In order to provide search users with a more high-quality and convenient search experience, Baidu Search has never forgotten its original intention, and is committed to working with content producers and resource providers to build a harmonious and healthy search ecosystem.
In 2013, Baidu Search released its first algorithm announcement.After 6 years of development and precipitation, Baidu Search has released a total of 13 algorithms, 35 announcements and algorithm interpretation articles. In 2019, in order to help webmasters to quickly and systematically understand search algorithms and rules, and to better improve the quality of their websites, we sorted out the algorithms and compiled them into [Baidu Search Algorithm Inventory] to help webmasters obtain the most complete information about the algorithms. .
1. Website content quality
1.1 Hurricane Algorithm 2.0 - Severely crack down on bad collection behavior
Hurricane Algorithm 2.0 mainly combats the following four types of bad collection behaviors:
1. There is a large amount of content collected and transported from content producers such as other sites or official accounts. The information is not integrated, the typesetting is chaotic, the article is poorly readable, and there are obvious traces of collection, which have no benefit to users.
2. There are multiple articles spliced together, the article logic is poor, it cannot meet the needs of users, and the reading experience is poor.
3. There is a large amount of collected content in the site, and the content productivity of the site itself is extremely poor.
4. There is a large amount of collected content that is not related to the theme of this site, the focus is poor in the field, and the traffic is maliciously obtained.For example, technology websites collect a lot of entertainment gossip, social news, etc.
1.2 Hurricane Algorithm 3.0 - Controlling Cross-domain Collection and Station Group Problems
Baidu search upgraded the hurricane algorithm in August 2019 and launched the hurricane algorithm 8.
This algorithm upgrade is mainly aimed at cross-domain collection and station group problems, and will cover PC sites, H5 sites, and smart applets under Baidu Search.For sites/smart applets covered by the algorithm, the display of search results will be restricted as appropriate according to the severity of the violation.
The relevant rules of Hurricane Algorithm 3.0 are described in detail below.
XNUMX. Cross-field collection:
Refers to the site/smart applet publishing content that does not belong to the domain of the site/smart applet in order to obtain more traffic. Usually, the content is collected from the Internet, with low content quality and relevance, and low value to search users.For such behavioral searches, it will be judged that the site/smart applet has insufficient focus in the field, and there will be different levels of restrictions.
Cross-domain collection mainly includes the following two types of problems:
The first category: The content/title/keyword/abstract of the main site or homepage shows that the site has a definite field or industry, but the published content is not related to this field, or the relevance is low.
Such as: food smart applet publish football related content
The second category: The site/smart applet does not have a clear field or industry, the content involves multiple fields or industries, the field is vague, and the field focus is low.
For example: the content of the smart applet involves multiple fields
XNUMX. Station group problem:
Refers to the behavior of constructing multiple sites/smart applets in batches to obtain search traffic.Most of the sites/smart applets in the site group have low quality, low resource scarcity, high content similarity, and even reuse the same template, which is difficult to meet the needs of search users.
For example: multiple smart applets reuse the same template, with low content quality and high similarity
1.3 Drizzle Algorithm - Beneficial Behavior in the B2B Field
In June 2018, Baidu Search released the Drizzle Algorithm to regulate and guide the benefits of websites in the B6B industry.
The drizzle algorithm is mainly aimed at the following problems:
1. Page title cheating and misleading, mainly in the following four categories:
(1) It is not the official website of the company, and the title contains words such as "official website", which misleads search users into thinking that it is the official website
(2) Similar keywords are stacked in the page title;
Such as - website construction - Beijing website construction - high-end website construction - SEO website construction - small program website construction - world top 500 website construction - top ten website construction companies -米国生活
(3) Martian characters or special symbols are interspersed in the page title, which affects normal reading;
Such as: Which one is better for website construction✳☆★★★★?Beijing's first choice for website construction米国生活official
(4) Interspersed benefit methods in the page title.
Such as: website construction SEO optimization one-stop service phone 4000110200/18612888118
2. Beneficial information exists in the content of the page body, mainly including the following three types of problems:
(1) The content of the text is incomplete and frequently interspersed and deformed.
Specifically, it refers to: the content of the text is a splicing of worthless content or the content of the text is incomplete, and at the same time, the contact information is replaced by "English letters", "homonyms of Chinese characters", "special matching", "full-width matching", etc., which are interspersed in the text content. .
Example: 5992 eight ⑧⑧⑧
(2) There are beneficiary contact information in the pictures attached to the article.
Specifically, it means that the contact information watermark area in the picture is large, which seriously damages the browsing experience.
Generally speaking, if you can't write a contact number, don't write it, including articles, etc. If there are too many contact numbers, there will be too many advertisements. Except for Baidu, all advertisements need to be punished. Do you understand?
(3) The content of the product or service information in the text is randomly collected and spliced, the layout is disordered, and the user cannot obtain the product or service information from the page.
1.4 Blue Sky Algorithm - Severely crack down on the sale of soft articles and catalogs
In November 2016, Baidu search launched the blue sky algorithm, which continued to crack down on news sites selling soft articles and catalogs, and also made users search for blue sky.Sites covered by the "Blue Sky Algorithm" will be downgraded in the Baidu search system.
Description of the problem: Buying a directory is a website with a relatively high rating by the website. It sells its own website space and its article publishing opportunities/html static page storage opportunities. There is nothing wrong with selling it. It may be suspected of making money from advertising. Baidu will give you permission. What do you sell for money.
