— What Independent Websites and Platform Walls Suddenly Made Me Understand

By Jane Sonnenschein · February 13, 2026

Background for Readers Outside China

Chinese internet users sometimes joke that today’s online world is divided into two parts: China, and everywhere outside China. This is not a literal description, of course, but it captures the long-standing boundary between the internet available in mainland China and much of the global internet. For most people using an ordinary domestic connection in mainland China, services such as Google, YouTube, Facebook, and Instagram cannot be accessed normally without a VPN or other technical tools. Within China’s own internet ecosystem, content moderation, restricted keywords, and the removal of posts by platforms are also part of everyday online life. These realities should not be denied. They form an essential part of the background needed to understand China’s internet.

China, however, did not always exist without Google. Google entered the Chinese market in the early 2000s and officially launched Google.cn for mainland users in 2006. To operate within China, the company agreed at the time to filter some search results in accordance with local regulatory requirements. During that period, Chinese users could choose between Google and Baidu, and Google held a significant position in the country’s search market.

The decisive turning point came in 2010. After cyberattacks targeting the company’s systems and some user accounts, Google also publicly expressed its opposition to the censorship requirements applied to its search service. It announced that it was no longer willing to continue filtering results on Google.cn. Google subsequently stopped operating its original mainland search service and redirected users from Google.cn to Google.com.hk in Hong Kong.

Strictly speaking, Google did not withdraw every part of its business from China. It continued to maintain some advertising, sales, and technology-related activities. Yet its locally operated search service for ordinary mainland users effectively came to an end. Over time, Google Search and several other Google services became unavailable through ordinary internet connections in mainland China.

After Google stopped operating its mainland search service, Baidu, founded in 2000, became the most visible domestic gateway for general search and was widely described as “China’s Google.” Yet Baidu’s later position cannot be understood simply as the story of a Chinese company naturally becoming an equally powerful replacement for Google. Baidu still has a large user base and remains one of China’s major general search engines, but it is no longer the automatic first stop for every kind of information in the way it once seemed to be.

Part of this change came from a gradual loss of trust. Chinese users have long criticised the volume of commercial promotion in Baidu’s search results and the difficulty of knowing whether the first result is the most relevant information or simply a paid placement.

The 2016 case of Wei Zexi became a symbolic turning point. Wei, a university student with cancer, found a hospital offering an experimental treatment through a Baidu search and later died after the treatment failed. A subsequent joint regulatory investigation concluded that Baidu’s paid-ranking system gave excessive weight to commercial promotion and did not clearly distinguish advertising from ordinary search results, thereby undermining the fairness and objectivity of the results and potentially misleading users. This case was not the only reason for Baidu’s declining reputation, but it exposed the conflict between search advertising and public trust with particular force.

Even so, explaining China’s search ecosystem only through Google’s departure, internet controls, or Baidu’s commercialisation would still leave out another equally important part of the story: China and Europe or North America did not enter the age of mass internet use at the same time. That difference in timing later shaped the development of websites, platforms, and search engines in very different ways.

The World Wide Web was developed at CERN in 1989, began spreading more widely among research institutions around 1991, and was opened more broadly to public use in 1993. From the early 1990s until the arrival of the first iPhone in 2007, Europe and North America had more than a decade in which the internet was built primarily around personal computers, browsers, and public webpages. If earlier personal computers, dial-up networks, and commercial online services are included, that period of accumulation was even longer.

During those years, personal homepages, company websites, online shops, forums, blogs, independent domains, hyperlinks, and search engines gradually became basic elements of how people used the internet. Vast amounts of information were placed on publicly accessible webpages. Websites connected to one another through links, creating the open environment that allowed search engines to crawl, organise, and index information over long periods of time.

China established a full connection to the international internet in 1994, but that date mainly marks the connection of national infrastructure, research institutions, and educational networks. It does not mean that ordinary Chinese households began owning computers and accessing the internet in 1994.

