How To A/B Test Facebook Ads
March 10, 2021
Performance tracking/testing is elemental in every internet marketing operation. A/B testing gives you that bird's-eye view over the effectiveness of your campaign and lets you know what to change for better outcomes. Efficiency may require changing the copy, images used, or the call to action.
A/B testing provides the clues. Here is how to do it.
What is A/B testing on Facebook?
A/B testing on Facebook is fundamentally similar to A/B testing done on Google ads. The simulation reveals which elements of your ad perform better for your KPIs and the target audience. You create varying campaign strategies of the same offering and test which one performs better per dollar of advertising.
A B testing/ split testing is not limited to the elements of your ad copy. You can A/B test your campaign to find out which audiences are more receptive to your message, or what time of day or region works better for your ad. The assays rely on a statistical analysis run by Facebook's algorithm or an external testing tool.
Why should you A/B test your Facebook ads?
A/B testing sheds light on the optimization needed to get a better return from your campaign. It takes the guesswork out of the marketing process by helping you make data-driven decisions on what your audiences want and what works better for them. By measuring the extent of the impact that subtle changes have on your ad metrics, you can tune your campaign into something that yields better results.
A/B testing shows the way to the winning strategy. For instance, by improving landing pages, making headlines clear, or shortening copy, you can significantly improve user experience and engagement levels for our ad. The experiment uses a negligible fraction of your budget.
How to A/B Test Facebook Ads
Do it one variable at a time
Most beginners fall into the trap of testing everything at ago, and then the results aren't clear enough for them to affect the change needed. Let's say you are a professional hairdresser that wants to improve footfall to your shop from your campaign's landing page. To achieve that, you might need to experiment with different visuals, calls to action, and headlines.
But if you A/B test all these changes at the same time, it will be hard to pinpoint what action led to better outcomes. And testing all these variables at once can take several hours. Results may even take several weeks, and that's time you likely don't have on your side.
It would help to create smaller tests that let you successfully track the measurable improvements needed for your campaign. This method is highly recommendable for small campaigns that need gradual optimization. An A/B test that's exclusive for the landing page reveals which layout is loved most by visitors. An image test helps you see which pictures are most appealing to your potential customers, and so forth.
If you deal with a high budget campaign or time-bound sales, we encourage the use of an external ad management tool. There are many tools out there to experiment with, but you can leave your big volume campaign to us. We have reliable tools that will automate the process for faster and clearer results.
The structure of your A/B test
You can A/B test your ad using a single or multiple single variation campaign sets. Use the Ads Manager feature of Facebook Ads to determine and set this structure. A single ad set has one major advantage when running an A/B test; the audience will not see all your multiple campaigns at the same time. It's good when you need to keep things subtle.
The one negative side to a single ad set is that Facebook may start to auto-optimize your campaign, which then corrupts the results. In that case, consider the multiple ad set. Each of your test campaigns goes to a different set. The social media platform will treat these campaigns as separate entities and will, therefore, not initiate auto-optimization.
Test something measurable
Many factors influence the response rate for a Facebook campaign, including timing, ad copy, hobbies and interests, social proof, and bidding strategy. Some of these variables are measurable, but others are not. To start an A/B test on Facebook, you first must create a specific hypothesis around a variable that is quantifiable.
Choose a testing hypothesis that can have a real impact on your ROI for a long time to come. Optimizing the size of your CTA button, for instance, can have a slight effect on response rates, but optimizing the CTA itself will have a bigger impact. You can also test the offer itself. The problem may not be your campaign or your audience, but the offer-test a paid versus free trial of the offer.
Choose the right audience and test at the right time
Your test audience should be broad enough if you are to get accurate test results. Do not use audiences from a concurrent Facebook campaign. Doing that delivers mixed results.
The more experiments you do, the more insight you get on how best to mold your Facebook ad for better conversion. The minimum testing period, as recommended by Facebook, is four days, and the maximum is 30 days. Tests that fall within that time frame generate more conclusive results. Anything longer may waste your budget.
At SparrowBoost, we determine the test duration period after carefully analyzing your customers' behaviors. If they take a shorter time to convert after seeing an ad, we shorten the A/B testing period to save your time and budget. Other statistical variables we consider in these tests include the number of visitors your page gets and conversion volumes. These factors help us to arrive at the best duration and budget for the test.
Determine the A/B test budget
The right budget for your split test will lead to results of a statistical significance. Facebook often provides its suggested budgets if you cannot confidently come up with a definitive figure. For us, the budget of an A/B testing endeavor depends on the number of versions we are running. Testing on many variables helps to come up with sufficient impressions for making an accurate decision on the winning ad. We can help to economize on budget while still getting the needed conversions for a 95 %+ statistical significance.
Run the test
After you create element variations in your ad and decide on the budget, the next step is to run the test. Facebook has automated the process, so there are fewer manual adjustments involved. Use the Split Testing feature provided to test the different components of your campaign for easy to understand results.
This Facebook feature lets you run an AB test on up to 5 versions of your chosen ad variable. You can schedule your test duration here and select the option to have a report after the test. When you start a test, wait for the scheduled run time to compete. Don't stop on day one just because you have hit your targeted sample size. Wait it out.
Review the results
If you are using Facebook's Split Testing feature, Facebook determines the winning ad for you. No one is sure about the nuances of their algorithm, but one significant thing they look at is the cost per conversion. The automated A/B testing process on Facebook also provides a near-real simulation of how the ad would perform if you run it at scale. The test results will be displayed in the Ads Manager table, and a copy of the report is also emailed to you.
If you are doing it manually or with an external tool, take the conversion results of the test and stack them up against each other. If you have a sampling threshold of 95 % or 99 %, it will be easy at this point to decide a winning ad. Before that, however, you must compare your current metrics with the original numbers before the test. If the new conversion rates are higher than the previous ones, the test was a success.
What now after A/B testing your Facebook Ads?
Now that you have A/B successfully tested your Facebook ad and chose a winner, what next? The point of split testing is to help you get better sales from your campaign. Here are some of our post-A/B testing best practices at SparrowBoost:
We strike out underperforming ads
Once we get precise data on the best campaign direction from the split test, we immediately pause the underperforming ads. Our extensive range of testing tools often makes it easier to see what is working and what's not working, regardless of the volume of the test.
We scale up the winning ad
We put your budget to productive use, based on the data from the test. We make incremental budget additions as we maintain real-time monitoring of how that affects conversion. By maintaining continuous testing, we can enhance the campaign even further by trying it on different placements.
We can save you from a conversion cul-de-sac
At SparrowBoost, A/B testing is just one of the strategies we use to create effective PPC campaigns. Our PPC management capabilities include data analytics and ML designed to generate the most return per dollar of your campaign budget.
Contact us for a free proposal.
About THE AUTHOR
After working for multiple digital advertising agencies and managing hundreds of client accounts, spending millions of dollars via Google Ads, Facebook Ads, Native Ads and Direct Media Buying, I took things out on my own and started SparrowBoost. Now, my tight-knit team and I continue to get smarter and more efficient at running our own campaigns and we share our knowledge with you.Learn more about SparrowBoost