Three Facebook Ad A/B Tests You Should Run ASAP

By — 02.23.18

As the Head of the Paid Social team at Adlucent, I see myself as equal parts marketer and experimenter.  With so many variables in digital marketing, it’s crucial to possess a curiosity to hypothesize, test, and optimize my work.  Not only does this process make me a better marketer, but it helps out my clients, too.  That’s why we regularly conduct tests here at Adlucent.  And since the goal of experiments is to become better at what you do, I’d like to share a couple of our tests with you with the hope that it will help you out, too.  So, let’s dig into the latest Facebook Ads tests we conducted.  

Question:
How do we optimize our efforts to ensure we are showing the best ads to the right audience, at the right time, for the right price?   

Hypothesis:
To improve results for our clients, we should optimize our campaigns for conversions and leverage automated solutions provided by Facebook.  This will retrieve the highest ROAS for clients and streamline efforts on my team.

Testing Process:
We ran three separate tests over a three month timeframe.

Test 1: Optimizing for Conversions vs. Impressions
We began by splitting our audience in half and running an A/B test.  The first group received ads that were optimized for Conversions while the second group received ads that were optimized for Impressions.  The difference between the two tests was that ads that were optimized for conversions were shown less frequently and cost more than their counterpart but were delivered to a more targeted segment of the audience.  At the end of the third day, we found that while optimizing Facebook campaigns for Impressions is great for generating inexpensive brand awareness, optimizing for Conversions allowed our clients to earn the interest of shoppers with the higher chance of clicking our ads on their way to becoming a customer.  Since our clients typically place more importance on click-through revenue, the test revealed that optimizing for Conversions was our winner. We would consider optimizing for Impressions for retailers with less site traffic, retailers who are click/view agnostic, or retailers interested in driving brand awareness with frequent advertisements to their target audience.

Test 2:  Automated Bidding vs. Manual Bidding
Much like Google AdWords, Facebook has designed a pretty reliable bid auction and ad delivery algorithm.  But we wanted to put these features to the test.  With a goal of conversions, we again split our audience into Group A and Group B.  Group A utilized Automatic bidding (meaning the bids were set by Facebook’s algorithm) and Group B utilized manual bidding (set by us).  This test took some time to achieve a statistically significant difference—2 months to be exact.  But at the end of the second month, we found that Facebook’s auction and algorithm provided a more reliable mechanism for bidding and serving ads to the correct audience.  With Group A winning the test, we became confident in Facebook’s bidding algorithm and felt more comfortable giving Facebook the freedom to automatically serve ads to individuals in our audience most likely to efficiently convert to customers.

Test 3: Accelerated Delivery vs. Standard Delivery
Dividing our audience again into two groups, Group A delivered ads at an accelerated pace—meaning ads were displayed immediately to target audiences without cost and timing constraints until daily budget was spent and ads were shut off.  Group B received ads in a strategic cadence, with Facebook serving ads to target audiences at the right time for the most efficient price.  Within a matter of days, we were able to see that Standard Delivery would become the obvious test winner.  Standard Delivery makes sure you’re being the best steward of your conversion-focused budgets. That said, if you’re under a tight timeline (for instance promoting a flash sale) or if you’re concerned about not driving enough impressions then Accelerated Delivery may be a good option for you.

Results:
With our focus on helping clients achieve optimal ROAS, these tests confirmed it’s best to optimize for conversions, utilize Facebook’s automated bidding function with its robust algorithm, and leverage Standard Delivery to reach the most qualified people at the right price.

Conclusion:
It’s important to note that these tests were conducted with the initiatives and goals of our clients in mind.  Your tests should always match the goals you’re looking to achieve, and advertising initiatives should be tailored to speak to the individuals you’re trying to reach (aligned to their current status within your marketing funnel).  By following these methods and regularly conducting A/B tests, you’ll have an optimized Facebook Ad program.

Have you conducted similar tests?  If so, what were your results?  Let us know in the comments below.


2 Responses

  1. Olivia says:

    Nice blog and excellent presentation you have here. I was just looking for some constructive points for my social media practice and there are some great resources here. I really love the way you describe your thoughts.

  2. Steve Bonino says:

    Thank you, Olivia! Glad to hear our approach helped you. If you run tests like the ones outlined in this post, let us know whether your findings align with ours. Hope you see some great results!

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload the CAPTCHA.

Scroll to top