Starting Fresh: Rethinking PPC Ad Copy Testing

If you are like us, then you can appreciate how useful AdWords can be to determine the right message to distribute to your customers. Ad copy testing is an important component of optimizing your campaigns and should be done regularly. The question is, what is the best way to do it?

In this article, we’ll be going through our preferred method as well as an additional method recommended by Google. We will spend time speaking to the pros and cons for each but plan to put this test into action and write a follow-up article on results for next quarter.

Our Preferred Approach (Impression Split Testing):

The most important thing for us, when it comes to ad copy testing, is to ensure each ad tested has received an even share of impressions. This is done by setting your AdWords campaign to ‘rotate indefinitely.’ This gives each ad the same amount of opportunity to entice the user to visit your site and convert.

The issue with this method is that sometimes one ad can throw off the impression split if its relevancy is much higher than the rest. This happens when a group of queries matches up better with one ad variant, then it does with another. If you refer to the table below, you can find an example of an ad test with uneven impression distribution.

If you were to make your decision purely based on CTR and Conversion Rate, you might think that Test 3 is the best ad. However, if you look at the impression difference between Test 1 and Test 3, you’ll see that Test 1 received almost 5x more impressions! This happens typically when one ad has a much stronger Quality Score relationship with the keyword than the others.

Some would argue that the most relevant ad, independent of ad-level metrics, should be the winner here. We would likely relaunch this test once more to determine if the results were merely a fluke.

The Other Approach (Optimize for Conversions):

One other method worth considering is using AdWords’ optimize for conversions setting. With this option, you can leverage Google’s algorithm that will automatically serve the ad that’s been deemed most relevant in most of the auctions. The benefit to this is that this setting will optimize your ads for clicks in each individual auction using signals like keyword, search term, device, location and more.

The downside to moving forward with this method as your primary ad testing setting is that you are trusting that Google’s algorithm will provide you the best results. You won’t be able to distribute an even share of impressions. This isn’t all bad since the goal is to achieve the strong results for the ad group.

Next Steps:

Being the skeptics that we are, we’ve decided to take our Google rep’s recommendation and put this method to the test. Our rep shared this article, and in it, the writer discusses a suggested methodology for testing using optimize for conversions. The test they suggest is as follows:

“Test an ad group with one ad (A) against an experiment ad group with four ads (A, B, C, and D) with rotation set to optimized. You can use drafts and experiments to create these two versions. That way, you’re testing to see whether or not more ads result in more impressions and clicks at the ad group level.”

For our next article, we’ll be putting this method to the test. The idea will be to determine if the optimize for conversions method yields more favorable results than our standard method. If you were interested in learning more about you should rethink PPC ad copy testing, please contact us by email at  or by phone at 781-591-0752.