In 1997, China had only around 620,000 internet users, approximately 299,000 connected computers, and roughly 1,500 websites. By the end of 2000, the number of internet users had grown to around 22.5 million. In a country with a population of more than one billion, this was still only a small minority.

In other words, 1994 was the year the internet was connected to China. The point at which growing numbers of ordinary Chinese people began entering the internet came mainly around the year 2000 and in the first years of the twenty-first century.

This timing was closely connected to China’s stage of economic development. In the early years of reform and opening, the country was still rapidly catching up in industrialisation, urbanisation, infrastructure, household income, personal-computer ownership, and telecommunications. During the 1990s, internet access remained concentrated in universities, research institutions, workplaces, internet cafés, and a limited number of urban homes.

Only after 2000 did school computer rooms, internet cafés, home computers, dial-up connections, and later broadband bring the internet into the lives of a much wider group of young people. Many Chinese people did not first become regular internet users because their families owned a permanently connected home computer. They went online in internet cafés, where they chatted, played online games, read news, or entered communities such as Tianya, Mop, Baidu Tieba, university BBS forums, and the discussion boards attached to major web portals.

China certainly experienced the era of portals, chat rooms, BBS forums, blogs, and desktop internet use. But mass participation in the PC-based internet was concentrated mainly in the early years of the twenty-first century, giving it a more compressed period of development than in Europe or North America.

It is also important to distinguish between a forum that operates as an independent website and ordinary individuals owning independent websites of their own. Tianya, for example, had its own domain and website infrastructure and was independently operated as a business. But for its ordinary users, it was still a centralised platform. Everyone entered the same community, posted under the same system, followed the same rules, and depended on the platform for visibility and traffic.

The blog services later provided by portals such as Sina, Sohu, and NetEase followed a similar model. Users appeared to have their own personal blog pages, but those pages remained part of a large portal rather than independent sites whose owners controlled the domain, server, structure, and source of traffic.

This differed from the early web culture of Europe and North America. Before large mobile platforms appeared, businesses, shops, media organisations, and individuals had already had a much longer period in which to build their own domains, websites, blogs, and online stores, all connected through hyperlinks. Search engines grew within this decentralised network of webpages.

China also had genuine personal websites and independently hosted blogs, but for most ordinary users, portals, forums, and large communities were easier to enter and more likely to provide an audience. Even during its desktop era, therefore, China’s internet already carried a comparatively strong tendency towards centralised platforms.

Just as mass PC internet use was beginning to expand rapidly in China, the smartphone era arrived. The first iPhone was launched in 2007, and China issued 3G licences in 2009. Mobile networks and smartphones then spread quickly.

By the first half of 2012, the number of people accessing the internet through mobile phones in China had surpassed the number using desktop computers. By June 2013, China had 464 million mobile internet users, representing 78.5 per cent of all internet users.

China did not skip the PC internet era. But soon after mass internet use began to expand, the country moved rapidly into an internet centred on smartphones and apps.

This overlap in timing profoundly shaped the form of China’s internet. Before smartphones appeared, Europe and North America had already had many years in which browsers, corporate websites, personal sites, online shops, public hyperlinks, and search engines became deeply established habits. For many Chinese users who came online later, however, smartphones, apps, and large platforms were already the most familiar form of the internet from the beginning.

A small phone screen also encourages users to enter a single application that has already integrated accounts, content, products, reviews, payment, and customer service, rather than repeatedly opening a browser and searching through many separate websites.

Mobile internet and large platforms therefore reinforced one another. The spread of smartphones accelerated mobile payment, e-commerce, social media, video, and everyday service platforms. As these platforms grew, they attracted merchants, creators, and ordinary users to keep content, transactions, reviews, payment, and social relationships inside their own ecosystems.

The more convenient a platform became, the more users gathered there. The more users gathered, the more necessary the platform became for businesses and creators. As more content remained inside those platforms, users had less reason to leave, while the importance of public webpages and general search gradually declined.

