This article studies the Zhihu platform from the perspective of how brands conduct content marketing. Not from the creator's perspective.

01 Why do content marketing on Zhihu?

Zhihu has at least 70 million monthly active users, but the amount of marketing experience shared on Zhihu is far less than that on Xiaohongshu.

We can use a formula to roughly compare the advertising scale of each platform. Advertising scale = K × daily active users × average stay time × number of content consumed per unit time.

Zhihu's daily active users are about half of Xiaohongshu's. Zhihu mainly has long content, while Xiaohongshu's content is shorter. Assuming that users stay on the two platforms for the same length of time, the average user consumption of content on Xiaohongshu will be 5 times that of Zhihu.

Therefore, the advertising scale of the two will differ by 10 times. Similarly, Douyin’s daily active users are 5 times that of Xiaohongshu, the duration of stay is 2-3 times that of Xiaohongshu, and the content consumed per unit time is 2-3 times that of Xiaohongshu. The advertising scale of Douyin will be at least 20-50 times that of Xiaohongshu.

Although the scale of Zhihu's advertising is not large, due to the large number of Zhihu users and their strong search habits, it is still very important to have a layout on Zhihu for products that do not require instant decision-making, require trade-offs, and require reference opinions.

For example, electronic products with a unit price of more than 100 yuan and medicines are often products that do not require immediate decision-making and often require reference opinions. When it comes to the purchase of these products, a relatively high proportion of users seek advice through search and other means, and decide which product to buy based on the search results they see.

I conducted a questionnaire survey with thousands of responses on an electronic product that I once managed. The average order value was between 200 and 500 yuan, and more than 70% of the buyers sought various opinions before purchasing, rather than making impulse purchases. Among the channels for soliciting opinions, Zhihu ranks third, accounting for about 15%.

For a user who is accustomed to purchasing some products by searching for opinions, his search tools are usually Baidu, Zhihu, Xiaohongshu, Bilibili, and vertical platforms. Users of different products have different search preferences.

02 Where does Zhihu’s traffic come from?

The first information page that users see when they open the APP is the main gathering place for traffic. The main source of Zhihu traffic comes from the recommendation page, followed by search results and hot lists. Similarly, a rough estimate is given for the reader's reference only to facilitate the understanding of the magnitude, and the estimate is not necessarily accurate.

Content types for recommendation pages: question answers, videos, column articles, and advertising cards . Answers and videos receive the majority of traffic, while columns receive only one-fifth to one-tenth of the traffic of answers.

03 The core logic of Zhihu’s recommended content

Open rate (number of reads/exposures)

Ratio of likes to views (number of likes/number of reads)

Interaction rate ((Likes, Comments, Collections, Likes)/Number of reads)

Complete reading rate (number of completed readings/number of readings)

Release time

Approval rate (number of approvals/number of votes)

Account weight (number of followers, salt value, excellent answerer and other labels, paid membership, real-name authentication, etc.)

I studied my multiple accounts with 1 million views and 300+ pieces of content, and analyzed the various influencing factors.

The interaction rate and like rate have little impact on the number of readings . Many answers or articles with high reading volumes have an interaction rate even lower than 0.5%.

The number of comments is strongly correlated with the number of reads . Content with more comments generally has higher read counts.

The correlation between each indicator and the reading volume is ranked from strong to weak: number of comments > number of likes > number of collections > number of thanks .

For the default answer ranking of a question, the higher the number of likes, the higher the ranking. The ranking is also related to factors such as interaction rate and open rate, but the number of likes is the core indicator.

For cold start of content, the core indicator is the open rate. Content with a high open rate gets a high initial exposure. Like Xiaohongshu, Zhihu does not publish exposure data, so the opening rate cannot be directly calculated.

The number of fans is helpful, but it is not decisive. For accounts at level 4 and with hundreds of followers, the probability of creating a popular answer is not significantly lower than that of big Vs. The reason why it is LV4 is that starting from LV4, one can obtain the authority to recommend content on one’s own. You can give your answer a certain amount of initial exposure to reduce the chances of it being buried. I'm afraid that someone will argue with me, so let me add that the core of Zhihu big V is still the ability to produce content. The number of fans will not have a significant help in content marketing. If a big V with 200,000 fans posts a hard advertisement, the number of likes will still be almost 0.

Here is the answer data of a Zhihu celebrity with 1.34 million followers in a certain period of time:

Zhihu has no obvious preference for new answers published recently.

Around 2018, a Zhihu algorithm was widely circulated. The interpretation of the Zhihu algorithm at that time was as follows.

This algorithm ranks answers to the same question. The more likes an answer gets for a question, the higher its ranking will be. The more likes an answer has, the more it will be affected by the dislike rate.

Extract 3 core logics:

1. Open rate

The opening rate of content is very important. It determines whether you can get higher initial traffic and whether you can get higher traffic after getting interaction.

2. Number of favorable comments

The number of likes determines whether you can get a better ranking for a question, and the number of likes and comments determines whether you can get more recommended traffic.

