Discussion on SNS attention mechanism: Network Atlas "Knowing"

Introduction: The original intention of writing this article is to let oneself "read" know. Knowing that it is an attractive product that has even shifted my interest in Weibo, I want to know what the logic behind it is and how it works. Of course, it is possible to talk about Quora, but I have not used Quora so I have no say, so I can only know the analysis.

One theory: Knowing is a kind of mixed network, which is composed of two kinds of nodes of individual node and knowledge node.

This network structure is due to a bold and interesting setting: users can either follow "people" or follow "issues" and "topics." The following illustration shows this structure:

The personal node (or social node) is the user active in the network, and the knowledge node includes two kinds of problem nodes and topic nodes, and a problem node includes both a problem and multiple answers under the problem. Social nodes can be http://www.aliyun.com/zixun/aggregation/8444.html with questions, answers, and topics at the same time "> information interaction, as shown below:

In this hybrid network, there is a rich and diverse information interaction between individual nodes and knowledge nodes, knowledge nodes and individual nodes, and the following diagram summarizes the main information transfer activities in the network:

One big difference with Wikipedia, Baidu, and other types of products is that information exchange between nodes is two-way and continuous, making it more frequent and contagious in a networked system.

Second theory: benefit from network mechanism, knowledge node can grow automatically

This is a kind of life mechanism: each knowledge node can continuously absorb the knowledge nourishment from the network, simultaneously can organize and optimize these nutrients. "Absorption of nutrients" is because every knowledge node can be linked to the topic and people and other intermediaries to let more people see, attract them to provide more quality answers; "Organization and optimization" refers to the voting mechanism for answers, highlighting the best answers in a self-organizing way, while the "no help" function can eliminate bad currency.

On the functional level, the individual node is equivalent to a special knowledge node, which goes further, is a "pseudo" social node.

The user's behavior is almost entirely knowledge-related, and the user's right to do other social activities is greatly limited. In a mixed network, the role of a human node in addition to the contribution of knowledge, is to cause attention to his users and more likely to be interested in the issue of association, so that the number of links in the network to increase rapidly, this role is similar to the "topic", so you can consider personal nodes as a special knowledge node. So I drew the following diagram, and abstracted the user into a knowledge structure that was placed throughout the knowledge hierarchy:

Not only that, with the continuous deepening of the use, I increasingly feel that the personal node is essentially a "pseudo" social node, and know that it is actually a "pseudo" social network.
My basis has the following two aspects:

1 is also the object of concern, "people", "Problem", "topic" These three in the system is not equal weight, "problem" is the most important, and "people" and "topic" is the second.

What is this remark? We look at the structure of the latest dynamic information, the title of each piece of information is very eye-catching "problem", although this message may be caused by the behavior of a user concerned, but its presentation form is problem-oriented. This means that the "problem" is the most important and essential node in the knowledge, not the "person".
More direct evidence is expressed in the distinction between "notice" and "dynamic". Notification is a form of strong push, and dynamic is a form of weak push, weak points imply that the product designers on the weight of information settings. I just used to know when the most curious is what information will appear in the notification, what information will appear in the dynamic, always find no behind the rules. It took a while to realize that the difference is (in general):

Information in notifications: all new developments related to the "issues" I am concerned with

Information in dynamic: All new developments related to the "Issues" user "topic" I am concerned about

For example, I pay attention to "Kai-Fu Lee", then Kai-fu Lee answered what questions, the system will not notify me; but if I focus on the question "Why is Lee so influential?" Then all answers to this question will be communicated to me. It is clear from here that the "problem" in the system weight is much greater than "people."
2 social incentives and ways of interacting have been greatly limited by the knowledge, and a few simple points:
1 can't upload pictures
2 The picture shown in the answer is very small
3 cannot comment on the responses
4 User profile is extremely simple, even the gender does not have to fill
5) ...
These constraints: Know not to fence, it is not just want to do the knowledge network and do social network, it is very clear what they want to do, is to do a great job knowledge network. The social network's position in the whole known system is from the attribute, is the means not the goal.

Summary: Do you know what you want to do? What other weaknesses?

The ambition is to establish a vast, borderless, rapid response self-organizing knowledge Network, here "Self-organization" contains two aspects: 1 nodes and links automatically, fast and reasonable increase; 2 node content automatically, quickly and reasonably rich. The dominant node of the network is "problem", the subordinate node is "people" and "topic", both of which are to expand the number of problems and enrich the content of the problem node. Therefore, although "people" and "topic" are secondary nodes, but they are the entire network system of the two power engines, know whether the future success depends largely on whether or not to establish a sound topic system and a complete expert system.

What else needs to be improved? Here I do not want to analyze from the perspective of the interface, this aspect of the small improvement must have a lot, such as many users complained about the search box problem, but want to still from the network mechanism of this deep perspective. I think the biggest mechanism that we know about the design flaw is: The existing mechanism is more advantageous to the old problem nodes than to the newly created problem nodes, so that these new nodes can not quickly enter the knowledge network system and establish enough connection to enjoy the knowledge nourishment from all over the network. To put it more bluntly, The newly raised question cannot be answered quickly. Why do you say that? As has been analyzed earlier, "notification" information is related to issues of concern, which is naturally an old problem, and if someone raises a new question at this time, it is not possible for other users to get it in the notification message, only in "dynamic".

Here's the problem: if you've got a lot of attention, then each login to see the new notice there are many, strong push of the notification information will be weak push the dynamic information drown out, to deal with the notice is enough, you have no time to find those new problems in the dynamic, you are forced to focus on the information is always those old problems. Imagine, if you know how to change your strategy, the weight of "people" and "topic" is set to be the same as "problem", so once the topic or the attention of the person added a new problem, users can be notified, then the new problem is quickly answered the chance of a big increase. Of course, "notice" is too bloated and huge, so a compromise improvement is: still retain the "people" and "topic" of the low weight, but leave a gap: all the topic or attention of the new added to the issue of "notice", the rest of the "people" and "topic" Dynamic information is still classified as "dynamic." In this case, the new problem node and the old problem node are placed on the same priority level, on the one hand, the new problem node due to the attention of people or the topic of the "referral" by more users to find a quick solution, on the other hand, the old problem of the content of every abundance can be concerned about its users benefit, So that the knowledge network is more reasonable and efficient.