Ugly or Pretty Website Experiment

Which makes more adsense?

[Sharper]

First, I'll admit that most of my sites aren't all that "pretty" in the sense that an art major would like. They tend to be standards compliant with lots of text, formatted using CSS in order to make the text more understandable and readable (as opposed to prettier).

My most popular site (in terms of visitors) is of the type where people find something on a search result and go directly to a content page. Visitors typically view 1-3 pages on average. This site normally accounts for about 60% of my Adsense impressions. I had to use this site to ensure that any changes in behavior were big enough to show up in the Adsense stats that cover all my sites.

I took that site and made it ugly. Since all the formatting for the site is in one CSS file, it was easy to do on a temporary basis.

I tried this experiment for 1 hour, checking updated Adsense and web server stats regularly before and up to 30 minutes after and recording the results to find the stats that corresponded to that hour of the reporting delay. During this period stats appeared to update about every 15 minutes and were 15-30 minutes behind.

How ugly? REAL UGLY. As far as I can tell from my site logs, not a single visitor went to more than the page they landed on. It was so ugly that there were probably some people scratching their head and wondering if the site was gonna finish loading or not, even though their browser said it was done. The only "obvious" thing to look at or click on that made any sense was the Adsense skyscraper.

For the 6 hours before this "change" (so, same day, same ad mix), let's say my overall CTR was (relative, not actual) 100 and my CPC was also 100. This is also a normal level within my usual variations. From what I've read of other's posting their usual CTR, mine is normally slightly higher than other's averages.

For the hour with the site uglified (impressions in the thousands, so should be a statistically relevent time period), the relative CTR was 235, but the relative CPC was 44.

This led to relative hourly revenue of 100 for normal and 101 for the ugly hour.

My conclusion is that while having an uglier site that people want to click off of can in fact raise your CTR significantly, it doesn't always make you much more money from Adsense.

I didn't try the experiment using a two-ad layout instead of the skyscraper. Doing so may have kept the CPC higher and managed to raise revenue. That's something I may play with at a future date just to see, if no one else does. Obviously, I prefer not to trash any site's appearance very often just in order to experiment with Adsense. A return visitor would probably have been convinced that for an hour that site was having server or code problems or something.

Anyone else have any data that relates to my experiment?

[killroy]

Hmm, so if I understand your findings correctly, people clicked more often in the ugly version, but usually clicked further down the list?

Hmm that on first thought doesn't make much sense. On second thought though, I'Ve found, both in AdSense as well as Adwords, google's greedy ranking algo causes the top spots be used by the highest bidders, though less relevant ads, while further down come the more relevant ads.

So, I specualte your ugly site gave them more reason to look at the ads in detail and then choose the most relevant, causeing them to click more often on the 3rd or 4th rather then the 1st or 2nd ad.

Jsut speculation, but I can hardly imagine it any other way.

Of course, to be conclusive you'd have to run it for one whole day, or even better, use a second account, and randomly pick the ugly stylesheet with one ad code and the other with the other code.

That way you could show the ugly styles only 10% of the time, causign you less loss of reputation and still ahve comparable results.

SN