This also explains why internet use in China later became highly concentrated. People searching for products often go directly to platforms such as Taobao or JD.com. Those looking for consumer experiences or lifestyle advice may enter Xiaohongshu¹. Short videos and tutorials are commonly found through Douyin, the Chinese domestic counterpart to TikTok, or through Bilibili. Many articles and forms of public content remain inside the WeChat ecosystem.

These large platforms resemble crowded marketplaces where consumers already know they will find products, content, and other users. Merchants and creators must follow the flow of people into the same marketplace.

By contrast, an independent website without platform recommendations, user reviews, or social distribution can resemble a shopping centre built deep in the mountains. It may exist, and it may even be beautifully designed, but few people will discover it naturally.

Building an independent website in mainland China is not impossible. However, websites hosted on mainland servers and offering public information services generally require an ICP filing, along with server management, technical maintenance, regulatory compliance, and continuous efforts to acquire visitors.

The more practical difficulty is traffic. When consumers are already accustomed to searching, comparing products, reading reviews, and making payments inside large platforms, an ordinary brand’s website is unlikely to become their first destination. Only brands that are already widely recognised are likely to have users actively search for their official websites. Most individuals and small businesses still have to return to large platforms to purchase visibility, accumulate reviews, and reach customers.

In such an ecosystem, Baidu faces more than problems of search quality and commercialisation. As increasing amounts of merchandise, articles, videos, reviews, and social content remain inside relatively closed platforms such as Taobao, WeChat, Xiaohongshu, Douyin, and Bilibili, the proportion of public webpages that Baidu can crawl, organise, and index also decreases.

Users no longer necessarily open Baidu before shopping. They may not use it first when looking for consumer experiences, and they are more likely to enter a video platform directly when searching for a tutorial. Search has not disappeared. It has been divided into e-commerce search, community search, video search, and search within social ecosystems.

Baidu still exists and still has many users, but general search is no longer the single, unified entrance to the Chinese internet that it once appeared to be.

The Chinese government has long framed internet governance in terms of cyber sovereignty, national security, public interest, and social stability. In my understanding, a country with an enormous population, significant regional differences, and a deep historical concern with national unity, social order, and sudden institutional disruption will inevitably consider questions of security and stability when opening its information boundaries.

This does not mean that every restriction should be beyond discussion, nor does it mean that the problems created by platform moderation, restricted keywords, and content controls can be ignored. This essay is not intended as a complete defence of any particular system of internet governance, nor does it attempt to resolve the political debate between censorship and freedom.

It asks a different question: when access boundaries, differences in economic development, a comparatively compressed period of mass PC internet use, rapid mobile adoption, platform business models, and user habits all overlap, how do they collectively create a search ecosystem so different from that of Europe and North America?

China’s failure to produce a second Google therefore cannot be explained simply by saying that Google cannot be used in China. China once had Google, and Chinese users once had a choice between Google and Baidu. After Google stopped operating its mainland search service in 2010, Baidu gained more room to expand. Yet it later faced both a crisis of trust caused by commercial promotion and a deeper structural change: the platformisation of the internet reduced the amount of publicly accessible web content and redirected search activity into separate apps.

Access restrictions explain why a boundary exists between China’s internet and the global internet. Differences in industrialisation, economic development, and the timing of mass internet adoption explain why ordinary Chinese people mainly entered the internet in large numbers after the beginning of the twenty-first century. The comparatively brief period of mass PC internet use, the rapid arrival of smartphones, and the concentration of content and transactions inside large platforms further explain why the online world within that boundary became centred on apps and centralised platforms.

This essay is not really about whether Google or Baidu won or lost, nor is it only about which company possesses better technology. Its central question is this: the form taken by search depends on the form taken by the internet itself.

