Friday, August 26, 2005
Yes, I agree there are philosophical differences between what I call a blog and what you call a blog but it's important to realize it's an issue that does actually matter. Especially in light of the recent popularity boom of blogs for marketing and PR purposes. Search engines like Goog have been working on this issue for some time but there have been no real significant improvements in the blog search & classification arena as of yet.
Without further adue, here are my year old thoughts on blogs, blog backlinks and initial conclusions. Although I have more to add and will add, I'd like your feedback on my thoughts.
In order to determine if the weight (amount of pagerank) a page receives from a backlink on a blog is different than from a standard website we must first determine if the link is coming from a blog. Honestly, I'm not sure any of the search engines are doing this very well other than Google. Because Google purchased Blogger, they know the ins and outs of a blog's architecture and can tweak their algorithm accordingly. So in order to determine if a site is a blog we need to look for the common elements of a blog that allow the engines to flag it as a blog and not a typical website.
1. Dates, dates and more dates
Blogs typically have dates all over the place, no matter what blog publishing app you're using, you'll have dates.
* Dates are found in the urls... "http://www.marketingshift.com/archives/2004_05_02_index.cfm"
Notice the "2004_05_02_index.cfm" (05/02/2004 is the date of this post)
* The homepage typically has paragraphs sorted in reverese chronological order and after each paragraph there is usually a link pointing to a url that has a date in the url.
For example: The close of each paragraph on my blog has "Posted by: Jason Dowdell 9:13 AM | Permanent Link" The permanent link points to a url containing a date in the file name like the example I mentioned above.
Different apps use different terminology to refer to a url where the post can be found.
Blogger - Permanent Link
MovableType - Permalink
2. Links In vs. Links Out
On a typical website's homepage there are no more than 5 links pointing out to another site. On a blog there are an average of 92 links pointing to external sites. This is based on the link analysis of the top 100 blogs listed by Technorati. So Google might look for a higher than avg. number of outbound links from a site.
Blogs have chunks of text in which there can often be found links pointing to external sites. Most e-commerce sites [the exception being spam web sites] don't link to external sites.
BlogRolls - common to have several outbound links in list format on a blog.
3. Blogs are text heavy
Most sites, especially commercial sites, are image heavy and text light. Blogs on the contrary are mostly text because they're all about information. Hey wait a minute, oh yeah, that's what the internet was invented for. totally forgot! doh!
Blogs always have archives, and ways to retrieve every piece of text that was previously listed on the homepage. Now if you think about your typical ecommerce website, they too have content that periodically is listed on the homepage. We call those "Sales" or "Featured Specials" and can usually be found a month later on a subpage. What makes a blog unique in this respect is the fact that when a blogging application archives it's data, it's usually text heavy and image light. Plus it's on a page by itself whereas a featured item is usually on a page with many other formerly featured items (we call this the sale page) and then disappears. A blog's archive lives on forever (or as long as the server lives).
5. Index Page
Last Update Date is more often than not no more than 2 weeks from the current date.
I think that's enough defining the architecture of a blog for the time being. Now let's move on to how we can determine whether or not the links from a blog count as much as the links from a normal website in Google's eyes.
First off, I'd like to point out that there are more than one kind of link on a blog. You're like, "what the...?" and I'm like "I did too just say that!" Yes, I understand that an a href is an a href is an a href but no two a href's are the same. Take for instance a link sitting in the middle of a paragraph on a blog. It's here today and gone tomorrow when the post is archived. It's a flash in the pan or 15 minutes of fame but it's definitely not going to stay around forever.
Then consider the links usually found up and down the right and left hand columns of a blog. Those links my friend are different because they stay around much longer than the Andy Warhol links in the blog entry paragraphs. So it would make sense that a link appearing on a blogs homepage month after month would be considered the same as a link appearing on any other site month after month for purposes of PageRank. Now if that is the case then everything is relative because those permanent links will be counted in the same way as permanent links on any other site. Since Google typically only updates the PageRank associated with a given url once a month at best, it's safe to assume the Andy Warhol backlinks won't show up in your backlink check for more than a month at best. But if you have a permanent link from a blog then it will show up for good, as long as you haven't pissed off the owner of the blog and they decide to take you off of their "Friends" or "Tech Patriarchs" link section.
To cement this theory of Andy Warhol links and their 15 seconds of fame vs. permanent links I'd like to go to the digital trenches and dig me up some dirty old data.
Well, well, well, what do ya know? Looks like my theory is partially correct. In searching through the backlinks of sites I give long term links to from the MarketingShift.com homepage and all subpages, I found that the links from the MarketingShift.com homepage weren't picked up as backlinks for those sites in my report. Instead I saw something much more bizarre. Even though every page on my site links to these external sites, only a handful of those backlinks were recognized as links. A possible explanation for this is since every page on my site links to these other sites, they could be seen as link spam and most have been disregarded. To test the theory a little further I decided to run the same type of analysis on a few other well known blogs and see what happened.
I'm going to use some of the same top 100 blogs as rated by Technorati for this data sample as well. Instead of running the analysis on the top 100 blogs I decided to run it on the top 10 blogs and ensure accuracy of the data. Because these top 10 blogs average 30,888 backlinks (on Google), the task is quite daunting. I decided to pull out slashdot.org since it's an outlier with 157,000 backlinks and that brought the average number of backlinks down to 17,720 backlinks.
Avg Backlinks: 17,720
Avg Links to Internal URLs: 107
Avg Links to External URLs: 120
Avg Links on HomePage: 227
From this data we can already see that the number of links to internal urls is approximately 89% of the number of links to internal urls. This would suggest that blogs have more external links on their homepages than internal links but I say this sample size is too small and should be reevaluated with a sample of at least 1,000 blogs or more.
Back to our search through the backlinks of the sites each of the top 10 blogs keeps in it's list of permanent links or friends or whatever you want to call the links that stay on the homepage for more than two months at a time.
Yes I know I cut the post off in mid-sentence but that's where I stopped working on it last July. If you would like me to revisit the blog backlinks research then post your comments here and I'll pick up the ball where I left off.
Blogs Are Not WebSites - Breaking Down Backlinks & Structure By Jason Dowdell at 02:01 PM