Zhihu institutional account operation and promotion practice

When companies are flocking to popular platforms such as Douyin, Bilibili, and Video Account to seek traffic growth, I choose Zhihu as the main battlefield for new media marketing. Because in my opinion, choosing a new media marketing channel that suits your product is far more important than whether the channel itself is popular.

It turned out that this strategy was correct. In 2 months, with the help of Zhihu institutional accounts , we still earned enough traffic for our products. Today's article is to review the entire operational practice, including:

1. How should Zhihu institutional accounts be positioned?
2. How to continue to find suitable topics (Q&A)?
3. What are the techniques for creating content on Zhihu?
4. How to do Zhihu Q&A SEO?

Next, let’s get straight to the point. Enjoy~

1. Positioning: The positioning of institutional accounts is to anchor products

The so-called positioning can be quickly explained with a formula: Positioning = the segmented population served + the segmented needs met. Applying this to the positioning of Zhihu institutional accounts means answering the question of "what kind of content to publish for whom".

1. For whom: Segmented populations served

This is a free point question.

Zhihu institutional accounts are essentially for serving enterprises and are a marketing and promotion channel for products. In other words, the target group it targets is actually equivalent to the target group targeted by the product. Therefore, we no longer need to go through the process of research and visits to quickly target the niche groups that Zhihu’s institutional accounts serve.

For example, if Shimo Documents segments its target groups based on Zhihu "topics", it can be divided into:

• Core users: Graphite Document, Graphite Document Enterprise Edition;
• Target users: online documents, multi-person collaborative documents;
• Potential users——office collaboration and efficiency improvement;
• All users——all users of the entire site (Zhihu + Baidu).

Core users are the group that is currently using/has used the product, while target users and potential users are the groups that we should focus on winning over (the above are only used as examples, in fact, Graphite Docs currently does not operate a Zhihu institutional account).

2. What kind of content to publish: satisfying segmented needs

Regarding this question, many people like to use differentiation as a general answer. But in fact, Zhihu institutional accounts do not need to talk about content differentiation like Zhihu personal accounts or other self-media accounts.

In my opinion, the biggest difference between Zhihu’s institutional accounts is actually the difference in their respective corporate products. What Zhihu institutional accounts need to do is to highlight and disseminate this differentiation (including functional and scenario-based) in content.

When talking about content, we cannot avoid a question, that is, what style of content publishing is adopted, or what kind of personality do we hope to create with the help of content?

My answer is that it is best to have an expert persona, which is determined by the platform attributes of Zhihu. It doesn't matter whether the expert is serious, playful, or neighborly.

Why is the positioning of Zhihu institutional accounts so important?

Because you don’t have any positioning, the content you post will be very confusing and your account will lack professionalism. This will directly lead to a result that Zhihu will not recommend stable traffic to your account, which will not only affect the natural ranking of your answers, but also affect subsequent SEO operations (if you are in a hurry, you can go directly to the fourth part "SEO").

2. Topic Selection: Keyword Library is a Prerequisite

Finding topics on Zhihu is basically the same as finding questions and answers. At this point, some people may say that this is easy. Just search for keywords on the Zhihu site.

In fact, it is not that simple. For example, what keywords should we search for? Among the questions and answers searched out, which ones should be answered first, and which ones should be answered later or even not answered at all? In addition to searching within the site, do we have other efficient and convenient search methods?

These are the questions we need to answer.

1. Build your keyword library

There are two benefits to building a keyword library:

1. Aim at the right target: You can accurately find potential questions and answers and carry out content operations in a rhythmic manner;
2. Check for gaps: You can check at any time which keywords are already covered and which ones are yet to be laid out.

How to build it specifically? There are also two methods (still taking Graphite Document as an example):

1) Find keywords for products and product functions

This is the most direct and effective method in the early stage of the account.

