Analyzing the hot-selling machine - the community ecology of Toutiao

As one of ByteDance's popular apps, Toutiao is a model of information flow apps, and it also has the native characteristics of the ByteDance system.

Compared with traditional news and information, Toutiao’s characteristics lie in its powerful recommendation mechanism and decentralized content production scenario. In this regard, Toutiao can be regarded as both a product of the first wave of artificial intelligence applications and a product of Internet 2.0.

This article intends to briefly analyze the community ecology of this product from four perspectives: content production, content distribution, embedded search services, and content monetization.

1. Decentralized content production mechanism

Traditional news and information applications are generally produced by a professional editing team. In this case, the content itself does not rely on traffic, and its seriousness and professionalism are relatively high.

In a sense, this is a supply-based production - the choice of the product itself does not lie with the consumers, but with the professional editorial team. The prominent feature of this model is centralization .

We can still see some similar traces of this model in Tencent News today.

The content production of Toutiao is completely decentralized, and all its content comes from self-media content producers. The platform itself can only use monetization mechanisms to guide the content production of self-media, but cannot achieve direct control.

This is a consumer-oriented production , and the content is designed from the beginning to cater to readers' preferences.

The resulting problems are the massive amount of content information, homogeneity, and uneven quality. Furthermore, it also encourages the trend of cross-platform plagiarism and copying.

The advantages are that it is more friendly, focuses on entertainment and practicality. The outstanding feature of this model is decentralization, which gives the right to produce content to the users themselves.

It is worth noting that decentralized content production does not mean that users have complete autonomy in content production. In addition to being affected by the platform's own review mechanism, self-media producers must also face the kidnapping of traffic. It is difficult for self-media that cannot be monetized to continuously output content.

In addition, this model does not mean that the content production is unprofessional. The fact is just the opposite. Today, as we dig deep into segmented vertical fields, self-media brands have even surpassed the past in terms of specialization in content production.

2. Content distribution mechanism based on recommendation

Although Toutiao is a real Internet 2.0 platform, this does not mean that the entire closed loop from content production to content consumption is decentralized. On the contrary, Toutiao is far more centralized than traditional applications. The secret of this centralization lies in its content distribution mechanism.

In the era of self-media, content production is directly related to traffic. Self-media that cannot generate traffic often face difficulties in monetization. Different from the traditional professional editing team, in today's era of massive and homogeneous content production, self-media content producers must rely on stable traffic to ensure basic economic sources.

When it comes to traffic, in addition to the quality of the work itself and the scope of its audience, the most important thing is the content distribution mechanism. There are three basic content distribution mechanisms: distribution mechanism based on personalized recommendation, distribution mechanism based on social network, and distribution mechanism based on page display .

Among them, the distribution mechanism based on page display is the most traditional method. Usually, the platform only needs to distribute the works to each sub-tab page in chronological order. If we modify it a little bit, we can create a hot list based on the number of views. For traditional content production methods, this distribution mechanism is basically effective. However, in the era of self-media, content production is massive. The article submitted by the author is likely to be buried in the massive amount of articles and not get any reading volume. Therefore, in the era of self-media, the mainstream distribution mechanisms are only recommendation-based and social-based.

The interesting thing is that Toutiao’s social support is not very good.

When readers click to enter the app, the first page displayed is the recommendation page, and the follow-up page directly related to social distribution is hidden on the left.

According to people's preconceived habits, the chance of clicking on the follow page is not high. In addition, users may be more accustomed to swiping to the right, so the follow page on the far left is even less likely to be opened.

In addition to the follow page, the circle function supported by Toutiao itself is more like Knowledge Planet rather than WeChat Moments. Originally, not everyone can create a circle, and many circles are charged. This means that the circle has little effect on newcomers, but is a monetization channel for veterans. This design arrangement makes the social network-based distribution model a kind of foil for Toutiao.

Because of this, Toutiao celebrity Sun Honghe warned newcomers not to pay too much attention to the number of fans - from a realistic perspective, Toutiao's fan stickiness is not high, especially compared to applications with prominent social features such as Kuaishou.

Therefore, the core of Toutiao’s content distribution lies in the recommendation mechanism - compared to cashing out fans, Toutiao is more suitable for creating popular articles.

