Operational Practice | How to use WeChat background data to optimize WeChat operations!

This article will start from the positioning and content initialization of the operation of this official account, and then transition to the data optimization part, focusing on how to optimize WeChat operations based on WeChat background data analysis . I hope it can be of some inspiration to friends who operate WeChat official accounts , and that will be enough for me!

Not long ago, I opened a personal public account, hoping to practice new media operations in my spare time. During this period, I also tried to use the data statistics function provided by the WeChat backend to optimize the entire operation. I have been doing this for nearly 20 days. Due to limited spare time and precise user segmentation, although the number of fans is not large, I have gained a lot of experience, which I would like to share with you now.

After some practice. I believe that the significance of WeChat backend data statistics lies in the following three aspects:

• Conduct verification analysis on the initial public account positioning and user accuracy to see if the initial strategy is on track and if adjustments are needed in a timely manner
• Analyze user sources/channels to determine the source path of fans so as to adjust the publishing channels
• Analyze images and texts to optimize content and find the right time to push them

Positioning of the official account

This public account I created is about operations. It was opened with the purpose of sharing my own operation experience and making friends with various operation experts. Since I have only been engaged in operations for a short time and my skills are limited, I positioned the audience of this public account as:

• College students interested in operations
• Employees who have worked for a period of time and want to switch to operations positions
• 1~3 years of operation experience

From the above positioning, it can be seen that the target audience is limited to novice operators or those who are about to engage in operations. Therefore, the number of people is limited, and fans focus on quality rather than quantity. After conducting online research and surveying some of the above-mentioned groups, we started from three aspects: the goals that the audience needs to achieve , the existing pain points , and the benefits that can be obtained by becoming a member of this public account, thus outlining the user portrait of this public account:

Overview/profile of the target audience of the official account

Next, based on the above user portraits, combined with the content of my work and the areas I am involved in, the content sections of the official account are divided into the following five sections. The future content push will focus on these aspects:

• Operational tips : share some practical operational means or methods
• Operational philosophy: Share important ways of thinking and concepts in operations
• Career planning: Share your personal views and experiences on career planning
• Operational cases: share practical operational cases
• Learning methods: share how to learn quickly at work, learning skills and tools

After completing the above public account positioning and content section, I named the public account "XXX" (to avoid suspicion of advertising, the public account name is not displayed), and the keywords highlight "operation" and "for newbies".

Initialization of the official account and channel delivery

After the positioning and framework of this WeChat account are established, the next thing to do is to initialize the content of the official account, that is, to arrange the "hall" to receive the audience in advance before they follow and enter the official account, so as to leave them with a good first impression and prepare for subsequent retention and push.

The content initialization of this public account focuses on three aspects, namely, the WeChat menu bar, automatic reply words and guiding pictures and texts to follow. Their construction must be subject to the tone of this public account and can represent the values ​​of this public account and the content to be pushed in the future.

Initialization of the content of this public account

WeChat menu bar

In the early stage, when there is not much content, the menu setting is as simple , intuitive and easy to understand as possible. Therefore, the menu setting of this public account is as follows:

Old news: If you have not obtained an invitation for the original function, click to view historical news so that subsequent fans can see previous articles;

Useful Information : Place some longer, classic useful information articles in the material library and link them here so that fans can directly view them after clicking on them.

Hook up: My WeChat account will pop up directly, allowing fans to contact me and interact with me, which not only increases fan activity, but also provides me with materials and directions for future content updates.

Menu settings for this public account

Automatic reply

The automatic reply settings are mainly divided into two parts:

(1) Automatic reply added

The settings of the added automatic reply are as consistent as possible with the previously set tone, targeting newcomers/peers (ok, I'm revealing my age), with a friendly and interesting language style, reflecting the following aspects:

• The target audience of this public account is newcomers. Let's go from 0 to 1 together.
• In the early stage of operation, there is not much content. Please pay attention to it later to dispel the doubts of followers.
• Instant interaction, you can hook up with me at any time

Automatic reply settings added to this public account

(2) Automatic reply with keywords

The purpose of setting up automatic keyword replies is threefold:

• Strengthen interaction with fans
• “Burying points” to check the popularity of articles (different keywords will be set for each article)
• Some articles are presented in the form of pictures, which are not clear enough in WeChat. Reply with the corresponding keywords to get a clear PDF document.

Guide attention to pictures and texts

Guide the readers to pay attention to the pictures and texts one after the other, so as to attract the readers' attention to the greatest extent, so it is divided into two parts:

(1) Guide the readers to pay attention to the pictures and texts before the article:

The positioning and content sections of the official account are shown in the figure below:

Preface image and text of this public account

(2) At the end of the article, please follow the QR code:

The purpose is to use the second level to attract those who have not been followed before but have finished reading the article, to retain readers to the greatest extent possible and convert them into fans. The content mainly reflects the name of the official account and a large QR code to guide people's attention.