• Product name: Graphite Document, shimo.in, Graphite Document app, Graphite Document mobile version, Graphite Document web page, Graphite Document enterprise version...
• Names of competing products: Tencent Docs, Youdao Cloud Notes, Evernote, OneNote, Kingsoft Docs, Weizhis Notes... Zhihu has a large number of software comparison questions and answers.
• Product features: online documents, multi-person collaboration, team collaboration; create menus, hierarchical titles, insert formulas, import PDF, collaborator name identification...

2) Find keywords for product application scenarios

When we have laid out all the questions and answers containing product and product function keywords, we will encounter another problem: there are no keywords to answer. At this stage, we need to expand the keyword library based on the product's application scenarios:

For example, Shimo Document can start from the perspective of office collaboration and efficiency improvement, and be refined into specific application scenarios such as year-end summary, file management, project management, HR recruitment, etc.

For example, XMind can start from the perspective of improving thinking and organizing knowledge, and then be refined into specific application scenarios such as the cultivation of structured thinking/divergent thinking, the construction of personal knowledge system, and SWOT analysis of career development;

For example, Chuangketie can extend the scene based on the graphic designs that the product can achieve, such as public account cover images, mobile phone posters, long marketing images, business cards, invitation letters, etc.

In summary, what we need to do is to discover the existing scenarios of users, supplement the scenarios that have not been discovered by users, and then condense them into keywords.

2. 6 search methods to find potential Q&A

With the keyword library, we can conduct targeted question and answer searches. Here are 6 search methods for you to share with you, and I will add more when I find more:

1) Search within the site

• Keyword search (assisted by Zhihu Q&A rating plug-in)

In-site keyword search is the simplest and most commonly used search method. However, this method has a shortcoming, which is that it is difficult for us to quickly and intuitively judge the value and potential of a question and answer in the result list (exposure is an important indicator).

I pointed out this shortcoming and of course also proposed a solution, which is to use a Zhihu Q&A scoring plug-in to assist in judgment.

After installing the rating plug-in on Google Chrome/360 Browser, search for specific keywords on the site, such as "website", and the corresponding rating will appear on the right side of each question and answer that appears on the results page. Questions with higher scores are worth answering first, while questions with lower scores can be put on hold.

Note: This plug-in will automatically become invalid after a period of use. You need to reinstall it and then register and log in before you can resume use. Although the whole process is a bit cumbersome, the registration does not require verification, which saves a lot of trouble.

In addition to the Q&A rating, we can also make a comprehensive judgment based on the number of Q&A views and the number of answers.

If a question has a high number of views but currently has few answers, it's worth answering first. Because it means that our answer has a great chance to reach the front row, and through subsequent SEO optimization, it can impact the top 3 positions to gain higher exposure.

As for questions and answers with high views and high numbers of answers, they need to be led by more substantial content.

• Related questions (system-related recommendations)

Many people don’t know that there is a “Related Questions” section on the right side of the Q&A because it only appears when it is opened on a PC web page. In this section, generally 4-5 related questions (sometimes not very relevant) are aggregated, and the selection strategy is the same as above.

• Zhihu system recommendation

For institutional accounts, Zhihu sets regular tasks every week, and as long as you complete the tasks, you can get corresponding rewards.

One of the rewards is "Hotspot Tracking for 1 Week", and the triggering condition is that the organization account completes 7 creations (including Q&A and articles) in one week. It will recommend hot issues in the management center of the institutional account, but most of them have little to do with its own products.

• Competitor account history question and answer search

Compared with the questions and answers recommended by the Zhihu system, manual search is still more reliable at present. In addition to the direct search mentioned at the beginning, we can actually dig out suitable questions and answers from the historical answers of competitor accounts. This is equivalent to joint screening by operators from both sides, and there is a high probability that it is valuable.

To take a step back, even if the Q&A itself is not of great value, from a marketing perspective of attracting more target users, we must occupy such Q&A.

2) Search outside the site

• Baidu keyword search

This is an overlooked, but extremely important search method.