For more information on the Blue Sky Algorithm, please refer to:
2016.11.21 Baidu launched the blue sky algorithm to crack down on news sites selling catalogs
1.5 Aurora Algorithm - Advocating attention to landing page time specification
2018Years5At the end of the month, Baidu search launched the Aurora algorithm, which aims to advocate the resource side to pay attention to the time specification of the landing page of the website.The landing page time factor is an important reference for Baidu search to judge the inclusion, display and sorting results of websites.
In order to allow users to obtain a more satisfactory search and browsing experience, Baidu provides more opportunities for inclusion and display of webpages that meet the time factor requirements of landing pages and have high timeliness, while reducing the display opportunities for websites that do not meet the requirements.
Baidu search landing page time factor specification
The landing page time factor is an important reference for Baidu Search to judge the inclusion, display, and ranking results of websites. Baidu Search integrates users' actual feelings about the time factor in the landing page, and finds that there are page time labels on the landing pages of a large number of websites on the PC and mobile terminals. Unfriendly behaviors such as unclear pages, no time on pages, etc., are not friendly to the user's browsing experience.In order to allow search users to obtain a more satisfactory search and browsing experience, we will give more opportunities for inclusion and display of pages that meet the time factor requirements of the landing page and have high timeliness, and reduce the display opportunities for websites that do not meet the requirements.
In order to avoid the loss of high-quality sites, an important partner of Baidu Search, Baidu Search has launched the "Baidu Search Landing Page Time Factor Specification", which opens the extraction specification for the landing page time factor to meet the browsing experience of search users and avoid webmasters. cause losses and achieve win-win results.
XNUMX. Time factor submission format: PC and mobile pages only support JSON-LD format submission time factor.
*The reproducible effect is as follows:
1. This code is the same set of code as the jsonld code used in the page modification in the background of the bear paw number. If the time has been marked in the page modification of the bear paw number, there is no need to remark the time on the page multiple times.
2. If the bear's paw account is not activated, the appid can choose "No""
XNUMX. Time factor field naming:
|Content release time||
YYYY-MM-DDThh:mm:ss datetime (please note that the date and time are separated by a "T")
|Content update time
1) If the content is created for the first time, please fill in the same time as pubdate;
2) If the page is Q&A or forum, please fill in the time corresponding to the first answer that the user can see
|Last reply time
1) If the page is Q&A, forum and has reply information, please fill in the time corresponding to the latest reply in this field
*Requirement: The mobile station requires the submission of the Bear's paw appid at the same time to improve the reliability of the time factor; the PC terminal does not require the submission of the Bear's paw appid for the time being.
XNUMX. Page submission time factor specification:
Depending on the page type, the time users need to refer to is different.Below, according to different page types, it is recommended to submit different times in priority.
Including main website homepage (such as Sina homepage, Netease homepage), sub-channel homepage (such as Sina Sports Channel) and so on.The characteristics of this type of page are that it has a variety of content and links, and the content is updated quickly, and it is sufficient to provide the latest update time.
page style example
|Sina Homepage, NetEase Homepage, Enterprise Official Website||Provides the latest update time (update) of all contents of this list|
Sub-channel home page
|Sina Sports, NetEase Technology, Sina NBA|
Positive teaching materials on the home page:
Baidu Investment Products: https://raventech.cn/
2. List page
Similar to the home page, this type of page also has multiple contents and multiple links. The overall page is updated quickly, and the latest update time needs to be provided.A special page is a special list page, which is generally a high-quality content aggregation page for the same event. In addition to the update time, the release time of the special page is also very important.
List page type
page style example
Novel/Video Cover Page
|Fighting against the sky, emergency doctor||Provide the latest update time (update) of all contents of this list;
If there is a release time, it is recommended to supplement the release date of the special page (pubdate);
*Update time (update) is required time
Search/Filter Results Page
|58 dog food screening, Jingdong s8 screening|
Post bar/blog/live and other personal homepage
|Live personal homepage, bar homepage|
|The Nineteenth National Congress of the Communist Party of China|
List of positive teaching materials:
3. Details page
These types of pages are mainly content-based pages, including article detail pages, video playback pages, and novel chapter content detail pages.
The interactive detail pages such as Q&A and forum are special and need to provide more content. See the table below for details.
Details face type
page style example
Article details page
|News articles, recipe introductions||Prioritize the publication time of the content (pubdate)
If the content is updated, the content update time (update) needs to be supplemented
*Publishing time (pubdate) is required; the time is accurate to the minute level
Video play page
Commodities/Products/Yellow Pages/Supply and Demand Pages
|Commodity, supply and demand page|
Company introduction page
|Company introduction page, company information|
Q&A details page
|Detailed pages with questions and answers, such as: Baidu Knows, Zhihu Q&A||Provide the issue publication time (pubdate)
Answer time visible to first user (update)
Last Answer Time (lrDate)
*All three times are required
Forum details page
|Forum post pages, such as Tieba, Tianya, etc.||Provide the post publication time (pubdate)
Reply time visible to the first user (update)
Last reply time (lrDate)
*All three times are required
Positive teaching examples:
For more information on the aurora algorithm, please refer to:
2018.5.7 Baidu search landing page time factor specification
2018.5.31 Baidu search launched Aurora Algorithm to advocate the importance of landing page time specification
This article is transferred from Baidu webmaster platform: https://ziyuan.baidu.com/college/courseinfo?id=2797&page=2
- Website Content Quality Algorithm - 4 Articles
- User Requirement Satisfaction Algorithm - 3+ Articles
- Search malicious competition algorithm - 1
- Station Security Problem Algorithm - 2 Articles
- Station Security Problem Algorithm - 3 Articles