MRI Blog
Working Around Facebook’s Lack of Split Testing Capabilities

Split-Testing in Facebook
Posted on Aug 27, 2014 in Social Media Comments

Split testing, also known as A/B testing, allows marketers to compare the performance of two campaigns with different variables. For example, the performance of an ad that features an image can be compared to a text only ad. Or the performance of several ads with different copy, or different calls-to-action, can be assessed. While you hear a lot about the importance of split testing Facebook ads, it’s actually not as easy as it should be.

Unless you’re using a browser-based extension, straight split testing on Facebook is limited. For whatever reason, this isn’t talked about very much. Let’s explain. 

For starters, if you place the recommended 3-4 ads into an ad set, the ad that generates the most clicks early on will be the ad that gets most of the available impressions. If you’re thinking this sounds great, since surely this must also mean THIS is the ad that’s converting the best, think again.

Not only is it common for one ad in a multi-set to get double the impressions of another, but also some ads in the set may never display at all.  If all the ads aren’t actually being shown, it’s virtually impossible to know if the ad with the most impressions is truly outperforming the others.

The good news is Facebook’s Ad Sets have made it much less cumbersome to split test than it used to be at the campaign level. Here are a few workarounds that can be utilized until Facebook begins to offer true split testing capabilities.

Place Ads in Different Ad Sets 

Rather than place all of your ads in the same Ad Set, create a different Ad Set for each one and run the campaigns separately with one starting right after the other. This gives you some analytics to factor in as you choose a campaign or two to move forward with.

Test Ads for One to Two Days 

Facebook’s algorithm (at least for the time being) sends an initial burst of traffic to the newest ads. This helps achieve actual ROI from campaigns that are still very much in test mode.

Run Two Ad Sets Simultaneously 

Running two Ad Sets head-to-head, targeting the same group, is the closest we get to replicating a true A/B test on Facebook. The flaw to this process is it drives up costs since you’re bidding against yourself. Use this approach sparingly when testing out an entirely new ad type or concept – not minor copy changes. And again only run the ad for a day or two to gauge results.

Facebook is continually evolving and making changes to its platform. Hopefully adding actual split testing capabilities is a feature they’re looking to add soon. In the meantime, these workarounds offer a few alternative ways to test different Facebook ad campaigns.

Share
comments powered by Disqus
Back to Blog Home
Gaming by State
Native American Gaming