What is A/B Testing and How Do I Do it Well?
A/B testing, also known as split testing, is a tactic by which you find the best combination of ad headlines, main text, images, and call-to-actions for your target market.
A/B testing is usually done over a period of weeks where one portion of the ad is changed at a time and tested against a control in order to find the best results. Every time one ad outperforms the other, one ad is shut off and iterations are made. The key here is to only change ONE THING at a time in order to know why an ad would outperform the other.
Perhaps an example will clarify:
In this example I will show how I would use A/B testing to optimize a fictitious ad for {LAB}
AD A:
Headline: The best resource for limo marketing!
Main Text: Are you struggling to market your limousine business? As a {LAB} member, you will get access to the best marketing and sales resources in the world. With access to Bill Faeth, a former limo operator himself, and the Limo University team you are no longer alone running your business. Sign up today!
CTA Button: Learn More
AD B:
Headline: Marketing help for the limo industry!
Main Text: Are you struggling to market your limousine business? As a {LAB} member, you will get access to the best marketing and sales resources in the world. With access to Bill Faeth, a former limo operator himself, and the Limo University team you are no longer alone running your business. Sign up today!
CTA Button: Learn More
The only thing changed between Ad A and Ad B was the headline. Now, we wait a few days and measure our Key Performance Indicators (KPI) to see which one does better.
For an ad optimized towards Traffic, our main KPI would be Link Click through rate (CTR). For sake of this example let’s say Ad A results in a CTR of 1.5% while Ad B gets a CTR of 2.0%. Clearly, Ad B is far out performing Ad A.
So, now you know which one is better but how do you make iterations in order to truly optimize? Your first move is to shut off Ad A but don’t stop there! You can continue to make iterations in order to try and better that CTR. Now that Ad B is running by itself, duplicate it and create a new test. The headline checked out so this time maybe you duplicate the ad, keeping the winning headline but alter the CTA Button. While it may seem like a small change, it can still have big impact on the result of your ads. For example:
Ad A (The new ad):
Headline: Marketing help for the limo industry!
Main Text: Are you struggling to market your limousine business? As a {LAB} member, you will get access to the best marketing and sales resources in the world. With access to Bill Faeth, a former limo operator himself, and the Limo University team you are no longer alone running your business. Sign up today!
CTA Button: Sign Up
Ad B (Stays the same):
Headline: Marketing help for the limo industry!
Main Text: Are you struggling to market your limousine business? As a {LAB} member, you will get access to the best marketing and sales resources in the world. With access to Bill Faeth, a former limo operator himself, and the Limo University team you are no longer alone running your business. Sign up today!
CTA Button: Learn More
Now, you have a brand new A/B test. Allow them to run for a few days and measure your results using the same KPI as before.
The important thing is to change one thing at a time and run one test at a time. Otherwise, results can get confusing and become much more difficult to optimize.
Without A/B testing, you never really know how well your ads are doing or if they could be doing better. With this skill in your digital marketing tool bag, you’ll be ready to improve results and command ROI with every campaign.