Google General Information
Google rolled out its first trial version in 1999 and is now perhaps the best known search engine on the net, and one of the largest with a claimed index size of
8,058,044,651 web pages as of June 2005.
Since the beginning of 2005 Google has been quite unpredictable in regards to both its update periods and how it is ranking pages for relevancy.
With a search share estimated at 48% of the search market share (January of 2005) Google commands by far the largest search share of any search engine.
The awesome market share has resulted in Google being an essential search engine for commercial sites, and has resulted a lot of spamming the index, decreasing the relevancy of the results. Google naturally responds to these attempts at manipulation of its results and the resulting battle is not helping Googles relevancy.
See the search engine relationship chart to see who Google supplies results to and gets results from.
Google Guidelines
Google's Technology in a nutshell
Google Quirks
Google has a few quirks you might want to be aware of:
- When displaying the number of backlinks using the link:URL command Google does not display all of the links it knows about, but only perhaps a random 10% sample. Keep this in mind when using this search function. It may be that some of the special Google search functions (such as allinanchor:) are being manipulated in a similar fashion.
- Google does not keep all of its pages in one large database but has several that are used in decreasing order
of preference.
- When looking for pages that are relevant to a particular search Google first searches in a special database called the "short barrels" which only contains information gleaned from page titles and anchor text links. If Google finds enough results to satisfy its desired ranking pool size it may only use this database for ranking.
- If there are not enough results to fill the ranking pool from step 1 Google will then go to its main index to get more ranking candidates.
- If there are still not enough ranking candidates Google may then go to the "supplemental index" for additional ranking candidates. It seems clear to me that pages which are relegated to the supplemental index are not going to rank well with the possible exception of very rare search terms.
- Google claim that PageRank is the heart of their search engine, but do not make public the actual Pagerank which they use in their ranking algo. Instead they provided two methods for "visualizing" the PageRank of a page
- The Google toolbar which displays a graphical green bar representation of PR on a scale of 1 to 10. In this scheme all possible real PRs are assigned to one of ten boxes (likely by some sort of logarithmic scaled scheme) so that a PR5 for instance covers a range of real PRs with no way of knowing whether you are in the top or bottom of the range.
- The similar PR display alongside a listing in the Google directory appears similar but is in reality quite different. The directory "green bar" is not divided into ten divisions but into eight divisions based on a total width of the bar at 40 pixels. Divisions are unevenly spaced at 0,5,11,16,22,27,32,38 and 40 pixels. There is again no way of knowing what True PageRank these graphics represent.
- Google will only spider and index the first 101 kbytes of any given page.
- Google have recently (July 2005) changed the information on their webmaster guidelines to advise that they no longer include pages with ?id= in the URL.
Google Ranking Factors
Googles ranking algorithm is perhaps the most complex of any search engine and changes often, making it difficult to analyze with any accuracy. Some of the known factors which affect Google rankings are:
Keyword density is not a ranking factor in Google per se, but you do have to have the search term on the page a few times and especially at the top of the page in a header, and in the first paragraph.
PageRank Owned by Stanford University but on a long term lease to Google PageRank plays some small part in Google rankings and is applied after the relevancy rankings. It is of considerable commercial value however.
Anchor text links Often considered the most important ranking factor in Google, there appear to be many variations in exactly how anchor text links are weighted in the Google algo. It appears the guestbook links are worth next to nothing, links form forums worth a bit more, and links from relevant sites may garner additional weight.
Page titles have been an important factor in Google since the early days, when together with anchor text were given their own database which is preferentially used when ranking a query. Google truncates the page title in the SERPs after about 60 characters, but may record more.
H tags Words in H tags are given better than average weighting by Google.
Font modifiers Google records and assigns an additional weighting to font modifiers such as bold, font size relative to the normal page font, underline and italics.
alt tags Google records the content on alt tags on images and gives unlinked images some ranking weight and gives the alt tag on linked images nearly the same weight as anchor text.
Searching with Google
Google has a number of special search operators which can give you more information than simple boolean searches.
Alternate query types
cache: when combined with a domain name displays Googles cache of that page.
link: when combined with a domain name shows a very abbreviated list of links Google knows about for that page, Perhaps on the order of 10%. Note that there can be no space between the colon and the URL.
related: when combined with a URL displays a list of "similar" pages but why they appear similar to Google is an open question. Note that there can be no space between the colon and the URL.
info: when combined with a URL will show information related to the page such as the page title and description and will offer to show more information such as related pages, links, etc.Note that there can be no space between the colon and the URL and that you can get the same information by typing the URL into the search bar.
Query modifiers
site: restricts the results to those websites in the given domain. Example Help site: google.com will find all pages in the google.com domain which have the word Help. This operator only works with no space between the : and the URL
allintitle: restricts the results to those with all of the query words in the title. allintitle: google search will only return results from pages that have both google and search in the page title.
intitle: restricts the results to documents containing the word in the title. Example intitle: google will return all the pages that have google in the page title.
allinurl: if you start a query with allinurl: Google will restrict the results to those with all of the query words in the url: allinurl: google will return only documents that have both "google" and "search" in the url.
Note that allinurl: works on words, not url components. In particular, it ignores punctuation.
inurl: with inurl:
Putting "inurl:" in front of every word in your query is equivalent to putting "allinurl:" at the front of your query: [inurl:google inurl:search] is the same as [allinurl: google search].
Inanchor: Was previously listed by Google as a way to find words in anchor text, but is no longer listed on the official Google search operators page.
allinanchor: Was previously listed as way to find a series of words all of which occurred in anchor text, but is no longer listed on the official Google search operator page, though it still seems to work after a fashion.
