to rely on Google to power the search feature of its websites? Would Google be able to negotiate safe passages for such deals? Even if so, will these passages be unfiltered? If filtered, who does the filtering? And can the filtered search service still claim to be Google search? Will Google lose customers and revenue because of these factors? The deeper one probes these issues, the more complicated the situation becomes.

What about the much touted R&D center? There has been much fanfare since it was opened. It was said that Google chose to lease a separate but smallish building instead of part of a larger building because in the larger buildings Google would not be permitted to build their famous kitchen, an essential part of the corporate landscape. (Well, corporate security could be another argument, but Google is known to lease space in non-standalone buildings, even here in China.) Google’s salaries topped the previous leader in this market, Microsoft. Every (well, almost every) young and old engineer dreamed to join Google. The swagger was apparent everywhere around the Google building in Tsinghua Science Park.

From the outset, Google US was terribly concerned about information leakage in China. In the early days, Google US employees who travelled to China on business were issued blank, new laptops and were told to keep their own laptops at home. Now with the much rumored inside job, the pressure to close the lid is greater than ever. As a comparison, Sun Microsystems always had its entire intranet open to its China R&D center, except for export controlled technologies such as micro codes and hardware designs. But Oracle never allowed its China R&D center to touch core technology. What will happen now that the two companies have merged?

At the same time, the pullout of the local services also reduced the need for paying sky high salaries to several hundreds of idling engineers. What will they do? They cannot all move to Facebook as their US colleagues did en mass.

Assuming a Google pullout, where does that leave the China search market? Common wisdom says this is a win-win for Baidu, which will have the China market all to itself. That may be, for the short term. In the long term, things may not be as rosy. Not having Google as a distracting target, Baidu could become the focus of those who are unhappy about all the “Internet sins.” But more substantially, the departure of Google could make way for yet another Chinese contender to the search throne. It is often said that Baidu can beat the technologically more superior Google because it can operate the China market more effectively. This is obviously true. But by the same token, Baidu is more vulnerable to another local competitor, which can be as fierce, as shrewd, and as connected as Baidu. Words have it that Tencent – the owner of the dominating IM client QQ and since has expanded successfully into adjacent markets — is coming along with its own search offerings. And Sohu, the well-known portal, was founded to be the Google of China (thus its Chinese name — meaning search fox) and never gave up its search dream. And of course, Bing is always ready to fill any vacuum left by Google. Interesting times? You bet!


This is really a interesting probe into this pull-out-of-China story.

It was once said that Google China employees are not allowed to access source code owned by Mountain View. Well, recently I got some words from some source, this is not true. They can access source code from Mountain View.

As a software engineer, I don’t believe software source code should be protected by segregation. The linux kernel code is out there, but not every body is linux kernel expert; Mozilla Firefox is open sourced, it doesn’t make it more vulnerable in competition.

People is always more important than source code in software engineering.

Leave a response

Your response: