WEBMASTER KNOWLEDGE

There are different major searchengines with different manners. To set up your website best for all of them you have to know how they are working.

With several millions of stored web pages it is easily comprehensible that search machines for most retrieval queries thousands of results supply. To bring the many found to sides into a meaningful order one calls a procedure ranking algorithm.

It means meaningfully that the results, which fit best the retrieval
query are to appear as far as possible in front. In other words: it
is to be sorted according to relevance.

This task causes several difficulties. Retrieval queries are very often not clear. Which results does someone, which looks simply for "download", expect?

Apart from it, it is problematic to evaluate the relevance of a text ( pictures and other multimedia elements are completely ignored by search machines ) only after the key terms occurring therein. So a Goethe biography can quite use the crucial word "Goethe" only completely rarely; in order to avoid repetitions, synonyms are used: More closely "poet" ,"writer". How do search machine recognize the fact that this text turns completely around Goethe, while on another sites (as for instance that, which you read evenly), the word Goethe is often used, but only has a very small relevance to the poet ?

The substantial trick - for the searchengines as for us Webmasters
- consists of considering not only the number and relative frequency
of the individual words but also the position. HTML is more a method
as well known much to structure text than web pages to arrange.
Therefore the appropriate HTML tags are suitable for using it for a
relevance regulation.

The most important Ranking criteria

Search machines evaluate websites as particularly relevant, if the appropriate search word occurs e.g. in the title or within a heading. Likewise become terms, which are located further above in the text, than particularly relevantly judges. For retrieval queries, which contain two or more words, about "Goethe biography", it is important that both search words stand as near as possible in the text and/or in the title. Terms, which are located deeply in interlocked tables, are tendentiell rather more badly evaluated.

Also the frequency plays a role. Mostly the relative frequency - often also called density - is more importantly than the absolute number. Only 20 words a long text, in which the word seems to "Goethe" four times, as more relevant the search word "Goethe" is therefore judged, than a text with 500 words with ten occurrencies. Therefore also short sites have usually a better position.

Also the occurrence within the Meta tags "Keyword" or "Description" will result in a higher relevance and so in better rankings.

For most search machines in the meantime also the URL of the side seems to be important . Not only the Domain is meant, also the path and the file name on the webserver. ( See my tips )

Not yet enough: the URL or the description and the description text for diagrams (old attribute img tags) is suitable likewise as additional relevance criteria. And some search machines evaluate obviously also the last update date, whereby often updated sites are better ranked.

Summary

The most important places within the HTML file for your key terms (with falling relevance):

Title
Meta Keywords and Description
URL
IMG tags

The terms should be located far above in the text, matching search words stand together also in the source text close.

The methodology presented here would work very well in a world, in
which all authors of web pages act " honestly " with title etc. On
the one hand spammers try to optimize their sites also with dishonest
means on the other hand many web authors do not worry at all in
the Web to be found. Because often are good pages in the network,
which use neither a suitable title, still Meta tags.

Therefore search machines go increased to over considering also
external information in the ranking algorythm. A central role
thereby take external links: Link Popularity.

Similarly DirectHit functions. Here it is counted how often search machine users click on a certain search result.