An interesting article looking at 'participant-free eye tracking' services. Conclusion here is that they seem to be pretty worthless.
Aga Bojko conducted an eye tracking study of the eBay homepage with 21 real participants and then compared the results with those from 3 automated services which claim to simulate human behaviour.
Eye Tracking Without Eyes - article by Aga Bojko
Of course, the things that people focus on when visiting a page will depend upon what it is they're trying to do and whether they've visited the site before. So whatever you try to infer from an eye tracking heat map without an understanding of the participants' motivation you're not going to get very far.
I recently posted a link to an article which looked at several eyetracking studies of the same image, where each time participants were posed a different question. It clearly illustrates the impact of user motivation on eyetracking behaviour.
Previous post: User test instructions affect behaviour (February 2011)