A brief history of Zhihu 2: How to iterate products in a community with more than 60 million users

From products like WeChat and Zhihu, I feel that the most important thing in making products is rhythm. For product managers , making the right things happen one after another may be the most difficult thing.

Socializing does not equal community.

Tencent excels in social tools , such as QQ and WeChat. QQ Space is more like an independent space rather than a unified community. So when it comes to community product learning goals, I chose Zhihu, which has 60 million+ registered users and has raised D round of financing.

I saw that the founder of Zhihu highly recommended a book called “The Death and Life of Great American Cities”, which talks about how to plan and build a vibrant city. I happened to have read this book, and I remember that the book said there are three key points for a city to be vibrant:

1. Do not plan the city strictly according to functional areas. The areas where office and living functional areas intersect are the most popular. There is no one in the CBD at night, and no one in the suburban residential areas during the day.
2. Keep streets as short as possible to create intersections between blocks
3. Preserve the old city, where lower rents can accommodate low-income people who provide the basic labor services necessary for urban vitality. For example, breakfast shops, waiters, cleaners, and security guards.

In fact, the community is very suitable for the metaphor of the city. Douban Community has previously tried a product form called Alpha City, but it ended in failure. This indirectly proves that building a virtual city is not easy and requires the coordination of various forces to make it start to grow.

Research Methodology:

• Use Python crawler on Tencent Cloud to crawl all answers under the user "Zhihu Products" + "Zhihu Butler"
• Use Python crawler on Tencent Cloud to crawl all the essential answers under the topic "Zhihu Product Improvement"
• Interviews with around 30 Zhihu users with over 10,000 followers (mainly those we met during the Zhihu Salt Club)
• Interviews with early and current employees of Zhihu.

1. What are the needs of users of knowledge communities?

• Creative users are estimated to account for 5% . Have relevant experience in the problem, have a certain amount of creative time, and can produce high-quality content.
• Social users are estimated to account for 15%. They are good at expressing their views and opinions sporadically, but are unable to form systematic output, such as this group of particularly talented netizens in the NetEase comment section.
• Browsing users are estimated to account for 80%. Most netizens belong to this type. They are not good at writing and do not have a strong desire to make comments and other interactions, but they are willing to browse and collect some high-quality content. The browsing and approval of this group of people will be an important source of sense of accomplishment for the top 5% of creators.

According to the frequency of all user needs (high -> low)

1. Purposeful search + aimless browsing (discovery)
2. Interactive functions for expressing opinions such as agree + disagree + comment
3. social contact
4. Output knowledge

A friend of mine who is a big V with 10w+ fans often uses: ask questions, answer, follow, like, collect, private message, not helpful.

2. Three categories of product functions on Zhihu homepage

Combined with the following screenshot of Zhihu's homepage, we can see the following three categories of product functions from Zhihu's homepage:

• Content (content generation + content distribution + content organization, such as topics, discovery)
• Social (social behavior + reputation system, such as following, liking information reminders + personal homepage)
• Cash (spending stream + income stream, such as my balance, my gift certificates)

2.1 Zhihu’s product mechanism in guiding PGC

• Content generation: questions, answers, columns, Live, topic index, public editing
• Information distribution: editor recommendations, big V endorsements, favorites, internal links between answers, Zhihu Daily, Zhihu Weekly, topics-essential answers
• Author incentives: likes (self-contained information distribution), fan attention, editor recommendation, excellent answerer identification (higher question ranking weight in the field), Zhihu Salt Club (offline)
• Commercialization: Zhihu e-books, rewards, Zhihu Live

We can also see directly from the areas of different modules in the web page that the content distribution function occupies the largest area. Interestingly, the area of ​​the content generation function is not the largest, but the location is the best, which is in the golden triangle area where users browse the web page.

2.2 Social (social behaviors, such as following and liking information reminders)

We can see the priority differences in the Tab page here:

• The highest is the content-related message reminder
• The second is the people who follow you
• Who gave you a thumbs up or thanks at the end?

From the information priority ranking in this regard, we can see that Zhihu prioritizes content function, followed by social interaction.

1.3 Social (reputation system, such as personal homepage, professional information, education information, number of followers)

3. In the past 7 years of Zhihu, what is the most important functional iteration every year?

So the question is, for the PGC community product Zhihu, what was the most important product feature iteration every year from 2010 to 2017?

