As we observed in Poynter's new
Eyetrack III study, in general the larger the banner ad on a news website, the more eye fixations it gets, and the longer duration of viewing. (A fixation is a very short pause of the eye. A normal viewing of an ad, say, would include multiple fixations and "saccades," which are the paths between fixations.)
Well, that seems obvious. But there's more to it.
For "half-page" ads (368 x 850 pixels) placed on article pages, they performed best in terms of number of fixations by people who looked at the ads at all; people looked at them longer. However, those ads were only seen at all in 38 percent of our test subjects' visits to pages containing the half-page ads.
The article-page ad type that got seen by the most people -- 56 percent of our test group -- was a 300 x 250 pixel ad inset into article text. It didn't get as many average fixations as the half-page ad (4.6), but was pretty close and was seen by more people. If I was an advertiser, I know where I'd want my ads placed on article-level pages.
There are more advertising observations from Eyetrack III
here.
Poynteronline did to me exactly what Dvorak accused newspapers of...