Hi all, I decided to make a little spreadsheet to play with the dimensions of the AdSense ad formats. I entered the width and height for every available format of text/image ad unit, and then calculated the area and aspect ratio for each. The results are shown below, sorted descending by greatest advert area: [B]Name Width Height Area Aspect Ratio[/B] Wide Skyscraper 160 600 96000 3.75 Large Rectangle 336 280 94080 1.20 Medium Rectangle 300 250 75000 1.20 Skyscraper 120 600 72000 5.00 Leaderboard 728 90 65520 8.09 Square 250 250 62500 1.00 Small Square 200 200 40000 1.00 Vertical Banner 120 240 28800 2.00 Banner 468 60 28080 7.80 Small Rectangle 180 150 27000 1.20 Button 125 125 15625 1.00 Half Banner 234 60 14040 3.90 Code (markup): Note that Google say that the three best-performing ad units are the Large Rectangle, Medium Rectangle, and Wide Skyscraper. Well, these are precisely the top three in this list. So, it looks as though ads that cover a large area on the page get the most clicks. This is, of course, intuitively sensible. However, I wonder if the trend continues down the list; this would make Half Banner the worst format available. Any thoughts would be appreciated.
CTR is highly dependent on your niche. For example, I have one website where I have only a banner (468x60) that is not even above the fold. It's only at the end of each article and you have to scroll to see it. CTR is always around 5% - 10%. The first monetization method that is above the fold is an affiliate offer and many visitors end up completing the action which brings me commission. For the rest is adsense below. Together they bring me a very good conversion rate. On another site I have two large rectangles above the fold. The website is of quality nature and I have a pretty low bounce rate. CTR is 2%. Niche is everything.
Well, I'm just glad that someone finally found this post of mine and commented on it. I've not personally got enough data yet to comment on the relative successes of the different ad formats; I don't think that I can draw any statistically significant conclusions from my current rate of 1 click per month...