We have no way of considering the specific details of Toutiao's recommendation algorithm, but we can briefly explain the principle - Toutiao's recommendation mechanism is implemented in a layered manner. The machine will parse the article, then derive keywords and determine the appropriate target audience. This kind of delivery is often divided into echelons——

Generally speaking, the machine will first push it to about 1,000 people to test the overall effect of reading. When the performance of clicks, readings, comments and forwarding is good, the machine will expand the delivery range to a group of about 10,000 people, and repeat this process continuously. For articles with poor performance, the machine will reduce the recommendation volume. Without enough recommendations, there generally won’t be enough readings.

Because of this, even if you don’t have fans on Toutiao, it is possible to create a popular article. From this perspective, Toutiao is indeed a recommendation-driven organization, and it is a veritable hit-making machine.

The disadvantage of this distribution mechanism is that content producers have very few rights to distribute content . The distribution mechanism based on social networks is generally not effective in Toutiao. This model has enabled Toutiao to achieve a high degree of centralization in the field of content distribution. Content producers must continuously improve the quality of their content under the official guidance in order to survive. In this regard, the headlines are indeed a bit overbearing.

3. Alternative source of traffic: embedded search service

Although the main source of content traffic is the distribution mechanism, active acquisition can also provide traffic for content producers. Toutiao’s built-in full-network search provides users with a channel to actively obtain information, and also provides content producers with a new idea for obtaining traffic.

From an experience perspective, there is not much difference between using the Toutiao APP to search and read recommended messages. Similar to Baidu, ByteDance also maximizes the use of its own information platform. The search features are generally similar to other full-network searches. The account information that has been authenticated and has a high weight is generally placed in front, and the accounts with the same weight and the same type of information generally select the one with the keywords at the front. Therefore, self-media people can also try to improve the long-term value of their own content through keyword optimization.

Generally speaking, the content produced by self-media can be divided into fast-moving and long-lasting types .

The former is more entertaining and a one-time consumption of attention, which is extremely important for the platform to strengthen user stickiness.

The latter is practical, has instrumental and reference value, and can provide long-term value to potential users.

The key to search is to provide space for further exploration of the latter.

You should know that most of the information recommended by headlines is relatively current. In other words, articles that fail to become popular at the time will find it difficult to gain readership in the future. This shows that search services can make up for this shortcoming to a certain extent and provide impetus for the production of long-term valuable content.

4. Diversified content monetization mechanisms

From content production to content distribution, it is not enough to constitute a complete business chain. Monetization is what platforms and creators care about most. The reason why Toutiao was able to defeat many competitors is closely related to its diversified efforts in content monetization.

Judging from the information presented on the platform, there are four main content monetization mechanisms provided by Toutiao: advertising sharing, knowledge payment, e-commerce sales and paid consulting. Advertising revenue includes official advertising and self-operated advertising, and may also include the Star Map Project. Knowledge payment is the so-called headline column. This is very similar to Zhihu. E-commerce sales are mainly through Toutiao stores, and also include some forms similar to Weibo Taoke. Toutiao Circle is similar to paid consultation.

The diversified distribution mechanism not only provides sufficient revenue sources for the platform, but also helps to motivate content producers. However, self-media ultimately follows the 80/20 rule. Wherever there is self-media, there will never be a shortage of cheap brick-moving laborers. Therefore, it is actually a question of how long self-media, especially long-form text and picture self-media, can last. After all, long-form blogs ultimately lost to Weibo.

5. Conclusion

Overall, Toutiao has achieved a complete ecological closed loop from content production and distribution to content consumption and monetization. This product ecosystem is characterized by decentralized content production and highly centralized content distribution. It is with the help of the latter that the platform is able to adjust quickly, guide the direction of self-media content production, and create popular articles. The flaw is that this planned economy-style operating mechanism is not conducive to personalization and social sedimentation, and content producers are just a screw in the whole machine.

Currently, Toutiao and Baijiahao are leading the creation of self-media content. The former is based on recommendation, and the latter is based on search. Unlike short video self-media platforms, there is no graphic and text self-media platform similar to Kuaishou. WeChat public accounts have prominent social features, but lack a discovery portal and recommendation mechanism, and are becoming isolated, so it is not suitable to compare them with Kuaishou for the time being.

If we call Toutiao and Baijiahao the front-line positions of graphic and text self-media, then WeChat public accounts are the military camp behind the scenes, which is important but deserted.

Recently, WeChat officials have also been making a series of adjustments. As for how effective it will be, it depends on how determined Zhang Xiaolong is to break the closed nature of WeChat.

Author: Shu Xiao Zhai

Source: Shu Xiao Zhai