Google Blog Search Quality Factors, new blog search patent application Options

[bragadocchio]

In addition to a relevance score for results in Google's Blog Search, a new patent application from Google shows a number of positive and negative factors that may play a role in the rankings of blog posts.

I wrote a long post detailing these at: Positive and Negative Quality Ranking Factors from Google’s Blog Search (Patent Application)

The patent application is:

Ranking blog documents

The positive factors include:

How many RSS subscriptions there are to the blog,
How often people click on a link to the post in search results,
How many blogrolls the blog is in,
How many "high quality" blogrolls the blog is in,
If the blog offers visitors the chance to tag posts, whether people are tagging them,
References to the blog by sources other than blogs,
Pagerank, and;
Others.

The negative factors include :

If new posts appear in short bursts or at predictable intervals,
If the content of the posts doesn't match the content of feeds from the posts,
If the content includes a lot of spam related keywords,
If a lot of content is duplicated in multiple posts from a blog,
Whether posts are the same size, or roughly the same size,
Link distribution of the blog,
If posts primarily link to one page or site, and;
Others.

What do you think of the results at Google Blog Search?

Do you think they are using any of these quality scores?

[kulpreet_singh]

I love the fact that Google blog search documents can be ranked by time frame, and then by relevancy.

So you can check documents in the last month, by relevancy. Sometimes if you rank "by Date" you don't get relevant documents, but if you rank by relevancy and restrict the time frame, it provides for great results.

Recently Google has also added this to their news search.


This post has been edited by kulpreet_singh: Mar 18 2007, 12:50 AM

[bragadocchio]

Being able to limit results by time frame is a nice feature, for blogs and for news.

I guess part of the reason why that works is because blog posts have time stamps associated with them, and news articles often have associated dates of publication, too.

The timestamps also help in determining the frequency of new posts, which can be used as a negative quality factor when there's some statistically significant aspect to it, such as posts being published every 3 hours and 35 minutes at the same blog.

[A.N.Onym]

I have tested G Blog Search against my blog and thought that:
- if G can't judge the amount of RSS readers for a blog, the data shouldn't be counted. Otherwise, the data is skewed towards those blogs, who show such data
- if you say tagging is a positive sign for Google, it bears no relevance to the quality of the blog. I offer no visitor tagging (though I sometimes tag posts myself), so it doesn't mean my blog is of less quality
- moreover, if I post once a week, on Tuesdays (which I try to do), I may be punished. What for? Sure, off the schedule posts save my blog Google reputation, but I don't think it is that necessary

All in all, I think age matters a lot, as I wasn't able to find any posts through blog search, though some of them are of superb quality and have a number of links from other blogs (and internal linking).

This post has been edited by A.N.Onym: Mar 18 2007, 03:02 AM

[bwelford]

I think it's critical to ping the news feed aggregators using Ping-o-matic or my favourite, Pingoat. I find that if I do that then Google Blogsearch will show the new entries within minutes. If not I've seen it not indexing for 18 hours. I've not waited longer than that before breaking down and using Pingoat.

As mentioned in another thread, it's not clear that WordPress blogs will correctly ping the omnibus pingers such as Pingoat. That's still something I'm looking into.

[A.N.Onym]

I guess I can switch from pingomatic to Pingoat for a month and see what happens. But I doubt it'll be that helpful.

[kulpreet_singh]

With Wordpress I found I didn't need to ping. Posts I made would show up on Google blog search within minutes or an hour at most.

This post has been edited by kulpreet_singh: Mar 18 2007, 07:30 AM

[bwelford]

That's interesting, kulpreet_singh. In your WordPress Admin, what are you using as Update Services under Options > Writing. The precise URLs would be particularly helpful.

[TheBiz]

I think Google blog search is improving and it
would not surprise me if they are already
incorporating some of those factors mentioned
in the patent. (Though I can not point to any
evidence of this)

Google already has a pretty
good overview of blog subscriber numbers
from the market share that Google Reader
has and it would seem obvious that feed
subscriber numbers and the rate of change (positive or
negative) in subscriber numbers is a signal that they could use
to determine quality.

[tamar]

> How many RSS subscriptions there are to the blog,

Interesting. I wonder what criteria they would use for RSS subscriptions. Would Google know about, say, Bloglines subscriptions, or would this be skewed toward Google Reader subscribers?

[bragadocchio]

I've been wondering that too, Tamar.

Welcome to the forums, TheBiz

I think that Google Blog search is improving, too. I'd actually like to see it as one of the tabs above the Google search box. Don't know if that would happen, but it would be nice.

I offer no visitor tagging (though I sometimes tag posts myself), so it doesn't mean my blog is of less quality

I don't know of a lot of blogs that offer visitor tagging. This does seem like a strange criteria. Interesting that they would include this, and not something about comments.

[bwelford]

I think that Google Blog search is improving, too. I'd actually like to see it as one of the tabs above the Google search box.

Back in May last year, I suggested Google might just replace the [I'm Feeling Lucky] button on the regular search with a button labeled [Search Blogs]. In other words the regular search would use the same buttons as Google Blogsearch but in the reverse order.

I'm sure that [I'm Feeling Lucky] button doesn't earn its keep. Given the simplicity of the search web page, that's expensive real estate.