When information is primarily stored on open webpages, general search becomes stronger. When information is divided and enclosed inside separate platforms, vertical search and platform-based gateways become more powerful.

The fate of a search engine is ultimately determined not only by the company that operates it, but by the structure of the entire internet around it.

Why Is There No Second Google in China?

Recently, questions like these have appeared frequently on Chinese social-media platforms:

Why is Baidu, once known as “China’s Google,” mentioned far less often today? Why has Google remained the dominant search engine around the world, while Baidu is so often criticised within China? Is the problem technological? Is the product simply not good enough?

To be honest, I did not have a particularly clear answer when I first encountered these questions. At most, I assumed that the explanation might involve technology, management, or excessive commercialisation.

Only recently, while learning more about the structure of the global search-engine market, did I begin to see the issue differently. I looked again at Google’s market share in Germany and worldwide, including its share of close to 93% of mobile search in Germany, and then compared that with the structure of China’s internet.

I suddenly realised that the real question might not be which company built the better product.

The deeper question is: what kind of internet was each search engine built inside?

A search engine is fundamentally a tool, and it is a tool that depends heavily on long-term accumulation. It is unlike a short-video platform, which can hold an audience through one viral video after another. Search depends on the scale of its data, the network of links it can follow, its ability to crawl webpages, the long-term training of its algorithms, and, most importantly, the trust of its users.

Its growth resembles a flywheel. More users create more data. More data improves the algorithms. Better algorithms improve the experience, and a better experience attracts still more users.

Search is therefore a deeply long-term business. But long-term accumulation requires one essential condition: a rich, open, and crawlable web ecosystem.

A search engine is more like a library than a stage for performance. Yet a library can exist only if there are books—and only if those books are openly placed on the shelves.

In Europe and North America, particularly in countries such as Germany and the United States, large numbers of businesses still maintain their own independent websites. Even chain pharmacies and online retailers usually have their own domains and product pages. Prices, stock information, and product descriptions remain publicly available on webpages.

Google can crawl those pages, build an index, organise the information into structured results, and sometimes compare prices directly in the search results.

In other words, the Web is still alive. Open webpages remain part of the central infrastructure of the internet.

Within this structure, search acts as an integrator.

China’s internet developed along another path.

China did not skip the desktop internet era, but its period of mass PC-centred internet use was comparatively compressed before the country moved rapidly into the mobile internet age. E-commerce became concentrated inside Taobao, Pinduoduo, and JD.com. Content gathered inside Douyin and Xiaohongshu. Social communication became concentrated inside WeChat.

When we open the web versions of many of these platforms, we quickly discover that the app remains the genuinely functional product. Much of the core content is enclosed through dynamic loading, login requirements, restricted interfaces, and closed technical systems.

This does not necessarily mean that search engines lack the technology to crawl the material. It reflects a platform strategy: traffic is intended to remain inside the platform’s own ecosystem, where it can complete a closed commercial loop.

When purchases are completed inside e-commerce platforms, when zhongcao² and product discovery turn directly into sales inside content platforms, and when discussion and distribution take place entirely within social platforms, the marginal value of general search naturally declines.

Search behaviour becomes divided into vertical search. People looking to buy something go to an e-commerce platform. Those seeking personal experience go to a community platform. Those looking for tutorials go to a video platform.

General search is no longer the only entrance.

When all the “books” are locked inside separate cabinets, what is left for a general search engine to organise?

That was when I began to understand why the claim that “Baidu simply became terrible” is too crude.

When a search engine operates inside an increasingly platformised and closed ecosystem, the proportion of open webpages available for crawling declines, while commercial pressure remains. The central asset of a search engine is not its algorithm. It is trust.

Once users begin to suspect that rankings are determined by advertising rather than relevance, that trust begins to disappear. What is lost is never only traffic.

Rather than saying that Baidu merely became a bad product, it may be more accurate to say that general search itself was pushed towards the margins of China’s internet ecosystem.

