How Google Could Determine Page Categories on Website Pages?

How Google Could Determine Page Categories on Website Pages? Since 2006, Google was the first to start showing links to websites in the first results of any query. In the recent Google live test, Google started showing expanded sitelinks within search results that have tabs that are above the sitelinks and show categories. The results of the experiment were reported this week by the SEO Consult SEO Consult team. SEO Consult in the article Google’s new SERP test Tabular Mega Sitelinks.

In my last blog post I asked What happens if Google include Categories for Search Results and let you edit them? I didn’t anticipate Google testing categories for web pages in their presentation of websitelinks however. I did find an interesting update to an earlier Google patent, which was released as a patent application in the pending stage regarding categories for ads of the AdSense type.

The continuation patent filing contained the latest claims section which explained the way Google could interpret web page categories in the context of displaying AdSense ads. It is possible that this is not the precise method Google could use to classify web pages into categories to determine the purpose of linking or even the specific categories that Google might display in search results in order to allow users to restrict the results they receive depending on the click on those categories. It does however illustrate possible ways in which Google might categorize websites.

The patent application reads:

Methods and apparatus for serving relevant advertisements
The idea was invented by Jeffrey A. Dean, Georges R. Harik and Paul Bucheit
US Patent Application 20120173334
Published on July 5, 2012
Filed: March 15th, 2012

Abstract

The relevancy of advertisements to the interests of a user can be improved. In one instance the content on the web page is examined to find a list of some or more subjects that are associated to that particular website. Advertising is thought as relevant for the web page when it is linked to keywords that are part of the list of at least one or more subjects.

A few or all of these advertisements could be displayed together with the web page or similar web pages.

While this patent is focused specifically on Adsense or search-engine marketing, it could possibly have an insight to SEO regarding how websites can be classified for different reasons also.

There are numerous ways Google could categorize pages in order to determine the ads that should be displayed on them:

How Google may decide on the Web Pages Categories

1. Every word on a webpage could be considered a potential subject.

2. A threshold-based method can be utilized to ensure that if a word appears more than a specific number of timeson the page, then it may be considered a subject for the page.

3. Terms that are used more often on a webpage than terms appearing less often could be given more importance.

4. Terms on the website that do not appear frequently across the Web may be considered to have more significance than words that are more frequent in the Web. For instance, the word “the” appears very frequently on the Web however the term “Chianti” appears much less often. In the event that the words “the” and “Chianti” appear on the same webpage in the same page, then the word “Chianti” may be given more weight as a possible subject to the web page.

5. Utilizing any method of weighting in determining the weight of the content of any page, only highest scoring words could be considered to be subjects for the page.

6. anchor text linking to a web page can be used to identify the category of a page. A page that is connected with the anchor word “Travels in Italy” might be considered to be about traveling to Italy.

7. It is possible that the name of a webpage could be interpreted as an indication of what topic that the page covers.

8. If an page on a particular topic is linked with the site that is being classified, the link could indicate that the pages have a similarity to each other and the subject of the first page could be used on the second page.

9. The top (1,4,10) the number of search queries a page is ranked for in results from a search could be used to determine the subject for that page.

10. The pages’ topics that are connected to each other in ways like being part of the same directory could be attributable to the page specifically targeted.

11. The history of search queries of visitors to the page could be used to determine the topic of the page. For instance, if a user is searching for “Italian wine,” and during the same search, the query session opens a page that talks about “Travels in Italy,” the topic of the page that is visited could help determine if “Italian” and/or “wine” could be subjects that could be used on”Travels to Italy. “Travels in Italy” page.

Take Away

Some of the methods that are described in this patent application could be employed in conjunction to create subjects or categories for web pages. For instance the number of links on a page may be a bit sporadic or absent, or perhaps not precise (like “click here”) or may not be enough to determine the proper category for a web page.

Or , the words that appear the most often on a page may be stop-words like “the,” or “a,” or “this,” and just aren’t particularly descriptive or useful as subjects.

Some websites don’t do an excellent job of providing descriptive titles for their pages. Some sites may have the same title on each page, or extremely brief titles, or even extremely long and insignificant titles.

A website that sells fresh steaks may include an affiliate link to a store which sells wines, to provide visitors with a suggestion for a fantastic dining experience. This is a wonderful use for linking, however the website selling steaks is distinct topics from the one selling wine. The topics on those pages aren’t sufficiently similar to make the same topic applicable to the steak website for the wine website.

However, if you take a look at several the signals (and other signals that could be utilized) and then use them all together to see how well they are in agreement with each other, it could provide a feasible method to create the subject or category of an entire page.

If someone is looking at a website page and is browsing, they shouldn’t be able to decide which category to use for the majority of pages. The search engine must be able to rely on clues that can be derived mathematically.

What do you tell the search engines about your website page’s categories when you choose the page title, anchor text that is displayed to pages, the choice of words that you use on these pages, as well as the selection of words you are trying to optimize your pages for?

These decisions could affect not just the types of ads that Google will display on your website when you use Adsense. They could be used to determine page categories on Sitelinks that link to those pages or even the categories Google could show in its search results.