The logic and method of operating a Zhihu account at zero cost

Zhihu started operating in September 2019, and it has been half a year so far. At that time, I started a business with my friends, mainly focusing on WeChat traffic. We tried no less than 10 methods of attracting traffic, and in the end only the method of Zhihu was left.

To be honest, I have only achieved 12,000 followers in half a year, and only attracted 100-200 fans to the official account every day (this period was quite special, with 1,000-2,000 fans attracted, and the number of fans of the official account reached 140,000). However, if you calculate the ROI, this conversion is still good.

Today I will review how I created Zhihu. Since I am currently working on K12, what I talk about may not be applicable to all industries. I will try my best to explain my logic and essence clearly.

Here is some rough data from Zhihu: the number of collections is about 4 times the number of likes.

This article is mainly divided into the following parts:

1. Zhihu Positioning
2. Operational Strategy
3. Drainage method
4. Some scalable guesses

1. Zhihu Positioning

The starting point of positioning lies in your purpose and your resources.

My purpose at that time was simply to attract traffic, and I was basically the only resource I had.

Resources meant it was difficult for me to produce content myself (at the time I was mainly doing quick testing and working in several directions, so it was difficult to focus on creating content, and I didn’t have the ability to do so), so the only thing I could do was move content around.

So what to carry?

Back to the purpose, to attract target groups, practical information is undoubtedly the best choice (currently Zhihu’s traffic is basically attracted by practical information), and in order to conduct rapid testing, the source of the practical information content you move must be large enough.

So I finally chose K12 learning materials and maintained an update frequency of one article per day.

2. Operational Strategy

There are three main strategies here: improving the weight of articles/answers , spreading keywords and finding hot spots

2.1 Improve article/answer weight

Zhihu's recommendation system is ranked according to weight. The higher the weight, the more likely it is to be recommended to the user group, and when users actively search, the article will be ranked higher.

The weight is calculated as follows:

If you don’t understand, you can just look at the picture below:

The main points I use to increase weight are as follows:

2.1.1 High-quality content

High-quality content determines the weight of your article, and will be of great benefit in both the cold start (the interaction and completion rate of the content obtained within 30 minutes of posting the article determines the first wave of recommendations) and the subsequent improvement of weight.

To dig out high-quality content, you have to start from the source of the content. I won’t go into details here and will just give you my own example.

My content sources mainly come from QQ groups, Baidu Wenku and tweets from some peers, so it is relatively easy to judge the quality. The QQ group refers to the number of downloads, Baidu Wenku refers to the ratings, and tweets basically refer to the number of readings.

2.1.2 Support Answers

Yang Dadao is a little trick I developed on my own.

It is mainly divided into three steps: guiding interaction - accumulating weight - attracting copywriting.

Step 1: Guide interaction

Beginning: Material name + Material download method at the end of the article + Guide to three links + Follow

Middle: Data content

Ending: Material name + If you need to obtain it, please reply to the material name in the comment area after three likes (this step is to induce interaction)

Step 2: Sedimentation weight

After the first step, if the content quality of your article is good (the quality of the content determines the user's desire to take the next step and interact with you), you basically have the startup volume. At this time, it will take a certain amount of time to accumulate weight (that is, to obtain a certain number of likes, thanks, collections, etc.). During this time, you basically have to manually send information to users, which is a rather tiring process.

Step 3: Lead generation copywriting

When your article has accumulated a certain weight, you can edit it again and change the copy to one that attracts traffic (traffic will be discussed in detail later). This cycle is generally 7 days. (No research, just based on feeling)

2.2 Keyword placement

My positioning is in a relatively niche field, and I am not among the core user groups on Zhihu, so it is difficult for me to get a lot of traffic through an article. In addition, I learned a sentence from an article shared by an Alibaba boss: traffic is hidden in relationship chains and keywords.

So I decided to use keywords, that is, no matter what information you search for, you must include me . The craziest keyword is that 8 out of the first 10 pieces of content are mine. Of course, it is not a random laying. There are mainly two steps.

2.2.1 Long-tail keywords

First define what long tail keywords are. Based on my own industry, I defined Chinese, mathematics and English as first-level keywords, and other subjects as second-level keywords. The categories contained in different subjects are further reduced to one level.

For example, mathematics is level one and physics is level two. For real math questions, quadratic equations are level two, and physics experiments are level three.

What I consider to be long-tail keywords are those after the second level. So at that time, I mainly worked on the second-level category. Several of my products ranked first when searching for this keyword, and there might be more than one in the top 5 that was mine.

The first three results for a certain keyword are mine

2.2.2 Find potential hits from core keywords

It is impossible not to touch the first-level keywords, because the traffic is very large and there are always points that can be divided. Since you can’t grab the first one for the first-level keyword, you can try the second and third ones.