The reason why it is important is that Baidu announced a strategic investment in Zhihu in August 2019. One important change brought about by this close cooperation is that the weight of Zhihu Q&A in Baidu searches has been increased - this is a traffic flow that cannot be ignored.

Therefore, in addition to conducting keyword searches on the Zhihu site, we can also conduct keyword searches on Baidu, and then give priority to the Zhihu questions and answers displayed on the first page of the results page.

This process can be carried out in conjunction with the 5118 Webmaster Toolbox plug-in, which can hide the ads that come out of Baidu searches and help us quickly find target results.

• 5118 Big Data Search

The last method is to use a third-party tool - "5118.com".

With the help of 5118's ranking monitoring, it can pull out all the Zhihu questions and answers corresponding to the search keywords, the result rankings in Baidu PC search, the Baidu PC keyword ranking list, Baidu PC search volume and other data at once, and support export to Excel.

What we need to do is to search for our own keywords and their corresponding Zhihu questions and answers in the exported Excel file based on our own keyword library.

However, this feature requires a paid membership.

Finally, I would like to add another suggestion on the topic selection for Zhihu institutional accounts, which is to try not to choose hot social events, political or military events. Because the institutional account represents the image of the company and its products, if the answer is inappropriate, it can easily cause a public relations crisis.

3. Content: “Why” and “How” are more important

After deciding on the topic, the next step is content creation. In this section, I will analyze the structure and pictures of Zhihu Q&A and answer the following two questions:

1. What is the general structure of the highly praised answers on Zhihu?
2. What are the tips and precautions for adding pictures to Zhihu Q&A?

1. General structure of top-rated answers

Highly-rated answers generally have a structure like this, which can be expressed by a formula: Highly-rated answers = giving results directly + making reasonable and well-founded analysis + ending with interactive comments and asking for three likes. How do you understand this formula specifically?

1) Get straight to the point and give the results

It means that we should summarize the answer in concise words at the beginning of the answer to create attraction, such as:

Recommend 12 completely free and conscientious websites, each of which is completely free and very easy to use, making you regret not knowing them earlier - Zhihu @木子淇, corresponding question: What websites are there that you are reluctant to share?

As a documentary fanatic and jellyfish lover, I have watched hundreds of documentaries, but only these 12 top documentaries attracted me. Every time after watching them, I would exclaim, "It's so good!" I want to watch it again. It covers history, humanities, and the universe. It is definitely worth a watch! Especially the two in the middle - Zhihu @DaisyWitch, corresponding question: So far, what documentaries have you seen that can be called "top documentaries"?

Delivering fried chicken, I can make a net profit of about 40,000 yuan a month, or hundreds of thousands a year. I don't know if this is considered a huge profit. Many people have eaten fried chicken takeout, but few know that this business can make so much money. Maybe this industry is not very eye-catching - Zhihu @林雁, corresponding question: Are there any other highly profitable industries that ordinary people don’t know much about?

The reason for writing this way, in addition to the well-known reason of "attracting users to continue reading", there is another very important reason, which is to attract users to click in. You should know that before Zhihu answers are expanded, their display logic is the same as the public account summary, and the content before the main text will be captured by default.

2) Make a well-reasoned analysis

Once you’ve attracted attention, you need to retain it with substantial content.

So what kind of content is considered substantial and rich?

My answer is that we cannot just introduce “what it is”, but also need to explain “why it is” and “how to solve it”. Zhihu users are not content to stay on the surface of the problem. They like in-depth, unknown and not easily accessible content. It is easier to gain their approval by telling them your deeper knowledge, experience or insights.

For example, this Zhihu question: What is the Feynman technique?

If you simply tell users that this is a "teaching as learning" method that can help you improve your knowledge absorption efficiency, truly understand and learn to apply knowledge, then the answer will most likely sink to the bottom. XMind, an institutional account, gave a good demonstration. It answered like this (the answer is too long, so I only broke down the main structure and key points):

What is the Feynman Technique?

