Comments on: Google 2009: The Painful Details

By: Thomas Eilander

Thomas Eilander — Fri, 16 Jan 2009 13:38:55 +0000

I wonder how long it will take till Google buys stumbleupon to use it for their rankings :o

By: Kian Ann

Kian Ann — Thu, 01 Jan 2009 23:09:16 +0000

@Ian – if you were to build your own search engine, how would you measure relevance of search results?
Like what you said, anything can be gamed. Even if bounce rates and time on site is measured together, there is cheap manpower available, and one might just spend some money to get a bunch of peeps to load do what the SE thinks is a good user behavior.
I think until the day we can really track eyeballs, I guess google has to rely on such measures.
P/S: Its kinda scary to think how people might game the SEs when we can really track eyeballs. Heh.

By: Ian

Ian — Wed, 31 Dec 2008 18:05:58 +0000

@Tom you’re spot-on on site A. Site C isn’t Conversation Marketing, though.

By: Tom

Tom — Wed, 31 Dec 2008 16:08:20 +0000

I am going to guess that site C is this one. And site A is your php hybrid milage comparison site that you mentioned once before. Site B no idea. Sorry about skewing your experimentation, I won’t be offended in the least if you delete/not post this comment.
All of the google updates sound exciting to me, but I am not really in the SEO business. I really like to see giant companies that change major parts of their business and not become in-the-rut dinosaurs.

By: Ian

Ian — Wed, 31 Dec 2008 07:16:46 +0000

@Erik I agree, but remember there are many, many other factors they’re weighing too. I think the real intent here is to add more, better measures of site relevance and authority. That makes things the algorithm more complex and harder to game. As I said in my first post on the topic, anything can be gamed. ANYTHING.

By: Ian

Ian — Wed, 31 Dec 2008 07:10:46 +0000

@Leo Definitely reserve for the long haul.

By: Erik

Erik — Wed, 31 Dec 2008 03:53:38 +0000

This is a very interesting piece of information. Many of us already had the feeling that social behaviour is of importance to your ranking.
But there is much more research required to know the social drivers of your ranking.
You mention bounce rate and time on site. But both are easy to manipulate and therewith less attractive to Google.
Bounce rate
Bounce rate measures whether a person opens a second page. If somebody quickly opens three pages on your site and then leaves he may not have found what he was looking for. Bounce rate is 0% however.
If a person opens the first page, finds what he was looking for and starts reading for the next two minutes bounce rate is 100% but the visitor is happy.
The problem above can be solved by calling the Google Analytics setvar with javascript after some time, for example 60 seconds. Everybody paying attention to a page long enough will not be counted as a bounce.
Time on site
The other measure you mention is time on site. Although this is in my opinion very valuable, I suspect it only measures correctly if you browse with one tab.
I use a dual screen setup with several browserwindows who each have several tabs. I open interesting pages in a new tab and start reading them in one go. This means that a page can be open for a long time without me looking at the site. Can Google measure which tab you are reading??
Basically, there is a lot of clouding here. Nevertheless, it is worth your time to look at these measures to try and improve your ranking by improving user experience. IN the end the golden key is user experience.

By: Leo

Leo — Tue, 30 Dec 2008 22:30:22 +0000

I know this may sound a little off topic but you mentioned domain age as a weight for trust benchmarks. What is your view on purchasing a domain long term vs. re-upping every year, at least in the eyes of google? Do you think that purchasing your domain for 10 years would help increase rankings (even marginally) b/c it shows that you are in it for the long haul?