How can a WeChat public account have a chance to be recommended by the system?

This article will unveil the WeChat official account recommendation mechanism for you, explore the process of article selection and recommendation, the recommendation process, and strategies for how to increase recommendation traffic, to help you go further on the road of official account creation.

Ever since WeChat official accounts changed their traffic mechanism, more and more people have started to run official accounts, because now official accounts have recommended traffic in addition to the original closed traffic, which means that there is traffic even if there are no fans.

However, the recommendation mechanism of WeChat official accounts is rather vague. Not all articles are recommended, and the amount of recommendation varies. Some articles may not be recommended at all. So how can one have a chance to be recommended by the system? Let’s talk about this topic today.

1. Process of article selection and recommendation

Any UGC platform will have a "security check" process, and the selection process for public account article recommendations is roughly as follows:

First, the publicly released articles can be filtered regardless of whether they have been fully notified or not.

The second is to exclude factors such as violations, including advertising soft articles, malicious marketing, water articles, etc. In addition to article violations, account violations may also not be recommended. Some accounts are often deleted or have poor article quality, and articles from such accounts will not be selected.

Finally, there are factors set by users themselves. For example, if the creator publishes the article in groups or chooses not to be recommended by the platform when publishing, such articles will not be selected.

After completing the above-mentioned condition screening, the system will extract keywords from the article (this operation may have been performed in the content identification stage, but it is placed in this stage for your convenience), and then start matching recommendations.

2. Article Recommendation Process

We open the data analysis in the background of the official account. In the single article analysis, there is a data section called "Recommendation Funnel Analysis" (if your article does not have this data, it may not be recommended). It has three data dimensions, namely the number of exposures (recommendations), the number of readings (clicks), and the number of attentions after reading (conversions) .

In addition, there are two other data: reading rate (click rate) and post-reading attention rate, which are also very important.

From the recommendation funnel model, we can know that when our article is selected for recommendation, the system will give a certain initial recommendation amount. This amount is not fixed and can be large or small. Then, the system will determine whether to make further recommendations based on the feedback data of the article.

Generally speaking , the greater the number of recommendations, the more likely it is that the number of readings and conversions will increase , which is a downward influence.

It also has an upward impact, because the recommendation algorithm is the same as that of other platforms. It cannot directly judge the quality of an article, nor can it know from the beginning whether the user likes the article.

Therefore, it also uses the data of the article to determine whether the article is worth continuing to recommend, and the reading rate is a very critical data. Generally speaking, the larger the value, the more recommended traffic it will get. Of course, there are also cases where the recommended traffic is low.

In addition to the reading rate, data such as the post-reading attention rate, reading time, completion rate, and post-reading interaction rate will also have some impact, but from the comparison of many cases, the impact is not as great as the reading rate. For example, the data of the article below has a high post-reading attention rate and a low reading rate, but the overall recommendation volume is not large.

In general, the recommendation algorithm for public account articles is not perfect yet and should still be undergoing iteration and upgrading. It also involves many traffic channels , including "Look", public account article flow, recommendation section at the end of the article, etc.

3. How to increase referral traffic?

From the perspective of data dimensions, we can manipulate the click-through rate, post-reading attention rate, and reading time.

1. For click-through rate

I believe that friends who often read articles by Watangren know that in this aspect, they can start with the title and cover of the article, but they cannot violate the rules, so they must control the scale.

2. Regarding the post-reading attention rate

This belongs to the conversion module, which is nothing more than a few commonly used routines. One is appropriate guidance, and the other is setting bait. For example, there is some information in the article that can be obtained, but you need to follow this official account to obtain it.

3. Regarding reading time

It actually belongs to the article quality module. We can make the article more readable, or describe the pain point scenario at the beginning, and then gradually enter the topic. This will keep users reading the article slowly, and the reading time will naturally be higher, and the completion rate will also be high.

At the same time, the conversion rate will also increase. If a user clicks in because of the title or cover and finishes reading it in a hurry, it is easy to ignore key information, not to mention the guiding words you left.

4. Regarding the quality of the article

Many people have proposed the principle that articles must be original, but after many tests and trying to post on different types of public accounts, I found that articles do not have to be original to be recommended, but it is not recommended to be exactly the same as others. You can revise it and add your own opinions, but of course it is best to be original.

Author: Tangge;

Source public account: Digging Pond Man (ID: 1090512)