At the end of this public account, please follow the QR code

Use backend data to optimize public account operations

The WeChat backend data sections involved in this article are user analysis, image and text analysis, menu analysis, and message analysis. Interface analysis and web page analysis are not involved because this official account has not undergone secondary development.

Several major sections involved in backend data analysis

Combined with the three topics mentioned at the beginning of this article that require background data analysis for optimization, there is a certain correspondence between these data analysis modules, as shown in the following figure:

Several major analysis sections corresponding to the discussion topics

User accuracy analysis/verification to check whether the previous operation strategy is correct

(1) Gender and language distribution

Gender and language distribution of fans

In terms of gender distribution, there are slightly more boys than girls, which is basically equal; the language distribution is mainly simplified Chinese, which is in line with common cognition. This part has little reference value for analyzing Topic 1.

(2) Geographical distribution of users

Geographical distribution of fans

From the above two figures, we can see that users are concentrated in the southeast coast and Beijing. If broken down by city, they are Beijing, Shanghai, Guangzhou, Shenzhen and Hangzhou, which is in line with the distribution of the Internet. The top few are almost consistent with the user distribution of "Internet" in the Baidu Index. At the same time, fans all actively follow and are strangers. This shows that the fans of this official account are consistent with the original audience positioning.

Baidu Index on the distribution of people using the keyword "Internet"

(3) User source and channel analysis to determine the source path of fans

Source of fans' attention

The source analysis of users, as shown in the red dotted box in the above figure, is mainly divided into public account search, scanning QR code, menu in the upper right corner of the picture and text page, public account name in the picture and text page, business card sharing, following after payment, and others. The user behaviors they indirectly reflect are shown in the following figure:

Analysis of fan sources and corresponding behaviors

The sources of fans of this public account are sorted out, and the data shown in the following table is obtained. In order to see it intuitively, the size of the value is displayed in the form of a heat map (the gradient from warm to cool to warm tones indicates that the value changes from small to large):

The source of fans of this public account is broken down

As can be seen from the above picture, the vast majority of fans of this public account come from the "public account name on the picture and text page", followed by those who follow by scanning the QR code (added after reading the article). In this case, combined with the "Analysis of Fans' Sources and Corresponding Behaviors", there are several explanations for fans' behavior of following mainly based on the "picture and text page public account name":

• The name of this public account is attractive and attracts fans to follow it
• Fans are prompted by a banner directing them to follow
• The article quality is high, and fans pay attention to it immediately

In addition, fans who come in through the path of "scanning the QR code" means that they follow after reading the article. This type of fans is more picky than those who come in through the "public account name on the picture and text page". It also reflects that this type of fans is more "loyal" and accurate, and of high quality.

On 7-11 and 7-12, it can be clearly seen that the number of fans who followed the official account through "public account search" increased. This is because the articles in the official account were reposted on other platforms during these two days, and the name of the official account and WeChat ID were noted at the end of the article. The fans found us through these two clues.

The sudden increase in followers gained through business card sharing on 6-27 and 7-3 may be related to the high popularity of the article on the previous day. This point will be expanded in the following section.

From the above analysis, it can be seen that the pre-text guidance image and WeChat account name settings for the content initialization of this public account are relatively reasonable, and there is not much room for adjustment.

Content analysis to optimize the structure and content of the official account

Content optimization judgment

Activity

In the message analysis, you can check the keyword replies for each day. I have a "point" at the end of each article and set the corresponding keywords. The number of keyword replies can reflect the reading and popularity of the article on that day, and can also be used as an indicator to determine the activity of the official account.

Message reply quantity distribution

Number of replies for each keyword (partial)

Diagnose problems and discover patterns

According to Graphic Analysis -> All Graphics, find the "Export to Excel" item under "Trend Chart", set the date range, and export the data:

All graphic and text data export location

In the exported graphic data, whether it is "picture and text page reading", "open from public account conversation", "open from Moments", "share and forward" or "WeChat collection", there are two categories: "number of people" and "number of times". In order to reduce duplication of work and streamline analysis projects , I only retained the part about " number of people " in each analysis indicator. After sorting and integrating the fan data and content data, I got the corresponding heat table of "Date & Fans & Reading & Content". I named this table "Published Content Information Diagnosis Table".

Release content information diagnostic table

The above table includes date information, fan information, reading information, sharing and forwarding information, conversion rate information and content information. Considering the dissemination channels of this article - pushing it within the official account and sending it to the operation-related WeChat group , I did not forward it to the circle of friends for the first time, so:

One-time conversion rate = number of people who open the public account / cumulative number of followers

Secondary conversion rate = number of people who opened the app from Moments / total number of followers

It should be noted that the colors in this table change from light to dark and from warm to cold, representing the corresponding values ​​changing from small to large. Therefore, you do not need to see the size of each value clearly, and you can intuitively see the changes in the data as a whole. The purpose of integrating the above information into a table is to identify the quality of the content released these days and the regularity of the release time. In short, this table can play two roles: diagnosing problems and discovering patterns .