——Learning through teaching.

Specific application method?

——Divide into four steps.

Why is the Feynman Technique so effective?

• Splitting and compressing knowledge;
• understand and simplify knowledge;
• Understanding and additional knowledge (with a mind map drawn by XMind inserted in the middle).

This is the kind of dry goods that Zhihu users like. Even if it contains personal opinions, they are still willing to vote for the answer (as of press time, the number of likes for this answer is close to 10,000, and the exposure it brings to the product is obvious).

Therefore, when you answer a question, if you can keep in mind the assumption that "will others ask me why when reading my answer?" you can write more detailed and useful explanations and descriptions. If you make a point, it will be very helpful to your readers if you explain why you think that.

3) At the end, please interact with the audience and ask for a triple like

Finally, encourage more interactions and encourage more users to participate, pay attention, and comment. You can also post your own product experience link (text links and card links are supported).

Note: Don’t be too aggressive! Don't be too hard! Don't be too hard! Important things should be said three times! Because it will be blocked, the following is a typical counterexample:

2. Zhihu picture matching skills and precautions

If your product is a software tool, you can choose to record a Gif animation when introducing the functions. It is more intuitive than a static image and can also increase the user's stay time. But it should be noted that the Gif image should not be too large (controlled at around 1M), otherwise the user may become impatient and jump out during the loading process.

In addition, for some horizontal screen images with a large amount of information (the images are generally blurred), try to change them into vertical screen images that can also adapt to users' mobile reading to improve the user's reading experience. Of course, if you want to attract traffic, then that’s not a big problem.

At the end of the content section, I would like to share with you a little tip: Zhihu supports answering two similar questions with the same content, which allows institutional accounts to quickly expand their reach. But I suggest adjusting the beginning and some of the content in a targeted manner based on the specific description of each problem.

Note: Don’t think about answering more than 3 questions with one piece of content, because the site management will delete the duplicate content, and serious cases may even result in a ban.

4. SEO: Zhihu also needs to do SEO?

Once the content is successfully published, we can move on to the next step - SEO. Some people may wonder, does Zhihu Q&A also need to do SEO? Isn’t it enough to just answer the questions?

——Of course not.

If we compare content to 1, then SEO is the 0 behind it. The latter is an amplifier of the former, which can bring greater exposure to the former and thus help corporate products obtain more sales leads. This is also the ultimate goal of Zhihu institutional accounts that we have repeatedly mentioned before.

Since Zhihu Q&A SEO is so important, how should it be done specifically? I have summarized 2 main tips:

1. Find high-powered accounts to like

Brushing likes is nothing new on Zhihu, but how to brush likes efficiently and unnoticeably requires some tips. But don’t be in a hurry. Before officially sharing the tips for getting likes on Zhihu, we have to first clarify one issue, that is, the sorting algorithm of Zhihu questions and answers, which is the "compass" for our subsequent operations.

There are two sets of rankings for Zhihu Q&A, the old one. The old one is simpler, based on "score = weighted approval - weighted opposition", but it brings two problems:

1. First: highly praised answers will dominate the list for a long time, and even new high-quality answers will hardly stand out;
2. Second: If a large number of malicious votes are cast against the answer, the answer score may even be negative, which means that the answer will be sunk to the bottom and it will be difficult to "turn around" again.

The emergence of the new version of the algorithm (Wilson Score) has solved the above problems to a certain extent, allowing new answers to have the opportunity to surpass highly-praised answers that were posted earlier - this creates room for us to implement SEO plans.

The above is the calculation formula for Wilson's score. It is very complicated and it would take a long article to explain it clearly. But I’m not going to talk about it here. Friends who are interested can go to Zhihu and search “How do you evaluate Zhihu’s answer ranking algorithm?” Many experts have already analyzed it from various dimensions.

