In Traditional Chinese Perplexity answers, news media, Wikipedia, and official brand sites account for more than half of the citations, while official sites are often the smallest of those three groups. This runs against the instincts of many marketing teams. They put the entire content budget into their own site without building paths into news and encyclopedic sources, then wonder why the brand remains underrepresented in AI answers.
How We Conducted the Study
During the first half of 2026, we sent Perplexity a set of Traditional Chinese queries based on questions Taiwan B2B and SaaS readers commonly ask. They covered four intents: product comparisons, definitions, selection advice, and industry trends. We classified every cited domain into one of seven groups: news media, Wikipedia and encyclopedias, official brand sites, forums and communities, government and academic sources, ecommerce and price-comparison sites, or content farms and aggregators. This is Tenten's own sample, not a full-web census, and the distribution changes with the topic. Treat the results as directional rather than precise market share.
Three Main Source Groups: News, Wikipedia, and Official Brand Sites
Although we used seven categories, citations concentrated heavily in the first three. The overall distribution in this sample was approximately as follows.
- News media, about 38%: Financial, technology, and industry publications formed the largest single source group.
- Wikipedia and encyclopedias, about 17%: Wikipedia led the category, followed by specialist dictionaries and reference sites.
- Official brand sites, about 14%: Product pages, blogs, and documentation published by the company.
- The remaining 31%: Forums, government and academic sources, ecommerce comparisons, and content farms, with no single category exceeding 10%.
The first three groups add up to nearly 70%. The citation opportunity is concentrated rather than scattered: a brand needs paths through news, encyclopedic sources, and its own site.
Why News Takes the Largest Share
News leads for three practical reasons. Publications update frequently and provide current, cross-checkable sources for trends and recent events. Mainstream media also has strong authority and clean page structures that make paragraphs and figures easy to extract. Finally, several outlets often report the same event, creating corroborating sources and increasing the chance that a model will cite one or two. For a brand, coverage of one product story by three financial or technology publications may create more AI exposure than 10 unlinked posts on its own blog.
Wikipedia's Less Visible Leverage
Wikipedia's share is easy to underestimate. It may not dominate trend questions, but it becomes a default citation for definitions, history, specifications, and people. The lesson for a B2B brand is not to create its own entry; most companies do not meet the notability threshold, and promotional entries can be removed. The useful work is keeping relevant category entries accurate and making sure their citations lead to reliable primary material. When a model reads Wikipedia and follows the references, a brand's official page or white paper can enter the citation chain.
Official Brand Sites: A Smaller Share, but the Only Source You Control
Official sites account for only about 14%, which can sound discouraging. Yet news depends on media relationships and Wikipedia depends on community consensus; the official site is the one place where a brand controls 100% of the content, structure, and update schedule. Citation performance depends more on extractability than volume. The same product specification is difficult to use when buried in sales prose and much easier to cite when presented as a direct answer, table, or definition with a clear figure.
Prioritize a few practical changes: put the main conclusion at the top of each page, use real reader questions as headings, add verifiable figures and dates to important claims, and apply FAQPage structured data where the page genuinely contains common questions. These changes preserve the brand voice while making the answer easier for a machine to extract cleanly.
A smaller official-site share does not make the site unimportant. It makes the site the only citation lever a brand can improve immediately without another party's approval.— Tenten GEO content team
Turn the Distribution into an Action Plan
If you do only one exercise, identify the brand's largest gap across the three groups. News: has any independent publication mentioned the brand in the past six months? Wikipedia: does the category entry cite your data? Official site: can the priority pages be extracted as clean answers, or are they still written as essays? Each question points to a different remedy, and resources should go first to the largest deficit.
To see how often your brand appears in Traditional Chinese AI answers and which source group is missing, track the live citations with Brand Radar or book a 30-minute GEO diagnostic session. We will run your actual queries and compare the brand's presence across news, Wikipedia, and official-site sources before recommending where to act.
