Relevance site SEO is the most important

Regardless of how the search engine changes, relevance, the factor that determines search quality, will only become more and more important. Whether it is the relevance of the content of the site or the relevance of the site links, these are our white hat seo basis and focus.

　　

Content dependencies

　What is content dependencies

Each site will have its own theme, even large integrated sites, will be around a few words to expand its content writing and publishing. Supplemented by the relevance of high outside the chain of support, over time, in this field, the site has become authoritative, with a high weight. For example, Yang SEO blog is based on SEO, website optimization and search engine for the theme of the blog. So supposedly in the search engine data, these three words occupy the dominant position.

This is the relevance of the content that we can see through Google Webmaster Tools:

　　

Through Webmaster Tools-optimization-content keywords, we can clearly see that the search engine for the overall content of the site to grasp, with the blogger himself for the site positioning is the same. Description, this is a site with a high degree of relevance to the expected keyword.

And a bad example is:

　　

We analyze this site, in the content page title, the general "XXXX Price/quote" as the title. But we see the relevance of the whole station keyword, "price" The word is only the importance of the tenth place, or even "Jingdong" and "Amazon" (participle reason).

In this way, in the white Hat SEO natural optimization, if the user to "XXXX price" for the keyword query, then the content page of the site is difficult to rank in the forefront, or in this long tail keyword, their efforts need to be more than competitors, before they can get rankings and traffic.

　Algorithm for content correlation

Among the 200 ranking factors known to search engines and several algorithms, there are many algorithms on the relevance of content and retrieval terms, and the TF-IDF algorithm described earlier is one of the first, in addition to:

Feature weight calculation: TF-IDF algorithm

Similarity calculation: Consine algorithm

A BM25 or bm25f model in probabilistic retrieval

...... Wait a minute. Of course, the above three types are more characteristic and indeed are used by commercial search engines in the actual retrieval process, in addition to a lot of open algorithms and models.

We SEO is not doing IR research, so the algorithm and model for us, the actual significance is not. But as SEO, basic understanding of these algorithms and models, will be the site content optimization has a huge role.

　Link dependencies

What is content dependencies

Link relevance is the same as the basic concept of content relevance, that is, access to the same industry authoritative site vote, the effect is far greater than the non-industry site.

This is why we say that the best outside the chain is your search for the site target keyword, ranked in the top 10 sites exported outside the chain, the best effect.

So often see some peer in the SEO forum to send some other industry outside the chain, I can not say no effect, but its efficiency is too low. Others send one, probably more than you send 10 or even dozens of. Of course, if it is mass, it is not within the scope of the discussion.

Now there are a lot of keywords based on the relevance of the outside chain tools, such tools should also be carefully selected. Generally such a tool mining is already known by the vast number of SEO, or even the Web site in the database. A regular and not very large number of releases, can produce a certain effect, but if it is a large number of mass, it is likely to trigger the penalty conditions.

Google some days ago in the update Penguin algorithm, some webmaster mass warning messages, the vast majority of the content of the message is about "unnatural outside the chain." This also shows that the search engine for uncorrelated outside the chain of mass, the penalty is very accurate and strict.

In addition, the links within the site are also relevant and irrelevant. If it is a large number of strong correlation within the chain, it will be within the site of the weight flow, produce very beneficial effect. If it is an internal chain with a weak correlation, it is likely to lose the voting effect.

　　Algorithm for link dependencies

In the link dependency algorithm, I also selected a few well-known and most likely to be used by commercial search engine models and algorithms, interested friends can search the relevant literature for their own research:

Google Patent: PR algorithm

Link Authority degree: Hits algorithm

Set of Compute objects: Salsa algorithm

Expert page subset: Hilltop algorithm

The latter two algorithms are complementary and perfect to the first two algorithms. It is characterized by finding an authoritative page with a high degree of relevance and then empowering other pages.

Based on this, we can draw a conclusion, that is, in the acceptance of the weight of the chain, the first thing to consider is the relevance of the chain, followed by the external link to export the weight of the site itself.

This article by Yangfan original in Yang seo, reprint please keep the link: http://www.seoyangs.com/relevance-seo-foundation.html