Editor's note: 2003 November 16 - this day may become an important milestone for search engine history. It is in this update called "Google Florida" this time, Google has made large-scale changes to its rankings and leads to many of the original ranking of a good website as a dream like a dream. Sun Mountain. Each update cycle may do some small changes to the ranking algorithm, but this big hand is changed, it is not allowed to let the website administrator and even the entire search engine optimization industry is shocked and the start. Not only many of the commercial websites that have been maintained in the spear have left the ranking, but also make many companies that do websites optimized have been impacted. However, because Google has almost occupying the big rivers and mountains of the search market, if you discard the business promotion, it is not a wise move.
What new mechanisms have been adopted for Google's new algorithm, is still the focus of SEO. There are currently two main theories. One is an "expert system", the Hilltop algorithm, that is, this article will be elaborated; the other is the TSPR (Topic-Sensitive-PageRank: Theme Page Level Technology) headed by Dan ThiesW, its probably: When the query occurs, Google will match the topics contained in the query condition and the CIRCA database, the less the Query and the topic, the better the theme page level score effect. In addition, since a given query condition may match multiple topics in the database, any small errors in the page level calculation will be flattened by a certain number of theme page levels, thus approximate theme page Level scores provide high quality search results. Detailed discussion on "TSPR Theory - How to Coping GOOGLE PAGERANK Ranking Algorithm". We believe that no matter what new algorithm in Google, its purpose is intended to provide users with more accurate search results.
HillTop (localscore) algorithm
Author: Atul Gupta
What is the true face of the GOOGLE new algorithm? This is still a fan. On the Google website, there is no new content in the content from its "reward high-quality content website and discharging the website with malicious technology". Nature, as a search engine, this approach is fully understood. The algorithm of the search engine should be confidential, because they will disclose their new algorithms for any comments, consent, or negative words of the new algorithm, and anything to the ranking system may be drilled by some people, and used Malicious manipulation results, thereby destroying the objective fairness of search results.
We have found some discovery based on some of the more feasible speculation and combine the actual research, test, and the development trend of search engine behavior and the knowledge foundation of history. First, let's take a look at some of the more worthy discoveries for Google's new algorithm:
1.Google new algorithm variety of rumors & guess
For Google's new algorithm, there are some rumors and guess:
Start using a dictionary temlor to enter;
The intention is applied to make the business site use the AdWords advertising service;
Start using "Bayesian-based SPAM filtering system";
Start penalty, reciprocal link, "Optimize" webpage, or "link text";
Among the many guess, "Google intention to apply the pressure to use AdWords Advertising service", obviously the wind. Many search engine optimization analysts believe that Google screens the business website through a confidential filtering system "blacklist" to achieve the above objectives. I really don't dare to find it. Explain the following ...
1.1 Will Google screen for search results if the query condition contains "commercial keyword"? Because the loss in this update is some of the business websites, there are many search engine optimization experts that Google begins to screen for query results generated by query conditions with so-called "commercial keyword". From the optimization expert to some of the initial test results made, it is indeed indicating some indications that seems to be able to confirm this speculation.
1.2 How does this move?
If this is true, why do Google do this? The back shield of this speculation is that the Google Algorithm update is just before the Christmas shopping season, and it is available on the market. Analysts believe that Google does this to apply pressure to the commercial site. If these business sites are looking forward to getting a rich profit in the annual American Christmas shopping season, they have to make an AdWords self-service advertisement. Free lunch seems to be. Also, some people think that Google is intended to apply a little bright color to its final performance (ie baseline) to touch its future investors.
1.3 So who has replaced the original chair of the commercial site?
From the results page, the top 20 results seem to be some official sites (.gov), education site (.edu), public welfare organization site (.org), directory and non-US local sites. Since these sites do not need to advertise, analysts believe that the website will not affect Google's advertising income, while also pressing the commercial site and forcing it to join Google AdWords buffet Advertising camp.
1. What is the list of "commercial keywords" on GOOGLE?
From the Google's AdWords advertising system, Google has a large "commercial keyword" database. Interestingly, is equal to the owner of the advertiser and the commercial site, which is "good business words" tells Google, and the advertiser's change in the bidding price of key words, you can see each key "Value" of the word.