Our focus this time is still on the impact of this new algorithm on our SEO. Let’s just say the observations:

1. The likes of high-level accounts in vertical fields have higher weights. For example, the like effect of a V5 account in the same field is stronger than that of 10 V3 accounts.
2. The effect of a like from a high-level account is immediate, and you can usually see the effect after refreshing the link after liking it.

In other words, our SEO task needs to be upgraded from the original 1.0 era of likes that focused on quantity to the 2.0 era of likes that focused on quality. So how do we do it specifically?

There are also two techniques:

1) Cultivate your own high-powered secondary account to like

This is not something that can be achieved overnight, but if it gets going, the mutual likes between institutional accounts and personal accounts can form a positive cycle, and the effect will be very significant.

It is worth noting that every like on Zhihu will appear in the account's dynamics. If we only like one account for a long time, it will be easy for users to find out and report to Zhihu officials. In serious cases, it will lead to the account being banned.

Therefore, likes need to simulate the normal user behavior trajectory. Do not like the same account continuously, and intersperse likes on some answers that will not directly compete with us in the rankings. Do not jump directly to the target answer after opening the question and answer link. Browse other answers under the same question as normally as possible, and sometimes make some simple comments.

2) Actively attract likes from high-powered accounts

It is not cost-effective to directly buy large-sized likes, and it is also easy to be reported. So how can we get high-powered accounts to voluntarily like our posts?

I have figured out a trick: quote some important opinions of high-authority V in the answer, and then @ the other party in the article. If the other party agrees with our content, there is a high probability that we can get a like from the other party.

Of course, the prerequisite is that our content must be interesting enough, which is the content we mentioned earlier.

For example, in the XMind case we mentioned earlier, it quoted the video content selected by @开眼科技 in its answer to the question “What is the Feynman technique?” and then @ed the other party.

2. Use third-party tools for quick sorting

What we talked about earlier was doing Q&A SEO on Zhihu, which is to improve the ranking of answers; but if we want to also improve the ranking of the Zhihu question with this answer in Baidu search, we need to use third-party tools for quick ranking.

Operations colleagues with a budget can try Traffic Treasure/Super Fast Ranking. After three or four weeks, they can usually reach the homepage of Baidu results.

3. Two Easter eggs that are not strictly SEO

1) Make use of the self-recommendation function

Zhihu institutional accounts can obtain a certain number of "self-recommendations" by completing tasks every week. The so-called "self-recommendation", to put it simply, is a function that allows the platform to help distribute content on its own.

Since the number of "self-recommendations" is limited, the best way is to combine background data to screen out recent potential content for self-recommendation, so that the excellent content has a greater chance of becoming popular.

2) Open Liu Kanshan invitation

Sometimes when you encounter questions and answers with low views, you can enable Liu Kanshan invitation and independently invite creators recommended by the system. The purpose is actually to let more users see the content you create.

5. Final Thoughts

Zhihu is a good traffic pool, but we must also realize that not all types of products are suitable for content marketing here. Perfect Diary came and left; Pagoda came and left; Miniso also came and left...

It’s not that these products are bad, nor that Zhihu is a weak platform, but that the “compatibility” between the products and the platform is too low, and they are not the “right people” for each other (for example, Perfect Diary is more compatible with Xiaohongshu).

The ToC tool products I mentioned earlier, such as Shimo Documents, XMind, and Chuangketie, are more compatible with Zhihu:

First, there is a high overlap in the user groups of Zhihu and tool products, as both are highly educated and pursue high efficiency; second, Zhihu users generally look for answers to specific questions, and if they see a suitable tool, they will generally buy it; finally, Zhihu supports placing product links directly in answers (links can be customized and user sources can be tracked later), which can greatly shorten the customer acquisition chain.

In summary, companies should decide which platform to operate on based on their own product attributes, user characteristics, and the tone of different self-media platforms, as well as what content format and operation methods different platforms should adopt. This is an important rule for enterprises to operate new media.

Author: One Punch Man's younger brother

Source: Nomadic Diary (ID: nomadic-diary)