3. Search open rate

Making your content the most clickable content in the search results will help your content get a better ranking in the search results.

In Zhihu’s search results, your content will appear with about 43 words near the keyword. It is usually not at the beginning of the article, nor is it necessarily the part where the keyword appears for the first time. A picture will be displayed, which is not necessarily the header picture. I guess the system's rules for grabbing preview pictures give priority to pictures with high clarity, obvious contrast, and clear subject.

Poor PR articles will be ranked lower and lower in Zhihu search results. Content with strong user feedback will be ranked higher and higher in Zhihu search results.

The Zhihu content displayed at the top of Baidu's search results is not sorted by the number of likes or other interactions. The content with the keywords at the front of the title will be displayed first.

There are several possible ways to rank Zhihu content in Baidu search results: based on the search result opening rate, based on the number of links to the content in Zhihu (i.e. Zhihu's related recommendations for the content). Based on how recently the content has been opened.

04 How does Zhihu push content to users?

The information presented on a Zhihu user's recommendation page is mainly based on the following three logics:

Based on the user's basic information , including school, gender, geographic location, etc.

Recommendations based on search behavior, topics of interest, questions of interest, content viewed or interacted with, and their related content .

Recommend content that has been liked by the people you follow .

Considering that the number of followers of an average Zhihu user is usually less than 50, or even in the single digits, there is very little relationship data, and there is little content that can be pushed based on basic information.

05 Example of content marketing strategy for Zhihu

Suppose we want to carry out content marketing layout on Zhihu for a new product with an estimated annual sales of 100 million yuan and a price of 200 yuan .

Some brands will choose to post dozens or hundreds of new product PR articles on Zhihu, looking for everyone from big V to small V, to test the waters.

I have compiled statistics on this type of advertising strategy for several brands. The average number of likes for each article is between 0 and 5, and even big Vs with hundreds of thousands of followers are no exception. These contents will receive very few readers and it will be difficult for them to rank high in Zhihu search results. You may not even find it after searching dozens of pages in Baidu search results.

Some brands will cooperate with Zhihu officials to organize an event. This method may not only have a low input-output ratio, but also test the team's planning ability. The single investment is relatively high, so few brands dare to do this.

Some brands choose to place ads through Zhi+ and other methods, and the cost of a single reading is around 0.5 yuan or more. Assuming the conversion rate reaches 1%, the customer acquisition cost is 50 yuan. For many brands this is unaffordable.

There are very few brands that have mastered Zhihu.

Based on our hypothetical new product, we first conduct a data survey.

Basic research. After investigation, it was found that the annual reading volume of questions related to the main keywords in this market segment is about 5 million. There are about 50-100 popular questions. There are two main competitors with a lot of content on Zhihu.

Strategic thinking. When users search for category keywords related to this new product, they are likely to see content related to this new product and develop a tendency to favor this new product or have a desire to buy it.

Continuity strategy.

1. Cover all popular and old questions , and provide at least one piece of content for each question. Focus on several key and old issues and lay out content that has the potential to become popular.

2. Ask new questions in 5-10 directions . This question should either be what the user needs to see in the search results, or it should be able to stimulate the user's desire to read or participate. Prepare 3-5 different pieces of content for each question in advance. And a potential explosive content is prepared for each question. Maintain these issues regularly to ensure that the answers to these questions are mainly optimistic about this brand and this new product.

06 How to evaluate the data of Zhihu’s content marketing?

Zhihu’s content marketing has a long payback period. Most of the reading volume does not appear in one or two weeks after the content is released, but for half a year or even several years.

In the short term, you can pay attention to:

1. The growth trend of content reading volume,

2. The content’s position in each question. A good position means a good growth in future reading volume.

3. Potentially popular content can be fueled by operations to make it more popular.

Evaluate the delivery effect based on the search result placeholders and answer placeholders under the questions in the layout content. Evaluate the effectiveness of content delivery by predicting long-term reading volume based on the content’s growth trend.

Zhihu is an important platform for maintaining brand reputation. Although it is difficult to measure the conversion rate, it is an important source of reference for people who need to rely on opinions to decide on their purchases. For brands with a scale of hundreds of millions but cannot afford to spend a lot of advertising money, they can do cost-effective content marketing on Zhihu. Don't hesitate, just go for it.

To summarize:

Zhihu’s traffic mainly comes from system-recommended answers to questions and keyword searches.

To create popular content on Zhihu, the first thing to pay attention to is the opening rate of the content , that is, whether the content can attract readers. The second thing to pay attention to is the number of likes and comments , which is related to the ranking of the answer under the question. The third thing to pay attention to is whether the content is what the searcher wants to see , which is related to the subsequent search ranking of the content.

Zhihu will recommend new content to users based on the keywords they have searched, the topics they follow, the questions they are concerned about, and the content they have browsed or interacted with. Therefore, the layout of the content should be focused, penetrate a keyword, penetrate a group of related issues, and penetrate a certain topic. These directions will allow a large amount of content to form a synergy and promote each other, thereby improving the efficiency of content marketing.

Author: Jiang Liu

Source: Jiang Liu