Community operation: data analysis model of community activity

The main point we will talk about today is relatively detailed, which is how to conduct data analysis and optimization in group topic discussions in community operations . It is the implementation level. You can also directly refine it, use it for your own benefit, and extend it to more areas.

1. Primary data analysis model of community activity

As we all know, community operation and product operation are the same, and they also have a life cycle. If the community management is in place but it is not promoted at the right time, the community will quickly die.

So, in community operations, group topic discussion is a good way to promote activity.

• Not only is the cost low;
• Moreover, the output content is sorted out again and made into mind maps, articles and other content for reuse and output (I have compiled an article on decentralized content output before, but I was too lazy before. If I get excited later, I will share it with you);
• It can also maintain user activity.

It is an active method with a very high ROI, but it is not something that can be done simply. Many community operators have only reached the level of doing it, without a systematic SOP planning (this involves a systematic SOP cycle planning for community operations, which will not be elaborated here. I will share it later if I get excited, hahaha) and data analysis.

Because the focus of this article is on how to conduct data analysis on group topic discussions as a means of community activity, the specific implementation process of topic discussions will not be elaborated. If you have experience in community operations, then you should also know the topics that need to be prepared for discussion.

There are two main classification indicators for topic discussions in community operations: scale and difficulty. Today, we will put aside the variable of scale and mainly analyze it from the perspective of difficulty. The community operation discussions below are collectively referred to as "group topic discussions".

Figure 2

Group topic discussions are graded according to the difficulty of the content and the degree of participation of the participating users, but grading alone is not enough, because we also need to conduct data analysis, so we need to quantify different situations.

If you can't quantify it, you can't visualize the data (data visualization is the focus of data analysis, especially in the face of huge amounts of data).

Figure 3

If you don’t have a very complete community coaching tool to provide you with some quantitative data, you can assign values ​​to different grading situations.

Figure 4

Finally, the content of the topic discussion is collected and graded and summarized into a relatively simple data table. If the amount of data is small, you can find patterns by analyzing it with the naked eye, but in actual work, it is often necessary to do more than 6 periods.

Figure 5

If the amount of data reaches this level or above, can you still analyze it with the naked eye to find the patterns? What if there is more than one type of group? Are there multiple types of communities that conduct different types of topic discussions?

So at this time we need to use data visualization to assist in judgment, but it is obviously not possible.

Figure 6

Generally, in this case, I will use a scatter plot for visual description (the following is a hypothetical diagram, just for your understanding). Generally, there are five situations when drawing a scatter plot based on your data: strong positive correlation, weak positive correlation, strong negative correlation, weak negative correlation and no correlation.

Figure 7

Basically, there are five situations as follows. I will not elaborate on the strength of each situation. The main issue is the correlation, based on the above model:

• Positive correlation: the more difficult the topic, the higher the participation of community users;
• Negative correlation: The more difficult the topic, the lower the participation of community users;
• No correlation: There is no correlation between difficulty and engagement.

If you have completed the above basic operations and data analysis, you have already mastered the basics.

Figure 8

If difficulty has nothing to do with participation, then what else affects the participation of community users? This also has to do with the time it takes to create the group and the intensity of operations.

2. Advanced version: Data analysis model variable replacement and cross analysis

[Data analysis model variable replacement] In addition to the difficulty level affecting participation, the length of time the community is established and the intensity of operational investment also have a great impact on participation.

Suppose, if you find that there is no correlation between difficulty and participation, then at this time, you can also replace one of the variables in the scatter plot. Find out the core factors that influence community user engagement. There are generally several aspects of operational investment:

• Capital cost: grab red envelopes, gifts, etc.;
• Human cost: how much manpower is invested, cross-team collaboration and communication costs, etc.
• Time cost: the time invested from preparation to execution, etc.;

Similarly, if detailed data quantification is not possible, each dimension can be graded and assigned values ​​based on the actual considerations of the operators to obtain quantitative data.

Fig. 9

[Cross-Analysis] If there is a correlation between difficulty and participation, and you want to further analyze more potential factors that affect participation, you can perform data analysis models such as cross-analysis on the scatter plot.

As shown in the scatter plot below, in addition to observing the difficulty and level of participation, you can also observe the distribution of operational intensity through different color levels.

Fig.10

Such data visualization graphics can help us quickly find cases with high ROI, such as points with high participation and low operational intensity. We should quickly extract and analyze such key cases and replicate them.

Fig.11

Advanced version: feedback execution and adjustment

The ultimate goal of data analysis is to assist us in reviewing and summarizing, as well as finding specific points, and ultimately coming up with solutions or replicable excellent solutions, which can be fed back into the overall operation system to improve overall operational efficiency.

• Find core data and key indicators;
• Conduct correlation analysis on key data;
• Come up with solutions or excellent cases and provide feedback to the execution system.

Fig.12

[The final super important point] Data analysis is an operational link that requires heavy investment. It does not mean that data analysis is required for all aspects. If we analyze every link indiscriminately under limited manpower:

• It will only significantly reduce the efficiency of operations;
• There is also the vicious circle of falling into analysis paralysis (doubting life).

Therefore, knowing which core links should be analyzed for data and invested in appropriately is a core underlying capability of the project leader.

Author: Operation Meow who loves drinking Coke

Source: Operation Meow who loves drinking Coke