I recently threw some random ad selecting code into a site I had and let it sit there until 100,000 visitors to each setup. Only one ad was shown with each pageview. The site's nothing amazing, but here's how the ads performed in comparison to each other: [800x600 box around content, blended inside] Skyscraper ad on the left was almost twice as good as a large rectangle on the top of the right side Right side rectangle worked around 20% better than a leaderboard on the top I thought the leaderboard'd do a little better myself
And your telling us why. every website and niche works differently. What may have worked for you may not work for someone else. The thing is experiment thats what you have done and it worked good for you.
Thanks for sharing your results. I'm sure everyone has different experiences, but the more knowledge the better!