Ranking Algo Behind Google Blogsearch

Halfdeck · 2007-03-20, 06:39 PM

I know Google Blogsearch still isn't that big, but I found this patent kinda interesting because I don't see alot of patents dealing specifically with blogs.

Note: as far as I know, this has nada to do with blogs ranking in regular organic search results. Also, just because something shows up in a patent doesn't mean its implemented. So I'm wrapping this entire post with a big *maybe* Good food for thought though.

You can read the summary here: http://www.seobythesea.com/?p=541

According to the patent, Google uses relevance and quality scores to rank blog posts in Google Blogsearch (shocking, huh?). Relevance is measured the old fashioned way (frequency of keywords in a post, keywords in H tags, TITLE, etc). Quality score is a little more involved:

Positive quality indicators:

- popuarity, based on number of subscribers
- implied popularity, as measured by clickstream data (when people click on a link while browsing Blogsearch)
- Your blog appears in many blogrolls
- Your blog appears in many high quality blogrolls
- Tagging blog posts - Assigning categories to posts mean you've taken the time to evaluate and categorize your posts (instead of assigning everything to "porn" or "uncategorized")
- Reference to your blog posts in emails, chat discussions
- PageRank of a blog

Negative quality indicators:

- Frequency of posts. Posts published in spurts or regular intervals (e.g. every 10 minutes) imply malicious intent by possible spammers.
- Mismatch between feed and blog content
- Repetitive text element across multiple posts (same ad text, copyright notices in a feed?)
- words/phrases used in posts - if your post contains a high percentage of words found in other posts flagged as spam
- size of post - similar / identical size posts imply automation.
- outbound links - a high percentage of outbound links from a blog pointing to just one or two sites imply an attempt at PageRank manipulation.
- ads - too many ads imply lower quality

http://appft1.uspto.gov/netacgi/nph-...DN/20070061297

2007-03-20, 06:39 PM	#1
Halfdeck You can now put whatever you want in this space :) Join Date: Oct 2004 Location: New Haven, CT Posts: 985	Ranking Algo Behind Google Blogsearch I know Google Blogsearch still isn't that big, but I found this patent kinda interesting because I don't see alot of patents dealing specifically with blogs. Note: as far as I know, this has nada to do with blogs ranking in regular organic search results. Also, just because something shows up in a patent doesn't mean its implemented. So I'm wrapping this entire post with a big maybe Good food for thought though. You can read the summary here: http://www.seobythesea.com/?p=541 According to the patent, Google uses relevance and quality scores to rank blog posts in Google Blogsearch (shocking, huh?). Relevance is measured the old fashioned way (frequency of keywords in a post, keywords in H tags, TITLE, etc). Quality score is a little more involved: Positive quality indicators: - popuarity, based on number of subscribers - implied popularity, as measured by clickstream data (when people click on a link while browsing Blogsearch) - Your blog appears in many blogrolls - Your blog appears in many high quality blogrolls - Tagging blog posts - Assigning categories to posts mean you've taken the time to evaluate and categorize your posts (instead of assigning everything to "porn" or "uncategorized") - Reference to your blog posts in emails, chat discussions - PageRank of a blog Negative quality indicators: - Frequency of posts. Posts published in spurts or regular intervals (e.g. every 10 minutes) imply malicious intent by possible spammers. - Mismatch between feed and blog content - Repetitive text element across multiple posts (same ad text, copyright notices in a feed?) - words/phrases used in posts - if your post contains a high percentage of words found in other posts flagged as spam - size of post - similar / identical size posts imply automation. - outbound links - a high percentage of outbound links from a blog pointing to just one or two sites imply an attempt at PageRank manipulation. - ads - too many ads imply lower quality http://appft1.uspto.gov/netacgi/nph-...DN/20070061297 __________________ Success is going from failure to failure without a loss of enthusiasm.