The most critical component of every content asset you create isn’t the asset at all, but the promotion of that asset. Without a paid media budget, your content is the proverbial tree in the forest. It might be exactly what someone is looking for, but it goes unfound. If you’re going to use a portion of your marketing budget for paid ads (and you should), you might as well gain some valuable information about your audience while you’re at it, right? The way to do that is by running strong A/B tests.
Enterprise companies use a/b testing to improve results on everything from landing pages to ad CTAs. Successful companies like Google, Amazon and Facebook perform more than 10,000 controlled A/B tests every year. When run correctly, they can dramatically improve results: In one example, Dell reported a 300% increase in conversion rate from A/B testing. Still, many marketers struggle to execute proper A/B tests. In fact, 56% of marketers don’t know the proper methods for A/B testing and 52% said they don’t have enough time to A/B test.
A/B Ad Sets Per Content Asset
The number of ads you create depends on the type of content asset you’re promoting. Shorter content, such as an infographic or a blog post, might only warrant 2-4 ad sets, while longer form content such as an eBook or interactive quiz will benefit from additional ad sets. I typically create one ad set per chapter or section, plus an additional set for the entire asset. So, an eBook with 6 chapters might have 7-8 ad sets. An “ad set” is two versions of the same ad: an “a” and a “b” version. The purpose of the test is to uncover critical information about your audience and gain a better understanding of what drives them to act.
The single most important rule for creating an a/b test is that you test only one element per ad set. This way, you’ll quickly be able to identify what your audience responds to and what elements are inconsequential. If you’re creating two sets of ads (2 “a” versions and 2 “b” versions), be sure you’re testing different elements. You won’t learn much if you continue testing the same thing over and over again. And lest you think you’ll quickly run out of tests, here are 28 a/b tests to get you started.
Image: photo vs drawing
Image: one person vs. group of people
Image: person looking at camera vs. person looking down or away
Image: background color
Image: static vs. animated
Image: person vs product
Image: graph vs. text
On-Image Copy Tests
On-image copy vs. no on-image copy
On-image stat vs. on-image ratio
On-image question vs. on-image statement
On-image text color
On-image CTA button color
On-image CTA vs. no CTA
On-image CTA button vs. CTA link
Post Copy Tests
Post copy short vs. long
Post copy question vs statement
Post copy: audience call-out vs. generic
Post copy: stat vs. ratio (90% or 9 in 10)
Title Copy Tests
Title copy: CTA vs. no CTA
Title copy: question vs statement
Title copy: short vs. long
Soft vs Hard (discover vs download)
Top vs bottom funnel (learn more vs. buy now)
Timely vs open (download now vs. download)
Formal vs Casual (register vs. save your seat)
Gate: landing page vs. LinkedIn form
Gate: pre-download vs. pre-results
Gate vs. no gate
What to Look for in Your A/B Tests
Consistent results across multiple ad sets means you’ve uncovered important information about what your audience responds to.
CTRs that are significantly different between sets means you’ve learned something new about your audience to use in future ads.
a/b versions with insignificant differences (.02% or less) means your audience likely doesn’t have a strong opinion on what you tested. Choose a different test.
Remember to report your findings not only to your leadership team and/or clients but to the content creators! The writers and designers should be using your a/b test findings to craft stronger copy and images that earn results.
Need some help managing a/b tests or your LinkedIn Campaign Manager account? Give me a shout! [email protected].