The essence of data thinking is to digitize some business status quo, with logic for judging the status quo of business data, and logic for reaction based on the judgment results.
What's the meaning?
That is, from 1) data collection, to 2) data analysis, and 3) feedback after analysis can be quantified as much as possible.
What should we do when it can be applied to user operations?
Looking back on data thinking, we must have been able to understand that it is nothing more than three steps:
-
Standardize the collection of user data and clarify user labels
-
Analyze and compare user tags
-
Based on data changes of different user tags, form operational feedback
For example, many platforms are now doing community operations. How should we use data thinking when users in the community operate?
First of all, let’s think about the value of the community. The value of the community is actually requiring users to discuss topics within the scope of the community (or users’ concerns), conduct high-value discussions, and avoid low-value content from appearing in the community. Otherwise, the community will be blocked by most users, and it will be difficult to achieve the transformation that you want to do in the community.
Therefore, around the purpose of this community, to operate users in a more refined manner, we usually require users to modify their names in the group.
If it is a community based on certain community characteristics, users are usually required to mark the region, industry, and name.
If it is a group based on a certain purpose in a certain area, the user is usually required to mark the company, interest, name/or stage based on this purpose. If it is a reading group, then it needs to mark the book they are reading, and if it is an English speech Group, you need to mark the segment of the English speech, etc. ~~
We must be able to see which groups have not been marked by users. Unless they are acquaintances, it is difficult to form discussions.
But once there are user annotations, users will have a good foundation for unfamiliar communication and operations.
This is a good first step for the infrastructure of a community, but from the perspective of community operators, it is not enough to collect user tags for community operations.
What other information should be collected?
-
Pay attention to the frequency and quality of the user's speech. The quality represents how much continuous discussion this user can bring in the group after speaking.
-
Pay attention to the number of users that users pull into the group, and the links between users and users in the group
-
Mark the content discussed in the group, such as gossip, dry goods, a certain group member, offline activities, etc.
-
Follow the active time period of the community
Usually, there are several key links in a group, and these nodes link the members of the whole group.
The above completes the data collection for community users.
Then the quality of the user's speech frequency and the link data between users and users are used to grade the users.
We continue to analyze the changes in the frequency and quality of users' speech at different levels after different types of content are released, as well as the data of links between users and users to judge.
-
What kind of content is constantly improving the link density of users in the group
-
What kind of content increases the growth rate of users in the group
-
What kind of content increases the frequency of users speaking
-
What kind of content is to improve the quality of user speech
-
Which content works best for users at which ratings
-
What kind of content works best at what time period
In this way, we have mastered a big picture of community operation that is finer than user groups, finer than user behavior, finer than active time, and finer than content labels.
At this stage, we have completed the data analysis, but this is not enough, because users are always fluctuating, and behind the fluctuations are the fluctuations of social hotspots, and the growth and changes of users themselves, so we still need to conduct continuous data collection. And data analysis, in order to know the big picture of the current community operation, let us know the community we operate.
So how to do the feedback of the next operation action?
When we find that users of a certain level are decreasing, which makes the user structure of the community two-leveled, we can find the content that is most interesting to the users of this level that we need to increase, and publish more.
When we want to convert some commercial projects and need everyone to increase their attention to the group and strengthen the links between users, we can initiate some content discussions that will help increase links within the group, so that it can be carried out at the right time. business transformation.
Of course, with our data statistics and analysis, we can definitely know which users have more user links and better output of community content. We also need to have a single point of communication and discussion with these core users to facilitate their understanding of Our community has a higher degree of recognition, and when the operation of the community is ineffective, we can communicate with these core users to perceive their judgments.
Data is like money. Without data, it is absolutely impossible, but data alone cannot solve all problems.
Data is more like the infrastructure for us to do a good job in user operations. With data, we can have a basic judgment on the current situation of users, the market and historical behavior of users, but the insight into users cannot rely on data alone. , In many cases, we need to communicate directly with users to find out the reasons behind it that are difficult to gain insight from data.
In order to facilitate insight, we can try to label the labels that were not data-based in the past as much as possible by extracting key elements.
#Special guest author#
Jing Qiu, everyone is a specially invited author of the product manager, a gold medal tutor of Qidian Academy, a former operation director of Ali Southwest, a senior product manager of Baidu, and has served as assistant to the president of Baidu and operation of the mobile cloud business unit.
This article was originally published by everyone is a product manager, and may not be reproduced without permission.
The title map is from PEXELS, based on the CC0 protocol