Google's Top 10 Choices for Search Results.
Posted Thursday, October 13th, 2005. Filed under Google, SEO Research.
A really late night post. When you run a search at Google, the top 10 results commonly follow a pattern. The average age of sites in the top 10 is around 6 years old, and a site newer than 2 years is very rare — your chances of seeing a site less than a year old in the top 10 for any commonly searched phrase are about 1 in 1000.
Typical top 10 slots include: a college paper (.edu) result, a government paper (.gov) result, a site with lots of .edu's and .gov's linking to it (Google's best guess at highest trust rank), a subpage of a huge site like Amazon or a shopping directory, an old page that hasn't changed in years with slow steady link growth, pages from historically top-ranking sites that update regularly with fresh content, new pages from old established sites that fade in and out of rankings, subpages that just got a bunch of backlinks coinciding with new content, an old authority site that keeps updating and getting new links, and occasionally a spam site that won't be around next month.
I think Google does this on purpose to show a variety of page types to the user. If you're targeting a phrase, start by figuring out what type of result your site will be, who your real direct competitor is, and what it will take to replace them.
October 15th, 2005 at 1:55 am
[...] Jim Boykin hypothesizes on Google’s Top 10 Choices for Search Results. [...]
October 15th, 2005 at 11:37 pm
Top 10 Google’s Favorite in Search Results
Jim Boykin has an excellent post about what he thinks Google’s top 10 choices for search results. Overall, the average age of the sites in the top will be 6 years old, and a site newer than 2 years old is very rare.
Your chances of seeing a sit…
October 17th, 2005 at 12:32 am
The Google SERP Party
This is the party-version of Jim Boykin’s Google’s Top 10 Choices for Search Results post.
Imagine you’re in a small town. Imagine throwing a party. Imagine coming up with a list of people to invite to that party. Now imagine the first page of …
October 18th, 2005 at 11:35 am
[...] I’d look at what Google might be trying to achieve in the top 10 results, and that will give you some clues as to what Google might be looking for, and just how little tweeks in their algorthym can cause such havoc in rankings. I’d also follow it by looking at link networks, the way Google might strive towards. [...]
October 18th, 2005 at 11:38 am
[...] Jim Boykin provides rather interesting insight as to what type of sites we are likely to find on the first page of Google. I don’t buy it hook, line, and sinker… only because we’ve gotten sites ranked into top 3 positions that don’t fit any of that criteria, but it does give some food for thought. [...]
November 11th, 2005 at 7:33 pm
[...] Google top 10 choices for Search Results - Oct 10 [...]
November 18th, 2005 at 10:32 am
Yeah, I think about Gooogle and what most people use it for, too.
(Sometimes I think it’s just a bunch of bullshit.)
But, you know, whatever.
Thanks, you’ve been great.
December 1st, 2005 at 4:08 pm
[...] I’m aware that things are constantly changing, and I’m aware of what Google likes today. I think the SERPS aren’t a mess at all. As engines get smarter and are better able to analyze things, SEO’s must get smarter as well. Overall Jagger was pretty good for most of our clients - probably due to our whiter hat methods. I believe that the engines are “rewarding” sites that appear to look like “natural resources” and so as an SEO, I’m trying to emulate this look on our sites as best we can. (via it being telling our clients to create better resources, and via trying to link more “natural”) [...]
April 13th, 2006 at 5:11 pm
If the trust factor becomes more and more important, how can non-english sites get trusted links from edu and gov sites?
June 13th, 2007 at 7:22 am
I’m just thinking, how in the world do you get a link from a trusted site, i.e. edu or gov?
October 21st, 2007 at 1:16 pm
If you clearly identify all of the possible word combinations, acronyms, semantic word stems and mix up the chronology when building links, you can essentially dominate thousands of searches from gaining traction from the long tail of search.
Add specific exact match link building to get a specific leg up in the SERPs and presto, achieving top 10 rankings become as common as sending an email. Take for example our website, it is only 1 year old, and we have over 4000 unique search compatible keywords. While most struggle to get out of the sandbox, we were ranking in multiple top 10 results, all as a result of authority and content development and using a diverse base of Class C-IP’s to get the juice.
Once you got it, everything you publish goes straight to the top.
October 29th, 2007 at 4:47 pm
I just followed today’s post back to this 2-year old post. You really think this is still true? I am not sure I believe the college paper one any more. And thankfully, it seems spam sites are turning up less and less.
November 3rd, 2007 at 7:26 am
I didn’t realize that age was such a factor. This is also helpful because I wasn’t sure if they liked .edu’s or .gov’s better. Is a page that is frequently updated better than one that is stagnant. I know that sites that are updated regularly are better than ones that aren’t but does that also apply to individual pages?
November 9th, 2007 at 11:25 am
Think about it. If it were up to you to write the algorithm that weeded out junk, would it not be logical to look at domain age, who is info, .edu , .gov links and other similar characteristics at least in part for determining rank. Over time, cheap little tactical approaches will be stifled… and quality links will hold more and more weight.
For thoughts on how to get links from .edu - try your college alumni site, they often post profiles of alumni and would be glad to give you a link. How about designing a few free websites for some local schools and government agencies and asking for a link in return - very altruistic and “not evil”, which should be supported long term by the mother ship.
February 2nd, 2008 at 3:01 pm
Agreed… .edu and .gov links don’t necessary imply that a site should be considered more important. Good analysis though.