Only then did I truly understand why Google has remained so powerful.

Its strength does not come from technological leadership alone. A deeper reason is that the structure of the internet in Europe and North America still allows general search to exist. Large numbers of independent websites maintain an open network of links, and the Web continues to carry much of the internet’s information.

Within that structure, Google acts as an information integrator rather than as one platform competing against other platforms from the inside.

China, by contrast, has entered an age of platform walls.

Search has not disappeared. It has been broken apart. People search for products within e-commerce platforms, for personal experience inside content communities, and for tutorials inside video apps.

General search is no longer the single entrance.

This is not necessarily a judgement about which path is morally better. It is a difference in development.

When the internet is an open network, general search becomes strong.

When the internet consists of platform walls, vertical search becomes strong.

Perhaps the difference goes even deeper, into the way attention is distributed.

Many European and North American business systems have a strong preference for turning activity into structures and models: the four Ps, the five Ps, funnels, frameworks, theories, and repeatable processes. These ideas can appear complicated, but they reflect a particular way of thinking—turning traffic, rules, and authority into mechanisms that can be optimised over time and repeatedly tested.

A search engine is itself a product of this kind of structural thinking.

A link becomes a vote.

Authority becomes trust.

Ranking becomes a rule.

The page that receives more references is placed higher.

What matters is the long-term persistence of the rule, rather than a sudden explosion of attention.

In the platform era, traffic is more often distributed centrally by algorithms.

One creator may be recommended today, another promoted tomorrow. Attention can be redirected very quickly. The rules are no longer publicly visible. They remain inside the platform.

Neither system is entirely right or wrong.

But each determines who becomes the primary distributor of attention.

When rules are central, search grows stronger.

When platforms are central, search is pushed towards the margins.

Perhaps the strength of a search engine has never been only a technological question.

It is closer to an institutional product, in which the rules themselves form the infrastructure.

Search depends on an environment in which rules remain stable over long periods.

When power over the internet shifts from publicly visible rules to internal platform algorithms, the role of search is naturally reshaped.

The reason search has remained powerful in Europe and North America is not that political life is stable for a four-year term. It is that the underlying rule structure extends beyond individual political cycles.

When the power to allocate traffic is concentrated inside platforms, the long-term stability of those rules becomes weaker.

Perhaps the real answer is that the fate of a search engine is not determined entirely by the company that owns it. It is determined by the form the internet takes around it.

The question has never really been about who won or who lost.

The internet grows into a particular shape, and search grows into a corresponding shape.

Now artificial intelligence is becoming a new gateway. Whether through Google’s integration of Gemini or through conversational search tools, the way information is organised is changing once again.

Perhaps the search engine of the future will no longer present a list of links. Perhaps it will provide an integrated answer.

The gateway may change, but one thing will not: human beings will always need a way into information.

When the structure changes, the gateway changes with it.

In the age of search, people actively went looking for information.

In the age of platforms, information actively comes looking for people.

Neither mechanism is entirely right or wrong, but each shapes a different rhythm of thought.

---

Notes

1. Xiaohongshu（小红书）: Literally meaning “Little Red Book,” Xiaohongshu is a major Chinese social-media and lifestyle platform. Users share product recommendations, personal experiences, photographs, short videos, parenting discussions, travel information, and everyday advice. It has become an important search and discovery platform for many younger Chinese users.
2. Zhongcao（种草）: Literally meaning “to plant grass,” this Chinese internet expression describes the process of making someone interested in trying or buying a product, service, destination, or experience through recommendations and appealing content. It is widely used in online shopping and lifestyle culture, especially when content directly encourages consumer desire.

This essay is also available in other languages:

Chinese version: 为什么中国没有第二个 Google？——从独立站到平台围墙，我突然明白了什么

German version: Warum gibt es in China kein zweites Google? — Was mir zwischen unabhängigen Websites und Plattformmauern plötzlich klar wurde