For the longest time, PPC analysts have been promoting testing ad copy within AdWords campaigns to identify the best performing ad copy in the account. The traditionally recommended way to achieve this is through Google’s “rotate evenly” ad delivery option.
However, setting ads to rotate evenly does not mean that ads will serve more evenly during the first 90 days of serving; it actually means that ads will enter the auction more evenly during the first 90 days. Because ads are not guaranteed to win every auction entered, balanced ad distribution does not always occur. How the ads actually serve will be influenced by a variety of factors including quality score, max bid, etc.
What Does This Mean for Your Ad?
Both winning the auction and being seen by users are necessary for the served percentage of an ad to increase in AdWords reporting. For example, if one ad variation shows up on the first page of search results and users see it, the ad's served percentage, reported in AdWords, will increase.
However, if a different variation happens to serve, but on page 2 or 3 of search results, and users don't see it, then the served percentage for that ad will not increase. This results in uneven (often by a lot) served percentages in AdWords reporting.
You may not compare ad effectiveness unless your ad copies are up for auction at a similar, statistically-significant number.
Testing Ad Copy
To fairly evaluate ad copy, decide whether to optimize for either conversions or clicks. Even though this removes from advertisers the opportunity to distribute the different ads evenly, it will optimize towards the specified goal. Both of these options are only available if the campaign settings are set to “all features.”
Keep in mind that the ad with the highest CTR (click-through rate) may not always be the one expected to get the most clicks. That’s because the number of clicks an ad is expected to get is also affected by how often an ad appears – that is, how often it is eligible to enter an auction. So a more relevant ad with a better landing page experience but a lower expected CTR could be eligible to show on search results more often. That can, in turn, lead to a higher overall number of expected clicks. Because the CTR takes into account the impressions made, it will be a more effective measure of success than simply clicks.
Which Setting Has the Highest Performance?
Recent studies conducted by Google have shown that using the Optimize for Conversions Ad Rotation setting leads to a 5% increase in conversions and using the Optimize for Clicks setting (for accounts without conversions) leads to a 5% increase in CTR.
Below is quick summary of when each setting is best:
|How it works||Who should use it|
|Optimize for Conversions||Prefers ads that are expected to get the most conversions||You have Conversion Tracking enabled and care about increasing conversions|
|Optimize for Clicks||Prefers ads that are expected to get the most clicks, based on past click through rates (CTRs)||You do not have Conversion Tracking enabled, but want to increase traffic to your site|
There are many different ways to approach ad testing in AdWords, which all depend on your end goal. Select your goal in advance. Once these settings are in place, you will be able to identify which ad copy achieves the specified goal better and then create new ad copy to reflect the strongest ad copy. Test ahead!
Matthew Gangnier, Paid and Social Media Analyst at iProspect, also contributed to this blog post.