1.5 What is the fact? Does Google are indeed filtered out the business site with the "Business Keyword" list?
I personally think that this idea is really contrary to the facts, and I firmly believe that Google is not like the "Filter" site thinking. And I believe that Google has no "business keyword" list. One popular "filter list" theory is that analysts have been drawn from their own signs, and in fact these so-called signs are caused by other reasons, but not "filtration" caused of. Behind I will explain and elaborate on this issue.
It is undeniable that we can understand the results of "filtering" from practice, but we have reason to believe that such results are actually a "by-product" of Google's new algorithm, rather than Google. The new algorithm is impact on commercial websites rather than public welfare websites. I will explain it in detail later ...
Scroogle.org, this website created by a group that doesn't like Google, which can be extracted from Google from Google, and provide users with "unfiltered" in Google. The tool of the result. However, since Google has now updated the algorithm to prohibit such queries, Scroogle.org has no such tools. However, the site still shows a list of so-called "Filter Hit" collected from many other searches from its website. I have studied this list. I think it can only prove that Google does not use any list. Otherwise, how to explain "California Divorce Attorney" is the first and "Books" or "Adult" be flexed at the bottom? Is it in Google's eyes, "California Divorce Attorney" is more business value than "BOOKS" or "Adult"? Google Any from normal ranking mechanisms to filter out the business site attempting to go to Google Brands, and even the credibility of its fair search results. If there is such an attempt, the google is really short, and it is worthy of its brand and its service. Google's stock market (IPO) and ultimate performance will be destroyed once. In addition, Google also said that its "Search Service" and "AdWords Self-Advert" service are divided into two different working sectors, and they are not bound to each other. I still believe this.
2. Several significant changes in Google Search Technology
2.1 Google starts using stemming technology (STEMMING)
Google has indeed adopted a stem technology in the search results. In the morning, if you search for a single query condition, such as "Home Garden", the search results will not appear in the keyword change in "Home Gardens", "Home Garden", and vice versa. For Google's users, this feature may be a good thing, because the results provided are more, but stand in search engine optimization, due to the sudden increase in search results, the competition for keywords Suddenly intensified.
2.2 Google uses spelling error correction tools in search results
Google begins to spell check tools for search results. For example, when the input query condition is "Search Engine Optimisation", Google will only prompt you if you are looking for "Search Engine Optimization", but it is still in line with the "Optimisation" search results. But now you can see the search results of "Optimization". Then: "E-mail solution" and "email solution" are consistent, and "e-Commerce" is consistent with the "ECommerce" search results, which increases the competitiveness of the data.
3. Analysis and estimation of HillTop algorithm
3.1 Why use a new algorithm?
Most of the net fans know that the page level algorithm is advocated by Google's founder Sergey Brin and Larry Page, which is used in the search results in search results. This is also one of the important factors since the acquisition of all the achievements since 1998 Google.
But the page level system also has a basic flaw, and Google knows this. It assigns an absolute "importance value" to the web page based on the number and quality of the site on a web page. At the same time, you will also take into account the page level of the link page. The higher the page level pointing to the external link page of a web page, the higher the page level value of the page page to the page. However, the "page level value" is not targeting the query, and thus a web page even if it is just a key word that is from a query topic, it will also obtain a relatively high rankings due to its high page level value. California in California. Krishna Bharat realized that this kind of scoop in the ranking system based on the page level, and put forward a new algorithm in 2000, he called it as "hilltop" algorithm, in 2001 The month applied for Hilltop patent, and Google became a transferee. It goes without saying that Google is already aware of this new algorithm to bring good results for Google's ranking system.
I am convinced that in Google's recent update, not only the Google's own page level and correlation algorithm are used, but also combined with the effect of using the Hilltop algorithm for intensive adjustment page level, only HillTop algorithm may have been very different from it. Watch.
3.2 What is HillTop algorithm?
It is interested in this algorithm to understand and want to dig the bottom of the person to see the Clever. The Hilltop papers written in Berharat and the detailed Hilltop patent information. But most of us, just know what the Hilltop algorithm is in the end, it is possible. Simply put, the page level determines "authority" of a web page, and HillTop (localscore) determines "authority" of a web page that matches a query condition.
