There are a lot of people will encounter such a situation, why there are many published articles in the recommended amount of media platforms from above, but the pitiful amount of reading it? Why sometimes recommended amount and proportion of the recommended amount of reading is normal, when suddenly stopped recommend it? Since the recommended amount and the amount of reading media platform in the end what does it matter? Why do some articles more than the recommended amount less some articles recommend it? For this series of questions, we first need to know is from the media platform recommended mechanisms and rules. If there is no media articles from the recommended amount, it means that no amount of reading. All from the core algorithm and recommendation mechanism media platform has been the direction of media people from the study.
Whatever you do, are recommended mechanisms exist. Today small just like we talk about media platform explosion recommendation mechanism from the text, whether it is done or drainage earnings are need to know. Different mechanisms platform recommendation algorithm is different, each platform different time periods will change with current events varies among the keyword recommendation algorithm. The recommendation system will be "content with the right amount of hot keyword article" During this time recommended lifting weights.
I believe we will have several questions:
1, Why the same article, published at different times is not the same effect
Without considering the platform and fans of interactive, publishing the results are not the same depends largely on a hot keyword grasp, there is a different reading habits of different users, then the result of articles published is not the same.
2, why the same article, the effect is not the same in different accounts released
With the self-rule constantly improve the mechanism media platforms, most stars are used to judge the quality or value of an account. The quality points will determine the recommended mechanism and display articles. The same article published the same time, high-quality account the effect of lower quality than good.
3. Why is the title of the party so well
In the flow of information from the media which publish articles mostly appear as a "cover headline +" to the user, the user does not open in the case of the specific content of the article is not known among the. All judges of the article comes from the title and cover image, excellent with the desire to have more easy for users clicked.
4. What data will directly or indirectly affect the system strength of recommendation of the article
Articles are clicks, the browser part-time view of the article, the article browsing speed, the number of Chan obtained, the number of comments obtained, the number of available collections to share the article (if there is to be read), bounce rate ......
I studied more fire from the media platform: Today's headlines, fish numbers, numbers that interesting headlines, and some other media platforms derived from some recommended mechanisms: generally speaking:
Submit a new article - to heavy - Audit - Recommended
Usually this is the process, I will look for this process is:
First, deduplication
Including content deduplication, deduplication similar themes, like the title picture with deduplication.
In fact, it can be understood as users like to see fresh content information. If the display information to the user's daily are similar, then the user is definitely not like it.
Second, review
After the de-emphasis, is the audit, the platform would be content with the title for review, should the illegal information platform will appear interception.
Illegal information includes:
Title: sensitive information, exaggeration, fraudulent use of the name of the headlines, there will be a tendency for headline unfavorable public opinion, with special symbols.
Body: sensitive topics, format incomplete, duplicate content, vulgar, worthless, excessive promotion information.
With map: a blurry picture, there is the promotion of watermarks, including two-dimensional code, the scale is too exposed, easily lead to ambiguity pornography and so on.
For the recommended mechanism, we have to focus on that: Whether audit mechanism and recommend mechanisms for each platform, are similar,
Push is a look at three of the principles of the second trial.
A. Preliminary recommendation
a. content is automatically submitted to the system after the audit released
b. The system extracts keywords preliminary recommendation (keyword heat, quality of content, originality of First Instance directly affects the amount recommended standard)
c. If you find published content for a long time does not pass, it is likely that the system after investigation and the sensitive word to push a manual review. Manual review will affect the recommended amount, if this is the case, please pay attention to the quality of their content, whether sensitive word violation.
II. The second round is recommended
It can be said, but preliminary recommendation system for the first time to test your content. The second round of the recommendation before deciding whether to become a critical text of this article.
The first two factors recommended by the main reference: reading rate (the recommended amount / amount of reading), reading rate, reader comments (interactive), the collection point forwards like (interactive), hot words of user interest.
III. Recommended and after the third round
After the second round of the determination, determines the intensity of the third round of recommendation. In general, the frequency of 2 to 30 minutes per platform push round. Each round of the recommended strength depends on a determination of whether the quality of the content. There are factors that determine reading rate (the recommended amount / amount of reading), reading rate, reader comments (interactive), the collection point forwards like (interactive), hot words of user interest. And so on.
Through the above article we can draw conclusions:
1, to improve reading rate, mainly to improve the optimization in the title and cover art above.
2, increase the time to read the user stays, the need to control the number of words and paragraphs good article.
3, reading speed, straight from the heart to the article, energetic, full of dry goods and valuable.
4, seize fields keyword, keyword relationship with the recommendation of the user.
Want to improve the recommended amount of articles on fundamental, then start from the content or direction, a high-quality article with a good title would certainly like to win the favor of users.
Finally, and to teach you a simple method, the use of aids from the media to improve the quality of their content, can collect and analyze different areas of industry critical text case from media tools by easily essays, extracts common ground among their own analysis Used.
