Properly Split Test Your Ads on AdWords
The purpose of split testing on AdWords is to always out perform your previous best ad. Doing so will lead you towards higher CTRs and lower bid costs. Now, the question becomes, how do you properly split test your ads?
Follow the process below and make sure that you are comparing apples to apples and not apples to oranges.
Once you have determined the winner from your 2 original ads (A and B), in this case A; what you want to do is then create 2 new ads, C and D.
Where ad C is the exact duplicate of the winning ad. And D is the new test ad that will be tested against your control ad (Ad C).
To start your new split test, just pause ads A and B and activate ads C and D.
You want to do it this way because, true split testing, means you are testing one variable that is different at a time e.g. Headline, Description Line 1 or 2, Display URL.
But you should also be aware that the time period is also a variable. By split testing ad D with A, the comparison would be incorrect. You need to reset the ads, by starting them all over again.
I know you can overcome the time period variable by filtering the results on Google AdWords by the respective date range. But I just find my way a lot cleaner.
Here is something that I do that you may want to adopt. You can delete ads A and B once you know for sure that ads C and D are running fine. But I just keep them paused. I do this for 2 reasons:
- To keep a history of my ads online
- To keep the ads visible on the offline AdWords Editor. When I edit ads I like to see a history of previous ads. If I delete the ads, they will not show up on AdWords Editor. Pausing is just a quick workaround for me.