I've seen a lot of very impressive research being done on the subject and many erudite papers being published. All of this is backed by a great deal of experimentation which involves complex equipment and analysis. Hence, when I saw the product being offered by Eyetrack I thought that it represented a fun, accessible way that everyone can answer the question about their own sites.
And, it's not the same for each type of social networking site. Take, for example, LinkedIn (below) and compare it with Facebook (above):
It would appear that visitors do actually read the text on LinkedIn, but very much at the "headline" level. All those extra bits of infromation added by LinkedIn to the right hand pane about who esle was viewed, etc seem to go to waste.
And then, of course, there's Twitter:
Are viewers reading the first one or two words of a Tweet only? That's all the time you have to capture their attention! But compare Twitter with Facebook: it does appear to be text not photos being scanned in Twitter.
One very interesting area that is worth studying is the deviation in viewing patterns caused by the client that is being used to view the page. With so many apps now available for viewing each of the social networking sites, it might be interesting to see if the viewing pattern is in any way governed by the way it is presented. Twitter in particular has a large variety of clients, so how does that affect the viewing:
So, thanks to Eyetrack for enabling us all to do a bit of research that is both fun and produces useful results.