BHARAT: When using the "Key Level" to find "authoritative" webpage, it should not be simply dependent on the size of the value, but it should focus on its correlation with the query topic; that is, not only the page page is required. I have to consider whether the page level of the page is commensurate. If a web page is only with the query theme, then even if its page level is very high, it is meaningless to users. In this way, the links of the document from "related topics" are more meaningful (BHARAT said "expert documentation"), all of the external "expert documentation" "The link constitutes" authoritative value "of the page. The significance of the Hilltop algorithm is that the link between the same subject website should have a higher value than the link of the non-related website.
The HillTop algorithm calculates "authority" of a web page in the following manner (this is just an example of extremely simple):
Taking this keyword to find a general query to find an expert document "文集". The definition of "Expert Document" has strict standards, because the "algority" should be a set of web pages that are easy to manage.
From the list of returned experts, remove the member site (see below *) and the mirror site.
According to the number and quality of the above expert documents, the web page is assigned with a "local portion". Then press the "part of the part" in the page. * Note: Member Site refers from the same domain, or the domain is the same, such as IBM.com, IBM.Co.uk, ibm.co.jp, etc., or means from adjacent IP addresses ( The top three IP addresses, such as 64.129.220.xxx)
In addition, if both sites are hosts under the same IP address, one is considered to be a member site for another site. (Ie, if ww.abc.com www.ibm.com is the host under the same IP address, I recognize www.abc.com is a member site of www.ibm.co.uk's member site) HillTop algorithm an extremely important feature The algorithm is invalid if the search engine is not found to be sufficient, and at least two votes are required), the algorithm is invalid, that is, the return result is zero. In other words, for highly clear query conditions (query words), the results of the HillTop algorithm are likely to be "0". This is a unique feature of the algorithm. As a result, most people in the SEO family are convinced that Google is indeed filtering out the business website with the "business word" filtration list. In fact, when the HillTop algorithm does not work, the "old" Google search results will be displayed. The collection of these query conditions is the "business word list" collected by the SEO family. This effect is unintentional but provides a powerful evidence that Google has indeed adopted HillTop algorithm. November 15, 2003, Google Based on the update of the new algorithm, an analyst pointed out: When the query is inquiry, if you add some "not" unintentional characters, such as "Car Rental - GHJKL ", Google will display the search results of the previous (before the change of algorithm), bypass the so-called" business word "filtration list.
For example, if you look for "real estate -hdfkdhgk", Google will try to return all about "real estate", while the content does not contain "HDFKDHGK". Almost not because the word "hdfkdhgk" is almost no, so according to the truth, the query result returned by Google should be the same as the "Real Estate" query results. However, the result is unexpected: Google's display results are the same as the ranking of search results before the new algorithm. In fact, the Scroogle.org website is to capture the search results of Google different in ranking, and thus derive a so-called "business word" filtration list.
What is the so-called "business word"?
We believe that "business words" filtering effect is just a derivative of Hilltop algorithms. Whenever the user performs a query similar to "real estate -hdfkdhgk", Google will Hilltop the entire query word. When you can't find a "expert file" containing such query words, the algorithm will return a zero value ( That is, zero effect). This means that we have wrapped the HillTop algorithm through this, and the other algorithms of Google have played a role in the process of Google, and the results provided by the query before the algorithm are updated. Yes! Later, Scroogle.org's reputation made Goolge found this bug. Later, Google surely divided the query into two steps, first put the query condition word, and sent the query condition to Hilltop, which, because HillTop no longer parsed the exclusion condition, so Ensure that it is working properly and provides the corresponding search results, then pass this result to the Google Algorithm. Google will eliminate the web page that meets the exclusion condition from the results provided by the exclusion condition in the initial query, and then display the results display Give the user. This is why we now use the above magazine to see the reasons for the search results of "old Google".
GOOGLE new algorithm meaning
The combination of Hilltop algorithm and Google's page level algorithm and page relevance algorithm looks a supercommission, which is almost impeccable. Its marriage has far-reaching significance for links to link popularity / page grades and links from expert filescore. I have listed a simple formula to illustrate the effect and impact of the new algorithm (for reference): "Old" Google Ranking Formula = {(1-D) A (RS)} * {(1-e) B ( Pr * fb)}
"New" Google Rank Formula = {(1-D) A (RS)} * {(1-E) B (Pr * FB)} * {(1-f) C (LS)})} where:
RS = Correlation Score: Based on the web header (Title). Meta tags, headlines, body text. The scores derived from the keywords that appear in the elements such as URL, Picture of ALT text, Anchor Text. PR = Page Level: (Based on the number of pages to your website and its PR value). The original formula is Pr (a) = (1-d) D (P (T1) / C (T1) ... PR (TN) / C (TN)), where D is the damping factor, which is generally set to 0.85; the formula can be described as: the PR value of the A web page is equal to the PR value of the link to the page, divided by respectively The sum of the number of external links. LS = Industry Score (LOCALSCORE): The score calculated according to the expert file.
