Forum Moderators: open

Message Too Old, No Replies

Theory on Google's new algo

just an idea that fits my situation

         

Spica

1:44 pm on May 20, 2003 (gmt 0)

10+ Year Member



I have a new theory about Google's new algo for those of us who have just lost ranking on their major targeted keyword or keyword combination. If I am right, that could be a really good improvement overall for the users, but one that unfortunately hurts me at this point. Fortunately, I can fix that easily for next time.

Here is my idea:

What if Google is now ranking sites based not only on the importance of the keyword on that particular page and in the anchor text of all inbound links, but also on the overall relevance of this keyword on all the pages that link to that page? In other words, extending the idea of THEME from just the anchor text of your inbound links to the entire pages that contain these links. Note that this would be a great way to discount irrelevant, free-for-all link exchanges and random guestbook entries.

Such an algo modification would explain my current situation: my site is all about "original blue widgets", which is the title of my index page, and these words come up several times on that page. My internal pages account for the vast majority of my inbound links. However, most of these internal pages show pictures of different blue widgets, and I didn't write "blue widgets" on these pages because it's obvious to the users that these pictures are all about blue widgets. Instead, I titled these pages and images "original round widgets", "original feathered widgets", etc. Since spiders can't see, they would not classify these pages as relevant to the "blue widgets" keyword combination. As a result, my site has dropped in the new index from page 1 to page 3 for a search on "blue widgets", but it is up for any other keyword combination that has blue widgets together with another relevant keyword (e.g., original, round, feathered) that also comes up frequently on my site.

That said, I still don't understand why they are using old data from the February crawl. That part is really upsetting.

albert

1:54 pm on May 20, 2003 (gmt 0)

10+ Year Member



There's another thread where somewhat similar thoughts were posted:

[webmasterworld.com ]

But it can't be proved, so far.

digitalghost

2:04 pm on May 20, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



>>Such an algo modification would explain my current situation

If you look for evidence to support your conjecture you're sure to find it. I can find evidence to support my current situation and my position is quite different from yours. So we have both have evidence to support two entirely different theories, which leaves us where? ;)