Therefore, I basically don’t answer the first answer to the core keyword. Instead, I look for answers that may become popular. I have summarized three characteristics of potential popular answers.

High views, medium attention, few replies . The corresponding data is approximately 10000-100-10.

This kind of answer is easy to get popular if the content is not bad. I also made a second post with several first-level keywords.

And keywords are divided into answers and articles (there are 4 types of content published on Zhihu, answering questions, publishing articles, publishing videos, and publishing ideas, the main content is articles and questions). Here I will give examples:

answer:

1. Search for high school politics to see what is being asked;
2. Check the answer data to see if it meets the three characteristics;
3. Find content related to the question from content sources;
4. Start raising answers.

For example, for the question above, the corresponding content may be a collection of high school political knowledge, high school political recitation mnemonics, etc.

article:

The content can cover all questions, so if some content cannot find matching answers, these contents can be used to lay out keywords.

What I do is to change a secondary keyword every day, and the content of the answer can also be synchronized to the article.

2.3 Finding Hotspots

Hot spots are the fastest way to increase volume, no doubt about it. The manifestation of hot spots is the change in popularity of keywords, which can be observed through Baidu Index and WeChat Index.

For details, please refer to my previous article on the issue of riding on the hot topic. Moreover, some keywords have never appeared on Zhihu. For example, during the mid-term, I judged that the traffic would increase, but Zhihu did not have the corresponding keywords, so my answer went up and eventually received tens of thousands of views.

To sum up: strategies are not separated from each other, but gradually complement each other. The second and third strategies are built on the basis of the first strategy, and content is the core.

3. Traffic diversion strategy

The key to traffic generation is to let users see you anywhere they can see you and remember you (provided you do not violate the Zhihu rules and become a Yanxuan member to ensure safety). It is very costly for users to go from Zhihu to your official account to search for your name and follow you, so you must give users motivation.

3.1 Personalized Signature

Needless to say, this is basically the most common ad position. But don’t just put a public account name, you must give users motivation, like me: public account name + update a learning material every day (for reference only)

3.2 Ideas

Ideas are actually similar to Moments, not public content, so Zhihu does not strictly control this area, making it a good place for advertising. And from what I have observed, my ideas have a good number of readers.

3.3 Zhihu Column

A Zhihu column is actually a place to collect articles. The introduction can also attract traffic, but the column is not efficient in attracting traffic, and the weight given by Zhihu is not enough.

3.4 Private Message

Zhihu’s attitude towards attracting traffic through articles has always been confusing, and I don’t understand it clearly. But sending private messages to attract traffic is a better way. Although it is more tiring, it can add value.

It has been mentioned in Yangda that to guide users to pay attention, you send information to it. During this period, the content itself cannot attract traffic, so you can standardize the private messages and just copy and paste them to send to users, but you must give motivation.

3.5 The content itself

This is the most common way to attract traffic to Zhihu, which is to promote your own official account in the article content.

For example, Daisy Witch will put some of Douban’s 100 selected e-books at the end of the article, and you can reply “balabala” in the official account to download them. After I finish answering, I will change the text to "If you need to download, please reply Balabala" on the official account.

So if your content is high-quality, users will be more motivated to pay attention to it.

4. Scale-up hypothesis

I spend about 1-2 hours on Zhihu every day, and I also spend time on the official account. So resources determine that it is difficult for me to do this to the extreme (the main reason is of course laziness), so Zhihu’s performance is not very good and there is no scale. Here I will put down how I think I can do it bigger.

4.1 Matrix number construction

The logic is similar to that of building a public account matrix, and it still starts from keywords, for example, I am working on junior high school materials. Vertically, we can work on elementary schools and high schools, and horizontally, we can work on question banks, policies, etc.

With a large trumpet, cold starting a small trumpet will not be difficult.

4.2 Accelerate the increase of keyword density

At that time, I published one article per day and answered one question every three days. If you are not too diligent, you can adjust it to 3 articles and 1 answer a day. This way you will have more opportunities to get exposure, but of course you will be more tired.

4.3 Building the Zhihu community

Zhihu currently attaches great importance to circles and has even replaced the original publishing button.

And under some answers you will be recommended to corresponding circles, which can be considered a bonus. I was in the second batch to register for the circle. Currently, there are only 720 people in the circle. At the beginning, the official would only help you promote it after a certain number of cold starts. I missed it. There was once a circle that gained 20,000 followers overnight. Friends who are interested can give it a try.

Here I mainly share with you the logic and methods I use to run Zhihu. It is not my turn to explain the entire Zhihu in detail. Some frameworks and details have been summarized by predecessors. The algorithm diagram I showed above was also cut from the PDF of my predecessors. If you want to understand Zhihu operations more deeply, you can download the PDF mentioned above.

Author: Zhang Yifan Y

Source: Zhang Yifan Y