A, B, C = adjustment control specific gravity: Google is used for sperm adjustment query results
D, E, F = Damping Control: Google is used to refine the results of the query. The "F" value may be "0".
FB = Factor Base: Google Toolbar Mountain Displayed Page Level Scores from 1 to 10, but it is not a linear function, but an index / logarithmic function. After analysis, we believe it has a base of approximately 8. That is, the value of the page of the PR 5 is 8 times more than the PR is 4 times (PR5 = PR4x8), similarly, a website with a PR value is 4,000 times higher than the value of a PR value of 4, and is more than 4,000 times. PR8 = pr4x8x8x8x8).
Advantages of Google New Algorithm / Advantages
The page factor is easy to be utilized and controlled, for example, Meta Tags that is used as ranking factor, as a result of the unreasonable keywords that many websites are flooded, they have to be shouted. In addition, at the beginning, due to the feelings of the scruples, the content of the website is not much, but it is also difficult to temptation, excessive optimization, and even on the page, there is many "invisible" content. Trick. The page factor does impartially impart a "malicious" control power to the website administrator. In Google's new algorithm, the proportion of each ranking factor is as follows:
Total "ranking" score consisting of three parts: correlation score (page factor), page equivalence score (non-page factor) and industry score (non-page factor): Correlation score (RS) = 20%, page level value (PR) ) = 40%, industry score (LS) = 40% of which:
RS (Correlation Score: Page Relevance) - is a conversion result of all SEO efforts;
PR (Page Equation Score: External Link) - is a conversion result of all efforts to establish links;
LS (Industry Score: Expert Document Link) - is the conversion result of the link to the expert file;
It is not difficult to see from the above results, and the score of page optimization only accounts for only 20%, and Google has clearly turned onto the non-page factor, and the right to rank control is increasingly taken from the website administrator. If a level of general optimization experts can get 10 points in RS, and a high level of optimization experts can get 18, 19 points, everyone can see: these two horizontal level of expertise The difference between the room is only 8 minutes! Does the HillTop algorithm run in real time?
The server architecture owned by Google is 10,000 Pentium-level servers distributed on the network. Once the HillTop algorithm is understood, it is difficult to believe that such a Pentium server can have such processing power: Imagine, first find "expert files" from thousands of theme files, then calculate the target page from these The residue of the link of the expert file, then returns the value to other ranking systems of the Google Algorithm and makes further processing - and all of this is within about 0.07 seconds - this will be completed in the search speed of Google. It is indeed incredible.
Running frequency and coverage of HillTop algorithm
We believe that to ensure that Google has always "lightning" search speed, Google will regularly run the batch of search frequencies (ie, the so-called "business word" blacklist) of the search frequency (hot), and store the results store It is used in the future. Google's database has a large number of query words for high quality query frequencies, collecting the keywords used in the real search and its AdWords self-service advertising system. Google is likely to set up the upper limit of keyword search times. Any search frequency higher than this threshold will be incorporated into the HillTop system, and then the HillTop system will regularly run batch of all high-query frequency keywords, possibly One month. The slightly small size batch of the incremental level may be frequent. At the same time, the results of the HillTop system after running batch after running the GOOGLE server, but the slightly larger batch database updates will be more frequent.
For those user query frequencies, this "honor" is included in the Query words of the Hilltop algorithm, and Google will still use the original algorithm and display the original ranking results. Thus for those who are highly clear or specialized, they are expected to maintain the original ranking because they are excluded within the scope of the new algorithm.
Why is Hilltop algorithm to use it for a long time?
Google has gained this patent as early as February 2003, but before the actual investment, you need to ensure that the new algorithm and Google are fully compatible with the page-related system, so it needs to be compatible. A large number of tests, then evaluate the results provided after the algorithm integration, then do the adjustment, then further recycling ... I think all of this requires a lot of time.
