Ad copy testing can be an insightful, yet difficult task to master. It can take time to see substantial enough results, and you could potentially be wasting money if an ad is performing poorly. But, when done correctly, ad copy testing can be a pivotal element in improving search campaigns, leading to reduced CPCs, higher CTRs and other major improvements to a campaign’s bottom line.
Marketers have their own opinion as to which types of ad copy perform better, but ad copy testing lets you show numbers to back the decisions.
Where to Begin
Start with a goal. Always have a goal in mind when running a test. Whether it be a higher CTR, or lower CPCs, it is always beneficial to know what data you are looking to obtain.
How Many Ads Should I Test?
I recommend testing only 2 ads at a time. The more ads that you include, the more variables that you will have to take into consideration when choosing a winner. Also, you will be able to reach statistical significance much sooner than if you were testing multiple ads at once.
What Settings Should I Use?
Make sure that you are utilizing the proper Ad Rotation settings.
Click the Settings tab and select the campaign that you are performing the test on. Next, under Ad delivery: Ad rotation, frequency capping, choose Rotate indefinitely, as the Ad Rotation option.
This will deliver the ads evenly, for an indefinite amount of time. It will allow the higher-performing ads to run as frequently as the lower-performing ads, until you decide to pause the ads or adjust the settings.
This method is preferred over Rotate evenly, because some ad copy tests will last longer than 90 days, and you do not want the options to automatically change without your knowledge.
Please note, that you should be continually checking the ads. You could be losing money if the lower-performing ad is not receiving any conversions, but is spending as much money as the higher-converting ad. While this setting is necessary for the test to be accurate, you could find out early on that money is being wasted.
What Should I Be Testing?
Ideally, you should only test one element at a time. Testing only one element at a time will allow you to isolate exactly what consumers are reacting positively/negatively to. This way, you will know which element led to an increase in clicks/conversions.
The two ads that you are testing should be identical, except for the one specific element that is being tested. If everything else is consistent between the two ads, you will be able to know exactly what is triggering the change in performance.
I recommend testing the following elements:
- Order today
- Buy now
- Shop now
- Landing page
- Specific product page
- Category page
- Display URL
- Value proposition
- Free Shipping
- Coupon Code
- 5% Off
- Dynamic Keyword Insertion
- Stagnant headline
- Description Line 1
- Description Line 2
Introducing Ad Customizers
Ad customizers are a great tool to increase ad copy relevance. Ad customizers are similar to the dynamic keyword insertion feature, in the sense that they can insert custom data, based off of numerous factors. While data proves that they are a valuable addition to any text ad, they can be another dimension that you can implement when testing ad copy.
One major benefit of implementing ad customizers is the fact that you can use more than one customization per ad. They also take into account user data, such as location and time zone. Ad customizers are an easy way to dynamically change ad text, while using real-time consumer data. Having this real-time consumer data allows for the use of additional customizers such as the countdown function, as well as location-specific text.
The countdown function is a great way to create a text ad that is both relevant and urgent. You can run one ad with static text, and another with the countdown function. This function can countdown to a specific date or time of day that a promotion will end. The settings can be set to the searchers time zone; this way the feature will be relevant across multiple time zones. When a consumer searches for your product or brand, an ad can be displayed telling them that the promotion ends in 3 hours, and can continue to count down until the promotion ends.
When you create ads with this countdown function, you will specify exactly what date and time the promotion ends. Once the promotion has ended, you can compare the results of the countdown function, in relation to the standard text ad without the ad customizer.
Another benefit of using ad customizers, is that when you make changes to the ad (price or end date) it will still preserve all of the historical data. This can help with ad copy testing if you are running the test over a long period of time, with multiple end dates.
How Much Data Do I Need?
There is no set number needed to reach statistical significance, but the more data, the better. Some search campaigns will run for a month or more to collect data, while other high-volume campaigns can collect enough data in a week.
The longer that you let the test run, the more accurate and abundant your data will be. This way, you will be able to eliminate the possibility of certain anomalies skewing your data.
How Do I Choose a Winner?
It is best to use statistical significance to make these decisions. Statistical significance shows the evidence needed to prove that one ad performs better than another, and not just by chance.
There are many online tools that can be used to help you calculate and determine if the data is significant enough to make a decision. All you have to do is enter the data (clicks, impressions, etc.), and the tool will do the calculations for you! It’s as easy as that.
The A/B Significance Test is an example of an online tool that will use your data to tell you whether your A/B test is statistically significant or not.
Who Does Ad Copy Testing Benefit?
Ad copy testing is beneficial to the success of any type of campaign, whether e-commerce or lead-gen. It’s a valuable way to gain insights as to what type of messaging is helping you to reach your overall goals.
As previously mentioned, ad customizers will help a text ad stand out by creating a sense of relevancy and urgency. I have seen great success using these in both the lead-gen and e-commerce space, but I feel that they make the most impact on e-commerce campaigns.
One challenge that ad customizers have solved for marketers, is the ability to dynamically change ad text, without the loss of historical data. This is especially relevant with e-commerce campaigns, where companies have a variety of daily and weekly promotions.
With the implementation of these ad customizers, you will be able to dynamically change the countdown to a promotion, the promotion specifications and more, while maintaining all historical data. Being able to preserve this historical data will help to ensure that your ad copy test is as accurate as possible.
Now that you’ve declared a winner, it’s time to start the next ad copy test!
After you gain a winner from the first test, you can use that information to test a new element. You want to constantly be running ad copy tests to find out what works best for your campaigns. There are so many different elements that can be tested, you should never stop after one.
- Follow the basic rules of ad copy- put your keywords in the headline and description. This will help to improve your quality score by making your ad more relevant.
- You want to have a fresh start- pause all existing ads, and introduce two (or more) brand new ads. This way you will have completely un-biased data.
- Don’t touch the ads- let the ads run untouched until they accumulate enough data to determine statistical significance.
- Check the test every few days- if there are a high number of impressions early on, let the test continue to run if possible. Longer testing will ensure that there were no anomalies influencing the data.
- Never stop testing- just because one test ended, doesn’t mean you’re finished. You can continue to optimize your ads, one test at a time!
Ad copy testing is an important investment in the future of your search campaign performance. It will take some time to see results, so give the test some time to run. After the test collects enough data, analyze the data that you find and then optimize from there. Once you are able to pinpoint exactly what is improving performance, you can begin your next set of testing.