Baidu, China’s largest search engine and one of its largest information technology companies, is on the move. In 2012 it announced plans to expand beyond Chinese borders and in 2014 began offering services in Vietnam, with similar ventures in Indonesia, Brazil, Egypt and Thailand. In 2012 the company invested in a major AI research centre in Singapore, focused on expanding its business to SE Asia. Compared to Alibaba and Tencent, Baidu is the weakest of China’s big three tech companies, but still enjoys 80% of the search market in China and significant market capitalisation.
Baidu’s expansion into Thailand began in earnest in 2014 and is part of its larger push into Southeast Asia, and a Thai language version of Baidu’s flagship search engine, baidu.co.th, appeared early that year. The country is one of the region’s fastest growing markets, with lots of room left for expansion. Google is dominant, and internet penetration growth rates are above the global average, with internet penetration itself still at just 67%.
This project traces Baidu’s expansion into Thailand, focusing on its search product. In doing so it builds on Jiang’s work on Baidu and adds to a growing body of commentary on Chinese tech expansion abroad. Much of this commentary raises fears of the expansion of Chinese tech companies on the grounds that it will lead to censorship and access to vital industries and resources. Other commentary, while noting the huge amounts of capital bolstering Chinese tech investment abroad, notes that Chinese tech companies have enjoyed cosseted conditions in China which have allowed them to grow, free of foreign competition, suggesting that success in foreign expansion may not be as forthcoming as might be expected. This project tests two hypotheses. The first is that for Baidu, success in China will translate to success abroad. It interrogates the mechanics of that expansion using Baidu’s flagship search engine as a test case. The second hypothesis, to be tested in the second half of the project, is that given the role of the domestic Chinese market in its business plan, Baidu will extend the information controls it practices within China to its operations abroad.
The first part of the project analyses the implementation of Baidu’s Thai search products, which cater to both Chinese and Thai speakers. It asks what the implementation of Baidu’s search products in Thailand can tell us about both the goals of Baidu’s expansion abroad and the company’s capacity to implement them. Perhaps surprisingly given the prominence of search in the company’s success in China, the nature of Baidu’s search product in Thailand has proven to be remarkably fluid. When Baidu entered the market in 2015, the search engine, baidu.co.th. was a standalone search box, much like the Baidu search page in China. Baidu sources suggested the page was simply a page for internal testing purposes, which seems unlikely. In 2013 the company had announced firm plans to unveil a standalone Thai version of its search platform.
However by mid-2017 the situation had changed. The URL baidu.co.th no longer resolved to a standalone search page. Instead, it redirects to th.hao123.com, a web portal modelled on the hao123 portal in China, but in this case the search bar defaults to Google and is branded as such. A similar shift occurred in Vietnam in 2013-2014. There, Baidu was ultimately forced to withdraw from the market after a consumer backlash. Without a standalone search product, researchers tested the search function in Baidu’s web browser, known as Baidu Browser, occasionally branded as Spark, which was also introduced into Thailand in 2014. Baidu Browser is derived from the Chromium browser (the core browser of Google Chrome), and so users may search using the address bar, but may also use Baidu Browser’s custom search field, which is similar in appearance and functionality to the custom search field in Mozilla Firefox.
Baidu Browser’s behaviour also changed over time in unexpected ways. When the research project initially began in March 2017, the search engine that Baidu Browser configured for the user was simply named “Search”, with an S inside a green square. Upon searching using this search field or the address bar, the user was redirected to a search engine named “Govome” – likely a form of malware. Govome is/was not a standalone Baidu search product: rather, in this instance, it piggybacked on Yahoo search rather then running from a Baidu-specific search. This means that all of the search results for users searching via Baidu Browser within Thailand come from the Yahoo search engine, with Baidu’s own advertisements placed amongst the results. Govome uses the “insppartner.com” API for providing search results. The domain “insppartner.com” is registered to Infospace Holdings LLC, a search engine corporation based in the United States which allows white-labeling and reselling of search services for any partner or affiliate.
Baidu does not appear to be focused on replicating the search experience which its users have in China for those using this product in Thailand, even when searches are conducted in Chinese. When making the ‘call home’ connection to Chinese servers, Baidu is removed entirely as a search option, meaning Baidu is not providing a version of its Chinese product to users of its product in Thailand Instead, the browser configures the default search options upon loading the browser for the first time and each subsequent time. Upon first installation of the browser, the default search option is set to Baidu. Until September 2017, the search option that replaced Baidu was the ‘Govome’ search engine. However, by October 2017, it was observed that this behaviour had changed, with ‘Yahoo’ being set as the default search option and Baidu being removed once again.
When using this search feature — Yahoo for those in Thailand — the search query is redirected via a domain “searchrfu.com”. The ‘searchrfu’ domain appears to facilitate Baidu’s tracking of search queries for users outside China in this case, potentially as a search referral profiteering mechanism or for facilitating the storage of search queries.
Interestingly, this version of Baidu Browser downloads a ‘whitelist’ of pornography websites to be used with an ‘automatic incognito mode’ feature. This mode ensures that access to these websites is not recorded in the user’s browsing history. The small list of domains correlates with regions outside China into which Baidu has expanded, with domains supporting English, Brazilian Portuguese and Thai.
The early stages of this research project suggest that in contrast to the company’s earlier statements search is not Baidu’s primary product in Thailand, and particularly not Baidu-branded search. The low-end technology applied to Baidu’s search products reinforces the judgment that search is not the company’s main goal in this market, while changes in the product over a short period of time also suggest that the product design is evolving in what appear to be unexpected ways. This is surprising given the centrality of search to Baidu’s business model in China.
Overall, the beginnings of this research project suggest that Baidu’s business model in Thailand is not a Chinese tech juggernaut in the way Chinese tech expansion more generally has been portrayed, at least when it comes to expansion of search products, although this may simply suggest that Baidu’s business model has adapted to the Thai market in unforeseen ways. These initial findings also contribute to our understanding of the impact of Chinese information controls on Chinese tech products outside Chinese borders. In this initial finding, the lack of emphasis on or capacity for direct control of search products suggests the desire or capacity of control in a manner similar to those within China is limited. The second forthcoming part of this research project tests those information controls in more detail.