Adverse aspects of Google New Algorithms / 瑕
After further analysis, we found that several defects and shortcomings in the algorithm:
The premise of Hilltop is that each expert document is completely fair, and there is no deception and manipulating ingredients. However, the situation may not be such ideal. A small stain of an expert document can have a great negative impact on the ranking.
Hilltop intention to determine a web page industry credibility through the link of the "authoritative" expert document, but it is not possible to ensure that these so-called "expert documents" "" quantity "is indeed a name.
Running a HillTop algorithm requires a large number of computer processing capabilities, so we guess the HillTop algorithm (possibly) is a batch of batch of all high-frequency (hot) queries collected every month. Since the Google Algorithm transfer 40% center from the center of gravity to Hilltop, the scores after the HillTop algorithm are maintained to the next batch, so the search results are probably not too large during this period. In addition, due to the "expert" voting method often does not change much, we may only see some "old" search results during a certain period. This is the purpose of the "providing the latest, best" query results. In addition to "Authoritative Website", users also want to see some fresh content, but for "new" Google, users only hope that the query frequency is not too high, so that the query word that is not included in the Hilltop system Fresh "query results. The difficulty of the new site to the popular query word will be more harmonized. It seems that the threshold of Google's rankings for new websites or new content is getting higher for those specially popular query words.
The commercial site is not difficult to link to the network classification directory, the trade association, government trading institutions, educational institutions, and public welfare organizations, so such sites will occupy the top ten positions of the search results.
Which sites are the most likely to GOOGLE new algorithms?
Member Site / Domain Cluster / MLM (Multi-Layer Marketing System Program) program under the same server.
Excessively rely on the "page" optimization factor.
Relying on a highly competitive query word to get a website.
Some suggestions for website owners:
The ranking rules have changed major changes, and the successful promotion of the website needs to start from the increase in PR value and get more from "expert documents";
To make more mainstream online classification directories (DMOZ, Yahoo, About, Looksmart, etc.), trade catalog, yellow pages, trade associations, and resource pages, etc.
Avoid using a domain cluster / member program or a system with member nature;
Avoid reciprocal links from suspicious FFA sites and link factories.
Several rumors need to be clarified:
The "Over-optimization will be punished" - Filter optimization itself is a spam, which is thus a matter of course. However, the ranking impact currently received by most commercial sites is not due to punishment, but by Google causes the center of gravity from the page factor to non-page factors. Good optimization of the website is still indispensable.
About "establishing external links is no longer important" - establishing a link facility is still more important as before, or even more important. Google Patent Page Level Algorithm still occupies a pivotal position.
About "Google starts using the 'Business Word' Filter List" - Google does not use any "filt list" to punish the business site. Google also has no intention to "punish" or "sieve" anything, it is only to create a system that recognizes the most relevant page.
The site is included in the DMOZ, Google Directory and Business Catalog Remaining Google thinks that the site is a commercial website, thus punishing it - the opposite, because these directorys are very high quality "expert documents", thus getting these The link of the directory will not only be punished, but will bring you higher income.
With the big update of Google's algorithm, we have to admit that if you want to get a good rankings, it is not easy. If you only use simple optimization techniques, such as meta or page optimization, there is almost no way to rank. We need to take into account a lot of algorithm variables, not just a meta tags. Search engine optimization will be more specialized. Old-fashioned optimization tricks have been talked for use or earned. It is now necessary to plan and implement a full range of SEO strategies. To ensure effectiveness, SEO experts need to do now is to invest more wisdom, research, analysis, planning and time. In this case, the optimization of SEO will gradually increase.
The price of PPC (paying clicks) will continue to rise. The PPC advertising mode will not only become a major component of the online marketing strategy, but also to ensure the corresponding search engine accesses. The external link and page level of the website are now increasingly important. It is more important to the impact of the ranking. The combination of search engine optimization and establishment of links is imperative. In addition, the update of search engine rankings will not only be more frequent, but also get more and more. Now, regular search engine optimization maintenance is particularly important. Search engine optimization for the website is not only necessary, but also is very beneficial. Many websites have lost their original ranking, so especially in searching for optimization experts. Every update algorithm is a shrinkage of the list of search engines. The optimization of search engines and rankings become more difficult, but it is more important.