I tried to use Python crawler on Tencent Cloud to crawl all the essential answers under the topic "Zhihu Product Improvement" and tried to answer this question with data. Of course, part of it is also my personal subjective judgment.

From data analysis, we can see the time rhythm of several major version iterations of Zhihu (draw a timeline), as well as the degree of user attention (the number of likes for related iteration discussions)

• 2010: Invitation to register, the threshold ensures that high-quality Internet entrepreneurs and programmers can answer questions seriously.
• 2011: New private message + discovery + collection functions were added to strengthen communication between Zhihu users and the acquisition of new information.
• 2012: The total number of likes and views of new users + the addition of featured answers under the topic page strengthened the social attributes and helped users filter out high-quality content.
• 2013: Optimize personal information flow , users’ topics enter the feed, dynamic is not time but algorithm sorting
• 2014: Question search algorithm and question ranking algorithm were released to optimize the quality of community content
• 2015: Anti-cheating system + copyright + secondary search optimization
• 2016: Column + Live was launched, and a more convenient first-level creation portal "Ask questions, answer, write articles" was launched

Zhihu Manager + Column articles from the Zhihu product team

An interesting phenomenon is that there are only 11 articles from Zhihu product team that have received over 1,000 likes, while there are 19 articles from Zhihu Butler that have received over 1,000 likes. It can be seen that the average Zhihu user is more concerned about community operations and atmosphere rather than product iteration.

But I personally think that if there is no product-level protection for high-quality information and crackdown on low-quality information, but the operation team relies on the operation team, the manpower cost required to maintain a high-quality discussion atmosphere may be far beyond what a startup company like Zhihu can afford.

4. The algorithms behind Zhihu’s product features

As for Zhihu's product iteration, the early stage focused on information distribution and interaction, and algorithm-level functions began to appear in the middle stage. Users may not be able to see them but they are very important to the user experience.

Content side

How does Zhihu sort answers? https://www.zhihu.com/question/19576738

Wukong system - anti-spam, likes identification https://zhuanlan.zhihu.com/p/19998740

Zhihu + Sogou search function cooperation on the creation of Zhihu search box – Zhihu column

User side

How does Zhihu calculate the weight of a user in a certain field? https://www.zhihu.com/question/21544084

Friendliness

What lessons should Zhihu operators learn from Mr. Ge Jin’s withdrawal? https://www.zhihu.com/question/24843376

5. Those unknown niche functions of Zhihu - Programmer Easter Eggs

Zhihu shortcut key: https://www.zhihu.com/question/19842222

In Chrome, press shift+ctrl+J

Zhihu once had a group function, but the content was of poor quality, so it was quickly removed.

summary:

A very important part of Zhihu's core functions is about the Feeds flow. Its value lies in allowing users to see content that interests them while also allowing them to see some of the most popular content on the platform.

At the product level, it can be considered that Zhihu's content distribution is mainly based on social relationships, and the Feeds in the "Discover" function are based on algorithms and editor recommendations.

Through a series of product functions, Zhihu has established a complete relationship chain between people-problems, problems-people, people-people, and problems-problems. In my opinion, this is also an important reason why Zhihu can become a content city.

• Questions and people, question-answer, invitation to answer.
• People and problems, positive: agree, thank, share, reward, negative: collapse, report, not helpful.
• People, follow, block
• Questions and issues, links, topics, indexes, redirects (for duplicate questions)

There was a very popular question on Zhihu one year ago:

"I started browsing Zhihu at the age of eight. How much knowledge can I have by the time I am 20?"

An answer with more than 10,000 likes mentioned a sentence:

"Your mom and I are going to square dance, what should we do if we can't keep up with the rhythm?

Son: With the low level of effort most people put in, it’s not even a question of talent.”

In fact, in addition to enjoying such witty answers, more and more users are now flocking to Zhihu, a Chinese question-and-answer platform.

In my opinion, the most important reason is just one sentence: Zhihu has a huge amount of high-quality content in different fields, and has a high-quality organizational structure, which makes it easy to find.

The subsequent Zhihu History 3 will consider analyzing from the perspectives of community operations ( attracting new users + promoting activation + atmosphere building + why some big Vs quit Zhihu, etc.).

Mobile application product promotion service: APP promotion service Qinggua Media advertising

The author of this article @贺嘉 is compiled and published by (Qinggua Media). Please indicate the author information and source when reprinting! Site Map