The latest Alertbox is on banner blindness (with a cool heatmap still and video). But the point that I want to blog about is his point on making ads look like content to unethically get clicks.
He says that these ads "work" because users look at them as if they are content and click on them because they resemble content. But how is he defining "work"? If the user realizes it is a fake and closes the window as it is loading (or just after), then the ad did not "work." The user probably didn't even see what was in the window. And if she did, she probably has a lower opinion of the company and its products. I hardly think of that as "working."