| The Google patent application submitted in March, | | | | patterns over time. How users got to the page in |
| 2005 has generated a good deal of debate among | | | | question, how long they stayed there, how many times |
| search engine optimization experts. The patent | | | | the particular page was clicked on when it was |
| document contains many general suggestions about | | | | presented in a search...a very impressive (bewildering?) |
| the direction Google wants to move their search | | | | array of factors. |
| criteria and ranking techniques in the near future. | | | | In fact this is an ingenious attempt to solve the "spam" |
| The document points out two areas in particular in | | | | and "staleness" problems at the same time. The major |
| which "there remains a need to improve the quality of | | | | assumption is that up-to-date "relevant" content -- the |
| results generated by search engines." (0009) These | | | | kind the search engines are supposed to be giving us |
| two areas are | | | | -- will be regularly updated, will be inter-connected by |
| (a) artificially inflated rank due to spamming techniques, | | | | an ever-increasing (and regularly changing) group of |
| and | | | | inbound links. In other words, links will come and go, |
| (b) stale documents that rank higher than fresh ones, | | | | changes will happen gradually, and "spikes" in either |
| and therefore "degrade the search results". | | | | traffic or increased link activity will be sure signs of |
| Google's ingenious proposal is to deal with both of | | | | spamming activity. |
| these problems by focusing on the history of web | | | | Conclusions |
| documents and web links. Assuming they have the | | | | Whether all of these measures will ever be fully |
| technology to record such a massive amount of | | | | implemented or not is beside the point. These |
| information, their objective seems to be to keep a | | | | suggestions make sense, and will be adopted to some |
| detailed record of the pattern of changes within web | | | | extent by all search engines. The future has been |
| pages. | | | | defined, and it is up to creators of websites and online |
| This should address the spam issue by revealing | | | | marketers to make the most of it. |
| unnatural patterns of change. Too many links too | | | | The most important conclusions we can take from the |
| quickly suggests "unnatural" linking activity has been | | | | patent application is that the history of our pages |
| taking place. Significant links that come and go might | | | | matters. In practical terms, this means: |
| suggest that expensive links are being purchased on a | | | | -- Rapid and wholesale changes in content will be |
| temporary basis and are not "natural". | | | | looked upon with suspicion |
| And it should address the "staleness" issue by looking | | | | -- Rapid increases in numbers of inbound and outbound |
| at the way specific pages have been updated. If a | | | | links will trigger red flags |
| page that has ranked high in specific searches has not | | | | -- Changes in anchor text that alter or remove its |
| been updated for a period of time, this will be seen as | | | | relationship to on-page content will be suspect |
| a reason to downgrade the importance of that page. | | | | -- Lack of regular and steady (but not radical) changes |
| Other pages with more activity, more up to date | | | | will get your pages labelled "stale" |
| information, and more linking activity, all other things | | | | -- Links that were valuable last year (or month?) will |
| being equal, will rank higher. | | | | not be as valuable this year (or month) because they |
| History is more important than ever | | | | are becoming "stale". |
| This means Google either already gives, or intends to | | | | In other words, webmasters and internet marketers |
| give the "history" of documents more significance. And | | | | must keep adding content, keep upgrading their pages, |
| not just the date when the document is created, or | | | | keep improving and adding new ones, continue to get |
| most recently changed. They also propose tracking | | | | new links, and freshen up their old ones if they can. |
| the pattern of the changes in content, changes in | | | | But they should not do any of it too quickly. |
| anchor text of links, changes in numbers and quality of | | | | Think of this "history" component as a method of |
| inbound links, changes in quality and number of | | | | measuring change. It may seem ridiculously vague, but |
| outbound links, changes in other pages within the same | | | | this is the reality we have to deal with. |
| associated group of documents, and even changes | | | | In the new world order, change has three speeds: Too |
| within the pages linking to a document. | | | | Slow, Too Fast, and Just Right. |
| On top of that, they propose tracking user habits and | | | | |