Analyze the data operation methods of social products!

Data is an important component and evaluation criterion for product analysis. Through data analysis, you can clearly see the user activity trends and characteristics of the product, changes in user activity, etc. Therefore, mastering data operations is very important for product operations. By doing this, you can continuously optimize operational strategies.

Whether it is WeChat or a blog, there will be a corresponding timestamp when it is sent. Its purpose is to tell users what the time coordinates are when the information was sent and how long it is from now.

Time is the essence of social information. Without time indication, the information users receive will be inaccurate.

Take dating as an example. When the known information is unclear, if you do not open the received message as soon as possible, you will not know when the user wants to date you (time, place, and person are the three main elements in the message).

To sum up, time and social interaction are positively correlated.

If the goal is profit rather than dissemination, then most of the functions of social products must prioritize GMV, and this will depend on "time".

Most people think that the way to monetize social media is e-commerce. In fact, just like commercial activities, the longer the retention time of social activities, the stronger the consumer mentality. Social media itself can be monetized through time - especially in a closed social environment, sufficient retention time is a hotbed for consumption.

At the same time, social products have specific scenarios and triggering conditions, which will generate a lot of data; and these data are an important component and evaluation criteria of product analysis.

It can be said that data is both the direction of the product and a ruler to measure the depth of the product; it can objectively and comprehensively reflect the product life cycle and overall product operation, so data operation is an essential operation method for the product.

In this data analysis, the Umeng+ platform has some key data indicators. This time, X Circle can obtain more detailed data through Umeng+ mobile statistics (U-App AI version), which can assist in more complete product monitoring.

The positioning of this X circle is that in addition to stranger social interaction, there are also circle relationships. We want to build a community by operating circle relationships. Its essence is a social consumption model rather than social e-commerce. By increasing the length of time users stay in social products, consumption conditions can be met; for example, point systems and membership systems, rather than simply introducing B-side merchants.

Next, I will break down some basic indicators to reflect the role of the data (the data is anonymized and not authentic, please do not treat it as real data).

1. Obtain user activity map through basic indicators

(Schematic diagram)

1. New users (today)

There are two types of new user indicators: one is the number of new users, and the other is the change rate.

Different from ordinary data, the rate of change can objectively reflect the degree of change in the number of users.

Taking April 28 as an example, there were 17,040 new users, with a change rate of +12.206%. Compared with 15186 on April 27, there is a corresponding increase. The graph generated by comparing the 24-hour chart every hour can intuitively reflect that the newly added users are accustomed to opening the X circle at 13:00 and 19:00.

Through the data of these two days, we can draw a conclusion: 13:00 and 19:00 are active time periods, which are just after lunch and when commuting from get off work. This shows that the new users are mainly office workers.

This is of great help in defining the scope of user portraits in the next step, and it also makes us realize that within the same social circle, stranger social interaction is not an occasional event, but a high-frequency and rigid demand that occurs in a specific time period.

2. Number of launches (today)

There are two types of launch times indicators: one is the number of users who launch the system, and the other is the number of users who launch the system repeatedly.

Taking April 28 as an example, the number of launches was 231,074, an increase of 8.96% from 212,060 on April 27. Combined with the previous number of new users, we found that the number of repeated launches has increased, and a higher number of repeated launches indicates that user stickiness has increased.

Combined with the newly launched function of customizing watermarks for PO pictures, we can conclude that "customized watermarks are popular among users."

If we work backwards from the functions to determine the demand, and then add the continuous data from the previous two days in March, we can see that the launch of a certain function has a significant impact on the number of user activations.

3. Cumulative daily activity during the period (today)

The cumulative daily activity in a period is the deduplicated data, and the deduplicated data per hour can reflect the actual daily activity of users, which can basically reflect the number of users and the daily peak - generally speaking, the number of users reaches a peak at 20:00 and starts to decline at 22:00.

This reflects that self-presentation is a common need of users but not a core need.

Through these three indicators, we can accurately draw a user psychological map, such as the user's usage time, favorite functions, specific user portraits, etc.

2. The overall trend directly reflects user changes

(Schematic diagram)

1. New users (7-day average)

By summarizing the number of new users in a week (on average), we can get the average daily increase in users in the past week. Different from the previous data, this data can get a steady increase in weekly/daily numbers.

For example, the average daily number of new users from April 20 to April 27 was 4,125, an increase of 4.36% year-on-year.

(You can also add different times below to compare versions)

2. New user retention rate on the next day (7-day average)

From this data, we can see that from April 20th to April 27th, the average next-day retention rate of new users reached 30.32%, a year-on-year increase of 3.27%.

The user retention rate this week has increased compared to last week, indicating that the current product form remains in an upward period. Although the increase rate is not high, at least there is no decline in retention.

By analyzing the product life curve, we can see that the product is in the seed stage. In the cultivation of construction-type seed users, the next-day retention data can help screen out the specific number of new seed users.

I won’t go into details about the other data, but will talk about the overall crash rate.

3. Total crash rate

The total crash rate is a very good indicator because, unlike other indicators, the total crash rate can be directly connected to development.

As a product risk control system, the crash rate is a problem that is often overlooked or most easily overlooked by operations, and the total crash rate can reflect the stability of the product.

We always think that only development is related to this matter, but in fact it is also closely related to product operation, especially when the product fails, it will greatly affect user retention.

Except for seed users, the loyalty of new users is very low and they are easily affected by the total crash rate.

But we also saw that the total crash rate of version 0.1.2 was 0.1%, while the total crash rate of version 0.2.2 was only 0.08% - the control ability of the development team is improving.

Similar to the visualization of operational data, the TOP version of the pie chart below can reflect the proportion of various user indicators in different versions.

For example, the percentage of new users for different versions can be represented by a donut chart - version 0.1.2 accounts for 21%, and version 0.2.2 accounts for 63%.

Through this process, we know that the product has become more competitive after version iteration, and the product operation direction has been further clarified.

3. Ranking promotion weight

(Schematic diagram)

1. Industry data chart

In the industry data section, Umeng+ Mobile Statistics (U-App AI version) can also provide rankings, and this ranking is of great significance to the X circle.

In the past, we could only estimate the industry ranking by the number of downloads of competing products; but now with the ranking, we can monitor/compare our position among products in the same category, as well as the position of the category in the entire market in real time.

(Schematic diagram)

2. User activity

Based on the visualization of user activity data, we can know the number of active users and the number of active days, which is of great help to us in accurately classifying active users.

(We generally classify users according to a 15-day cycle, with 5-day active, 10-day active, and the highest level being 15-day active. There are three user groups, A, B, and C, divided by user activity levels. They are generally maintained by dedicated personnel and reactivated in the community.)

4. Some thoughts on data operations

As a data operations veteran, what is the most important thing?

Actually it means: the ability to integrate data.

The data visualization provided by Umeng+ greatly facilitates data statistics and aggregation. Visualization makes user trends very obvious. It can also monitor the status of new versions in real time and track and feedback data changes in a timely manner.

This was unimaginable before, because in the past, symptoms were determined through data walkthroughs or spot checks, which would waste a lot of valuable time. The key to data operation is speed, which is also the benefit of data visualization provided by Umeng+.

Related reading:

1. Data operation: How can operations train their data thinking?

2. Data operation: 8 essential data analysis methods for operation!

3. Data operation: How to use data analysis to achieve user growth?

4. Data operation: How to build a data indicator system?

5. Data operation: How to analyze data more efficiently and more valuablely?

6. Data Operation: How can big data make users more willing to pay?

Author: Wang Shiqin Source: Wang Shiqin