Diagnosing the problem

Outlier: On June 27, the number of people who read the picture and text page reached 1,784, but the total number of fans was very small. Combined with the tweets from the previous day, the secondary conversion rate was relatively high, reaching 65.63%. Therefore, this outlier is worth noting, and then call up more detailed data:

Detailed data for outliers

From this, we can see that "other" in reading sources accounts for the largest proportion, which is 79.34%. This well explains the phenomenon that the number of fans is extremely small, but the reading volume is huge. According to WeChat's settings, sharing outside of WeChat software is considered "other", such as on QQ, Maimai, Douban and other channels.

Next, let’s analyze the conversion rate of this article:

Conversion rate of outliers

It can be seen that at this time, due to the small fan base, most of the reading volume comes from secondary dissemination.

Discovering patterns

Discover the right content direction

It can be seen from the above table that during the period from July 3 to July 7, data such as fans, reading volume, and sharing conversions all showed an upward trend, which was quite impressive. When the release time and channel are fixed, it can be determined that the changes in content have brought about the synergistic growth of these data. Because the releases of the past few days are:

Content release status of this public account from 7-3 to 7-7

As can be seen from the above table, users are interested in operating cases and operating strategies, and have a high demand for systematic and easy-to-understand operating knowledge. In addition, users prefer pictorial presentation, which makes the reading experience more intuitive and friendly. The above experience in content selection and format can serve as a direction for searching, writing and arranging materials for articles of this public account in the future.

Sort out a reasonable menu structure

The data provided by the menu analysis section can help us discover high-quality menu columns (menus with more clicks) and unpopular menu columns (fewer clicks), so that we can improve the menu and optimize and update the "buried" structure and content when the content was initialized in order to improve retention rate .

Menu click trend chart

The trend chart of the menu section can show the click status of menus at all levels in the past 30 days. The TOP5, i.e. the top five menus in terms of cumulative click count, can be displayed in the trend chart section. As can be seen from the above figure, the first-level menu "Old Article Review" is clicked the most times, and it appears on the 7th to 14th. The day before and the day of this day are the start days of external communication, that is, the articles of the official account are published on external websites, and the ID and name of the official account are provided. Interested parties can add it through search. This also indirectly reflects that the user activity on that day is good and the external communication channel is beginning to show results. Similarly, “勾达小喵”, that is, contacting the author, also saw a steady increase after starting external communication.

Detailed data on menu clicks for each version

The above picture reflects the detailed click data of each version of the menu bar. Click "Download Table" in the upper right corner to download more detailed click data. After downloading and processing, the following table is obtained:

Heat map of menu clicks for each version

From the thermal situation in the above table (the depth of color and warm and cold tones represent values ​​from low to high), we can see the changes in the number of clicks in the menu version update dimension of the time dimension. Consistent with the situation presented in the above table, the number of clicks increased after 7-1, which is directly proportional to the growth in the number of fans and the growth in the number of readings.

Then use the pivot table to sort out the data, unify the menu settings and menu levels of each version, and count the cumulative clicks of each menu, as shown in the following table:

Heat map of cumulative clicks for each version of the menu

From this we can see that the two menus "Review of Old Articles" and "Contact the Author" (Hook up with Xiao Miao and Contact Xiao Miao) have the most clicks. Therefore, it can be judged that the settings of these two menus are reasonable. Other menus with fewer clicks need to be optimized, such as the micro mall and BD development. It is time to update them.

Optimize the release time period

In All Graphics -> Hourly Report, export the hourly report data from 6-24 to 7-13 to get the following table:

The original data of this public account 6-24~7-13 hourly report

Such data is disorganized and contains no valuable information, so the above table needs to be cleaned and reorganized. Same as above, only keep the indicators/data related to "number of people", use the function to convert "hours" into "time periods", such as "5:00~6:00", in the form of a pivot table to filter the dates. There are two points worth noting when processing data:

• Take the mean value of each transformed field;
• Remove the date outliers mentioned above to avoid interference with regularity identification.

Finally, we get the following two tables:

Processed hourly newspaper reading volume distribution table

Processed hourly newspaper reading volume distribution map

Processed weekly reading volume distribution table

From the above hourly newspaper reading volume distribution table and line graph, it can be seen that the peak reading period is between 18:00 and 23:00. Therefore, publishing articles within 15 minutes of the beginning of this time period can effectively increase the number of readers.

From the weekly reading volume distribution table, we can see that Thursday is the day with the largest reading volume in the week, because high-quality articles can be arranged on this day in order to attract more target readers to read.

Conclusion

In the ever-changing Internet era, our operational strategies and methods will inevitably encounter frequent trial and error and repeated iterations. The basis for us to adjust our strategies and methods is often external market demand and internal data about products. External market demand is elusive and difficult to obtain, but internal data is available, observable and easy to analyze. As WeChat operators, we must make full use of the data analysis tools in the WeChat backend, conduct in-depth research on existing data, and continuously optimize topic selection, layout design, and push timing in order to achieve better operational results.