At the earnings call for the new quarter, Zhihu CEO Zhou Yuan summarized two priorities for Zhihu in 2025: first, financially, achieving significant loss reduction throughout the year, potentially even approaching profitability; second, operationally, enhancing user experience, increasing user engagement, and improving Zhihu's reputation and credibility.

Financially, after several quarters of cost reduction and efficiency improvements, Zhihu's losses have narrowed significantly. In Q3, Zhihu reported revenue of 845 million yuan, with net losses narrowing from 278.4 million yuan in the same period last year to 9 million yuan, a decrease of 96.8%. This corresponds to a significant reduction in various expenses, with income costs decreasing by 35.6% from 473.7 million yuan in the same period last year to 304.9 million yuan.

However, increasing revenue is a more challenging task than cost reduction. Financial metrics and targets are relatively easy to quantify and execute directly, whereas growth in operational indicators, besides resource investment, also requires consideration of market insights, external environments, and other uncertainties, and often have a lagged effect.

If we divide Zhihu's challenges into two timelines: one is the intensifying competition for internet attention, where users have numerous options for media consumption, and short videos, which dominate the top of the user engagement pyramid, continue to have a siphoning effect, making the content Zhihu provides no longer unique. The other is Zhihu's long-standing struggle between commercialization and user experience, facing intense competition while under pressure to monetize.

The intertwining of these two timelines has led to overall negative perceptions such as "watered-down Zhihu," "too many ads," and "creators leaving," pointing to issues like content homogeneity, excessive commercialization, and damage to the creator ecosystem, ultimately shaking the platform's positioning.

Against this backdrop, Zhihu seems to need to reanswer the question "Who am I?" and provide users with a clear usage scenario. Clearly, Zhihu is also seeking change – under the premise of maintaining differentiation, continuing brand value, and not abandoning the label of professionalism.

01. Content-Driven User "Diversion"

If Zhihu currently has a need for "self-reinvention," the ongoing adjustment in its revenue structure can be seen as the first entry point. In the third quarter, Zhihu's paid reading revenue was 459 million yuan, marketing services revenue was 257 million yuan, and vocational education revenue was 105 million yuan.

Advertising is the most natural monetization model for traffic. For a long time, advertising revenue has accounted for a significant portion of Zhihu's revenue. However, the contradiction lies in Zhihu's positioning as a professional community, making its users particularly sensitive to advertising content. Under high expectations, excessive commercialization significantly impairs the user experience.

Zhihu has experienced periods of aggressive commercialization. Looking back at its product development history, Zhihu launched in 2011, abolished the invitation system and opened registration on the web in 2013. In the following years, user base growth, Zhihu Live, paid Q&A, and Yanxian Membership launched, with product upgrades and commercialization explorations remaining relatively coordinated.

In 2020, Zhihu launched the advertising product "Zhi+" to help advertisers connect more efficiently with creators within the community. Undoubtedly, higher expectations were placed on the advertising business. However, after the rise of short video forces represented by Douyin and Kuaishou, the competitive environment suddenly became severe. According to QuestMobile, short video users' time spent online accounted for 25.7% in December 2021, surpassing instant messaging to become the internet format with the highest proportion of user time spent online. In the same year, Zhihu went public on the US stock market.

In this quarter's earnings call, management's summary of the advertising business centered on one core idea: reducing low-quality, untrustworthy commercial content and leveraging Zhihu's differentiated advantage for high-value brands. Actively screening advertising partnerships and focusing on the alignment between community content and advertising can be seen as a correction to the old structure that was overly reliant on the advertising business. According to official data, the exposure of low-quality content on the recommendation page decreased by over 40% compared to the beginning of the year, and negative user feedback decreased by over 30%.

After replacing the main engine, content payment took the lead.

For Zhihu, higher user expectations shape the platform's tone and evaluation. Starting as a "knowledge-sharing" and "professional community," users' primary expectation of Zhihu has always been high-quality knowledge and insights. In other words, content is the platform's core asset, and directly converting content into commercial value can better coordinate the production-sales cycle of the community and maintain the creator ecosystem. Paid reading and vocational education both fall under the category of content payment but differ in form, purpose, and audience, each serving specific user needs.

In the first quarter of this year, the paid reading business became Zhihu's largest source of revenue, accounting for 46.8%. This quarter, this business's revenue share exceeded 50% for the first time, reaching 54%.

Paid reading targets a broader audience, presenting a comprehensive content library. In recent years, the "Yanyan Stories" section has stood out the most. Native creators incubated within the Q&A community can earn revenue through the membership system. This unique content supply system has spawned more compact, value-oriented content, allowing Zhihu to capitalize on the trend from short stories to short dramas.

