It's been written about quite extensively. Google stuff like "eye tracking study conversion optimization" and you'll come up with a ton of hits. Lesson #6 on th

It's been written about quite extensively. Google stuff like "eye tracking study conversion optimization" and you'll come up with a ton of hits. Lesson #6 on this[-2] blog post also has another example. A/B test everything, and, while this isn't a hard rule, generally the model's eyes should be looking where you want to viewer to look. The reason I think 37Signals successfully breaks this "rule" is because the headline is by far the most dominant thing on the page.[-1] Not so with OP's design.[0] On the Basecamp site, it's virtually impossible not to be able to see the headline first and read what it says before moving on to anything else on the page. 37Signals would be wrong to have the model's eyes looking in any other direction. If she were looking to the left she'd be looking away from the call to action. If she were looking to the right she'd be looking away from the social proof. This is one of those times that they understood the "rules" so well that they were able to break them for greater effect. [-2] http://www.quicksprout.com/2013/08/01/7-conversion-optimization-lessons-learned-from-eye-tracking/ http://www.quicksprout.com/2013/08/01/7-conversion-optimizat... [-1] http://i.imgur.com/0XWFn9v.png http://i.imgur.com/0XWFn9v.png [0] http://i.imgur.com/HAW2JTv.png http://i.imgur.com/HAW2JTv.png