A/B Testing with Digital Ads: A Beginner’s Guide

Reading Time: ~3 Mins

When you’re running digital ads for your company or client, you want to make sure you are running the best possible version of your ad(s) as possible. You may be asking yourself, “How do I know which is the best version of my ad?” A/B testing allows you to answer this question.

What is A/B testing?

A/B testing is also known as split testing. It is the act of testing a small variation in your ads or ad groups to find out which performs better. Once you have the specific variable you’re going to test (ie. testing ad copy with emojis against ad copy with no emojis), you’ll show each version to two similar sized audiences. After a period of time (ideally 3 weeks or longer) you’ll be able to determine which one performed better. 

If you are A/B testing on the ad set level, you may be testing your ads against two different audiences, such as an audience with the age groups 18-24 vs an audience with the age groups of 25-34.

If you are A/B testing on the ads level, you may be testing out variations of images, headings, descriptions, call-to-actions (CTA) or ad formats (such as single image versus a carousel). When you are testing on the ads level, it is important to only test one single variable at a time. This will help you determine what change was making the improvement in your ad.

Why should you A/B test?

A/B testing lets you find the best variation of your ad as possible, as well as finding the best audience to show your ads to. This can significantly increase your ROI.

Not only does it increase your ROI, it helps you understand your target audience. You’ll gain insight as to what content and language resonates with them. You’ll also be able to increase your conversion rates. By knowing what works in your ads and what doesn’t, you can easily set your ads up for success.

What can you A/B test?

In social media ads there are a variety of variables you can test, such as:

• Ad copy (short or long text, use of emojis, quote or question style, numbered list, punctuation, etc.)
• Creative (different images or image vs video)
• CTA (Learn More, Shop Now, Apply, etc.)
• Ad Format (image or carousel)
• Target Audience

If you’re running search ads on Google, variables you can test are:

• Headings
• Descriptions
• Bidding strategy (maximize clicks vs maximize conversions, etc.)
• Audience targeting
• Keywords

Analyzing your A/B testing data

Before you end your A/B test, you want to make sure you’ve captured enough data for reliable results. Typically, you’ll want to test for 3 weeks or longer. Once your test period is complete, it can be quite clear which one is the winner. This will be noticeable in the amount of conversions (or other KPI you identify) each variation had.

Occasionally, it can be unclear which one is the winner. To help analyze the data, compare your link clicks and overall conversions for each ad. If they both had the same amount of conversions, but one had less link clicks, the one with less link clicks would be deemed the top performer, as the conversion rate would be higher (less clicks needed to drive the same amount of conversions).

If you find that your results from your test are inconclusive, test another variable of the ad. The more you test, the more data and insights you will gain about what is working and what isn’t working. 

Key takeaways

Ask yourself what your goal is for your A/B test and why you’re testing particular variables, know how to analyze your data and always remember to only test one variable at a time.

Written by Cheryl Goeres

---