The Scientific Process of Digital Ad Testing
How to Run an Effective Paid Search Ad Copy Test
We search marketers spend a lot of time on finding unique and creative ways to appeal to our customers. Ads are the gateway between customers and conversions, so understanding what helps them convert is incredibly important. Does creating strong ad copy assure good performance? How do I know when my copy is good? How about when it’s great? The answer my friends, is ad copy testing!
It’s All About the PROCESS!
When it comes to Ad Copy testing, it’s all about the process. Just like the scientific method, it’s about asking a question, running experiments, making observations, and adjusting your approach based on what you find. The first step is arguably the most important. You need to decide what to test!
Decide What to Test
Think about your product/service, how your customers use and search for it, and use this information to inform what component you want to test. Ask yourself the big questions first:
- Are consumers price sensitive? (Discounts vs No Discounts)
- Does Location matter? (Geo vs Non Geo)
- Does Shipping matter? (Free Shipping vs Rush Shipping)
- Does Selection matter? (Large vs Exclusive)
- Does Quality matter? (Cheap vs Premium)
- What is your best Call to Action? (Book a Visit vs Visit a Showhome)
- Should I use an Ad Customizer vs a Static line?
- How about DKI (Dynamic keyword Insertion) vs Non DKI?
Once you understand what questions you want to ask, it’s time to set up your test!
Set Up Your Test
The first step to setting up the actual test is to select the ad groups/ keywords to be used for testing. I would recommend starting with the area of your account that can deliver the most performance insight quickly. There are two types of use cases:
- Single Ad group Testing – This is best suited for high traffic and high spending ad groups. High spending ad groups are often the best suited to gain efficiencies through ad testing. However, this approach will only tell us how ad copy performs across one set of keywords, which may not be the whole story.
- Multi-Ad group Testing – This is best for small accounts where a single ad group does not have enough search volume to carry out a proper test and where you want to test the impact of your changes across multiple ad groups with the different set of keywords.
Now that you’ve selected your Ad Group approach, it’s time to set your goals!
Set Your Goal
Ad testing is pointless unless it leads to your desired outcomes. Do you want more clicks and traffic to your website? How about more conversions at a lower cost per conversion? Selecting the metric you want to test based on your campaign objectives is always the way to go! Here are a few examples to get you started:
Always make sure that your test metric is aligned with your campaign objectives as it will make the test more valuable and relevant to your overall goals. Once this is set, we can pin point the ad element that we would like to test.
Select Ad Variable to Test
It is always recommended to run single variable ad testing so you truly understand the impact of your changes. Simply put, when you run two ads testing only one variable you’ll know what element in the ad accounted for better or worse performance. I like to start with with the headline, as this is the first element a user will read when they see your ad in the search results. The headline should stand out and show exactly what you are offering. After I test and select the best possible headline, I move on to other elements, like the description line, the call to action, and the display URL (one by one). When running these tests, you’ll want to make sure that you’re adhering to a few testing principals and parameters:
- Ads included in testing are active for the same set of ad groups and are live over the same period.
- Device preferences for the ads are same.
- Ads are not edited during the test. Any changes to the ad between tests make the test result inaccurate!
- The ad rotation settings for the campaign is set to rotate evenly. The purpose is to expose an equal number of users to each variation.
Your ads are now adhering to best testing practices and your continuous observations have led to some great findings. What’s next?
Measuring Test Performance
Another important safety measure in a test is to set abandonment points—minimum and maximum data guidelines for the test. It is necessary to have statistically significant data to pick the ad test winners, and if your ad test doesn’t have enough data, then you should wait until it does. Do not pause ads or make any adjustments until you hit that statistical confidence level, because you will disrupt the methodology of the test and end up with inaccurate results. Once the minimum data is achieved with a sufficient confidence level, it’s time to take action. For best results, you should keep a confidence level of 95-99%.
Always Be Testing!
Despite being one of the most important forms of optimization, ad testing often gets overlooked by search marketers. Not all test results will lead to extreme impact on your account, but regardless of the outcome, it’s important to have a process around ongoing testing as you do often stumble upon efficiencies that would have otherwise gone unnoticed. Use this methodology, continually test your ads, and stay on top of your competition. Happy testing!