Convert more of the traffic you've already paid for with Mutiny.See how Mutiny works
Browse dozens more proven playbooks from other B2B marketers.See more playbooks
What you’ll learn
What you’ll need
CXL is a learning platform for marketers. They offer mini degrees that are in-depth training programs taught by leading marketing practitioners.
Ognjen Boskovic, Growth Lead at CXL, knew that before a customer buys a CXL training they like to consume some of the free content first. When the prospect is ready, they'll immediately think of CXL because they've already gotten so much free value from their content.
One of their most popular channels is their newsletter. But organic newsletter growth is slow and unpredictable. He needed a way to predictably grow the newsletter audience, while still staying profitable.
Ognjen and his team ran this experiment with one hypothesis in mind:
Is it possible to deliver content in-feed through Facebook Ads as a way to get email signups for less than $2 a lead?
Facebook had recently increased the limit to the amount of copy (words) allowed on an in-feed ad. With a small window of opportunity, Ognjen had the brilliant idea to create in-feed Facebook ads that were copy-heavy rather than relying on a visual design to stand out.
Ognjen also knew the maximum cost per lead he was willing to pay was $2.
He did it, and here's how you can too:
The first step is to create different kinds of target audiences. For CXL, Ognjen created 4 different targeting variations; agencies, newsletter subscriber lookalikes, B2B SaaS startups, and digital marketers.
Lookalike audiences are a powerful tool within Facebook’s ad platform. By uploading your existing email list, they can find more Facebook or Instagram users who have similar personas.
The next thing is to make content with different creative styles and pair them with a target variation. In the case of CXL’s desired audience, they’re targeting a few different types of startup roles.
Therefore, he needed to be specific with the language used in the ads. For instance, the ad targeting the digital marketing segment needed to be clear enough of what they were signing up for, but also eye-catching enough that they wouldn’t ignore it. Ognjen opted for a clean look and used copy that called out that they were advertising but not selling you anything.
Here's what that ad looked like:
CXL's top performing ad uses pattern interrupt to get the attention of marketers in their target audience.
Once you've let the ads run long enough to get statistical significance, begin following this framework to continually optimize both your targeting and your creative:
Declare 2 targeting winners and 2 losers (at the ad set level)
Within the 2 targeting winners, declare 2 creative winners (campaign level)
Turn off the losers, start optimizing the winners
Replace the losers with new targeting and creative bets. This could be iterations on existing winners, or totally new ideas.
Make your decision to kill or continue an ad based on the cost per email signup, while also keeping an eye on cost per thousand impressions (CPM) and return on ad spend (ROAS).
"With this setup, we were able to see which creative styles and targeting segments work best, and also which creative style works best with each targeting segment. By doing this, we were able to maximize the learning." said Ognjen.
Comment section went crazy. In my media buying experience, that's almost always a good sign." said Ognjen.
But here's what the numbers from this experiment ended up looking like: 1,592 email signups with an average cost per email of $2.75. They also wanted to use this campaign as a brand awareness play, so they were also happy with the $3.89 CPM over 350k impressions.
Despite not selling anything directly from the ad, they saw a few leads from this campaign actually make purchases on CXL’s website. Taking that revenue into account, it led to a $1.65 cost per email and a $2.33 CPM, making this experiment a success.
Learn from Fabien David at Notion how to personalize landing pages based on campaign (UTM), prioritize what to test based on your data, and the importance of starting small and testing quickly.