What We Learned Sitting Down with the Facebook Experts

by | Jan 16, 2019 | Social Media Marketing Tips

Facebook AdsThe team at AIM recently sat down with a Facebook optimization expert at Quality Score during which we received a wealth of information on how to best approach advertising on Facebook.

I want to share some of our takeaways with you today. But I have to warn you, this is not a quick read. Set aside at least 15 minutes to dive into this article.

The Low-Hanging Fruit

Let’s get started with the low-hanging fruit.

These ideas are useful, often quick to implement, and sometimes overlooked. However, this doesn’t mean that they will only have a low impact – quite the reverse is often true.

Better still, some of these can be implemented right now.

  • Lookalikes: When you’re using lookalike audiences, you don’t necessarily need to segment your audience by demographic features. In fact, Quality Score recommends that you don’t do this at all. Facebook is skilled enough at creating audiences, where segmenting audiences by demographic just narrows your reach, instead of qualifying your leads.

    Our thoughts: If you can’t bring yourself to remove demographic targeting, consider instead changing the demographics to a slightly younger audience – 40+ rather than 55+.

  • Audience characteristics / lookalikes: Over time, your audience characteristics will collapse in on themselves as you target more and more customers who share the same characteristics. This can make lookalikes less effective at reaching new audiences over time.

    Don’t rely solely on increasing the % range – instead look at seeding existing lookalikes with new data or testing new lookalike bases.

  • Conversions: Make sure your conversion window is on the 1-day click setting. This will ensure that Facebook immediately finds people who are more engaged and also more likely to become valuable customers.

    Our thoughts: This seems like a really great tactic for 2-step lead generation, due to the need for quick-turnaround breakeven. But for 1-step, we think the 7-day click setting would still be viable. Test your campaigns to see which is most effective for you.

  • Testing / Ad Creative: Don’t trust the native A/B testing tool in Facebook. If you want to get meaningful, replicable results, set up manual tests. Duplicate your ad set and run unique ads in identical ad sets.

    Make sure you know your top written creative and your top visual creative. Test variations on images with your best copy, and variations of copy with your best visual.

    Keep moving the needle on ad creative. Don’t let creative go stagnant.

  • Ad Formats: Implement video ads. They’re one of the most responsive formats right now. Quality Score recommends a 10 second video ad.

    Our thoughts: Test a variety of video ad formats. “Selfie – face-to-camera”, interview, slideshow, moving image with text overlay, scribbles. If you have ad creative that is already working well for you, think about ways to adopt it to a different or new format.

  • Scaling / Ad Testing: Use Dynamic Creative Optimization. It’s as easy as ticking a checkbox, and will allow Facebook to algorithmically (dynamically) mix and match your creative elements to find the winning combinations and put them in front of your audience.

    You’ll only find this option for Conversion, Traffic, Video Views, Reach, Brand Awareness, and App Install campaigns.

    Dynamic Creative Optimization

    Our thoughts: We’ve implemented a few tests with Dynamic Creative Optimization, and so far, we are on board! We’ve seen CPLs drop from $4 to $1 in one of our tests (with a CTR of 9.3%), and in another recently launched test, we’re seeing a CPL drop of £.50.

    In the past, we’ve also used Smart Display campaigns and Responsive Search Ads, which are both Google products. Although the results can be mixed, we’ve generally achieved good results from these products.

    While we don’t have down-funnel results yet to review, early tests look very promising. If the lead quality remains steady, this could have a significant positive impact on ROI.

  • Scaling / Interests: There’s a simple way to add more volume to campaigns that target interests.

    Underneath the Detailed Targeting selection box, you’ll find a check box to “Expand interests when it may increase conversions at a lower cost per conversion.”

    You must be using the conversions, app installs, lead generation, post engagement or traffic marketing objectives for this to be an available feature for your campaigns.

    Expand Interests

    Our thoughts: Although Quality Score recommends this, we recommend you do your own testing first – we’ve heard mixed reviews from other sources.

  • Scaling / Budget: We’ve heard this one before, but it’s good to hear it from a third-party expert who understands the inner workings of Facebook as well as our own expertise. If you want to scale a campaign that is performing well, increase your daily budget by 15-20% every couple of days to scale it up. You minimize shock to the Facebook algorithm while giving the budget room to grow.

    Our thoughts: We’ve seen this work in the past, so we are wholeheartedly behind this recommendation.

  • Ad Formats: For a quick test that may boost your conversion rates, try an A/B split test with one ad running without the “Learn More” button.

    The theory goes that this allows the ad to appear more like native content, generating more clicks.

    Learn More Button

    Our thoughts: This takes about 5 seconds to set up and could take as little as two days to test. Why wouldn’t you want to test this trick?

  • Testing / Bidding: When you want to test new targeting options, try using manual bidding with a bid set at approximately 30% more than your target CPA.

    This bidding strategy is less locked down and you give Facebook the freedom to aim at your target while learning quickly. When you use automated bidding, Facebook is locked in to certain performance metrics, and can quickly default to lower quality leads or give up if there’s an initial struggle to find the right prospects. Using manual bidding will allow you to determine the quality of the leads acquired, and then bid accordingly.

Ideas open for debate and consideration

Of course, not everything can be that simple.

Although we always recommend testing new ideas, some of the suggestions raised some eyebrows with us and we are still debating whether they are likely to succeed.

Nonetheless, I still want to share them with you for your consideration, but the final recommendation will be down to you.

  • Post IDs: Social proof can be a big needle mover for ad creative. Theoretically, when you pull the same Post ID across all your ad sets, you’ll benefit from the social proof and post history.

    Our thoughts: We’re skeptical about this one. We’ve tried it in the past, and the negative comments piled up and piled up on that Post ID, followed by numerous ad rejections.

  • Negative Comments: Speaking of comments (especially negatives ones), what do you do with all the comments that show up on your ad posts?

    According to Quality Score, you mostly just need to hide negative comments. You might keep and respond to a few of them, but not all.

    Too few visible comments, and you’ll trigger a higher possibility of the dreaded “x-out”. This is when a viewer chooses to hide or report your ad. It’s not a metric Facebook will allow us to see, so there’s some room for interpretation here.

    Our thoughts: Similar to our thoughts above, we think this one is a bit risky, simply due to the build-up of negative scores (and the hidden nature of the x-out). But since negative comments will continue to exist, it’s worth testing the idea of leaving a few, and hiding most.

  • Scaling / Targeting: When you really want to scale, especially with a 1-step campaign, use Broad Audience targeting on Facebook, and only limit with demographic and location factors. Create a lookalike of this audience as well.

    Our thoughts: I’ve not yet used this feature, but it’s certainly something I would test for 1-step.

  • Advanced Day Parting: If you see a reason to be using dayparting, you should try a campaign with a lifetime budget and use day parting to run campaigns in the morning only, or the afternoon only. Whichever period your data shows might benefit from separating it out.

    Our thoughts: We’re entirely in theoretical mode here, but we also like the idea of pairing this with different messaging for different times of day. For example, in the morning, you might run creative that has a positive message “take control of your finances” and in the evening you might run a more timely “escape the rat race” creative.

From the tests we’ve already run, we know there are some potential home runs in here.

And maybe a few assists that will help your campaigns become just a little more profitable.