A/B Split Testing for Your Google AdSense Ads
October 29, 2010
Posted in AdSense & Contextual Advertising
Leave a Comment
Google AdSense & Contextual Advertising
While quick and dirty testing of your Google AdSense ads can give you some answers over a long stretch of time, A/B testing can give you answers in a much shorter period. How? By testing two different formats, colors or placements simultaneously.
How To Do A/B Split Testing
Google has provided some code that will allow you to make a change to the ad you're already running, letting you run two ads, each one appearing 50% of the time. This will give you more accurate data than making changes one after the other.
There are a few things to remember when doing A/B split testing...
Make only one change at a time. Change only the format of the ad, or only the color of the ad, etc. Never make more than one change. If you do, you'll never know if the difference in numbers is due to one change over the other, or a combination of the two.
One of the two ads you test must be the one you are already running. You need a baseline (or control) from which to measure results. If you've had a 10% Click-Through Rate (CTR) for the past several months, it's safe to assume that the same ad will receive the same CTR during the test. This will allow you to compare the new ad's CTR with the baseline number.
If, however, you start with two new ads, you will have no baseline from which to compare one ad to the other.
The tested ads must run concurrently, to avoid seasonal and other fluctuations in CTRs.
The scripting that Google has provided in their AdSense blog, along with the use of Channels, takes care of the third point. You will have to manage the first two.
Depending on the number of ad impressions you normally receive, log in to AdSense one or two weeks after you started your experiment and check the number of impressions for each channel.
If there's a large discrepancy between the two ads, you have two options...
Use the best performing ad only from now on. Replace the A/B testing code with the code for the best ad. Start earning more income!
Test the best performing ad against another change.
A/B testing can be done on any number of pages on your site. This could be particularly useful on pages that have high exit numbers and no other monetization options.
Use A/B testing to see if you can reduce the exit numbers. Since those people were going to leave your site anyway, you may pick up some bonus income from them.
The cost to you? Perhaps 30 minutes of your time.
The benefit? Some AdSense formats, colors and locations have reported CTR improvements of up to 100 times higher than that of other ads.
But the only way to know which ads perform 100 times better is to test, test, test. This A/B testing code makes it so simple, you'll be earning more AdSense income in a matter of days.
For more information about Google's A/B testing suggestions, see...
and this AdSense Help Center article about it at...
For additional information, review related articles in the Google AdSense series.