Users with higher content requirements, emphasizing personal skill enhancement and application, are categorized as "professionals," with their needs met by the vocational education business. Currently, vocational education is still in the stage of improving efficiency and accelerating loss reduction, with management aiming for break-even by 2025.

Thus, Zhihu's strategy has a new composition, still centered on creating value through high-quality content but attempting to differentiate between light and heavy use scenarios, promoting commercialization without compromising the experience. On top of this structure, AI, as a special variable, may systematically affect its operational efficiency and long-term planning.

02. AI Bound to the Community, Zhihu Explores Differentiation

The emergence of AI Q&A and search products undoubtedly poses a direct challenge to Zhihu, a "Q&A community." For example, the star product Perplexity AI, established only two years ago, already has over 10 million monthly active users. In July 2024, Perplexity answered approximately 250 million questions, half of the total search responses for the entire year of 2023.

A simple comparison between the two shows that AI's advantages lie in its broad coverage, with large models integrating vast amounts of internet data, allowing content expansion at a much faster rate than UGC content; and its immediacy, accuracy, and efficiency, directly generating answers based on user questions, reducing the process of connecting questioners and answerers and handling large amounts of data. In contrast, the advantages of traditional Q&A communities like Zhihu seem to be limited to the word "community."

"If you can't beat them, join them" was the first impression of many people regarding Zhihu's "Zhihu Direct Answer" product. Since 2023, Zhihu has successively launched the "Zhihai Map AI" large model and the community-native AI feature "Discover·AI Search." Zhihu Direct Answer is seen as a "critical leap" for Zhihu in AI application. In October this year, Zhihu Direct Answer launched the "Professional Search" feature.

An unavoidable premise is that at this stage, most internet companies have launched AI applications tailored to their respective businesses, with chatbots, search engines, and design tools being the main types. However, due to the chaotic competitive landscape, no overwhelmingly dominant or governing application has yet emerged. The standardization and ecological integration of AI technology are still in the exploratory stage, changing user habits while adapting to their diverse needs and seeking new paradigms.

In an interview with Tencent Technology in October, Zhou Yuan discussed his understanding of AI search. "The typical scenario for traditional search is opening dozens of tabs in a browser, reading through them yourself, and then summarizing. AI search reverses this process; you see the summary first. If interested, you can either delve deeper with follow-up questions or further research around the source."

In this broken-down process, Zhihu's differentiation lies in the second half: after seeing the AI-generated summary, users can trace back to the community that provided the content, ask follow-up questions, or conduct further research. It attempts to address two issues with AI search from the side: first, returning the judgment of answer quality to users to avoid illusion interference; second, addressing user stickiness issues after the interaction process is simplified by guiding users to the community to stimulate more communication and interaction.

Combined with the product descriptions in the earnings call, the current launch of the professional search feature aims to introduce more professional knowledge in vertical fields to meet the needs of different workflows. However, it also supports users in uploading their own research materials for customized Q&A and in-depth interpretation, and its future upgrade direction remains integration with the Zhihu community.

Based on this, we believe that Zhihu's AI search is more like a community "activation" tool that uses technology to connect the community with front-end users, ultimately driven by creators and content. As of the third quarter, the platform's content volume continued to grow, with Zhihu's cumulative content creation reaching 855 million, a year-on-year increase of 14.9%, and the platform's cumulative content creators reaching 77.7 million, a year-on-year increase of 11.6%.

If some of the negative evaluations of Zhihu in the past stemmed from the complexity of information filtering and the limitations of interaction forms, AI can indeed provide a solution. The premise is that it can reorganize the path between content creation and consumption, making "browsing Zhihu" popular again and encouraging more creators to return.

03. Concluding Remarks

In the Q&A community space, Zhihu has already experienced a battle. In 2016, Toutiao launched "Toutiao Q&A" and renamed it "Wukong Q&A" two months later, investing significant resources to try to capture the market, including poaching big Vs from Zhihu and offering high subsidies.

However, in 2021, Zhihu's tenth anniversary, Wukong Q&A ceased services, and Zhihu went public the same year. When reflecting on this competition, it was frequently mentioned that Wukong Q&A failed to form a community effect, could not provide continuous content production, and creators struggled to monetize their content. Even with the flow of Toutiao's traffic and high subsidies, it was difficult to quickly establish a new Q&A community.

Grasping the "community" aspect allowed Zhihu to win the industry competition. Now, when the very existence of Q&A communities is being questioned, its chosen solution is still to rely on the community, support creators, and trust users.

After all, essentially, there is still a group of people on the internet who are unwilling to give up the territory of long articles and deep reading.

*The title image and images in the text are sourced from the internet.