Beyond the Basics: Advanced Facebook Advertising Techniques

Beyond the Basics: Advanced Facebook Advertising Techniques

By Will Morel • June 23, 2015

Whether you’re a user acquisition specialist at a large mobile gaming company or a marketing manager at a small business, Facebook advertising can be quite intimidating. The social juggernaut’s platform capabilities have grown extremely powerful, and there is a tremendous amount of knowledge required to wield them effectively.


Facebook is a fairly self-aware organization, and recently introduced Blueprint, a series of e-learning modules designed to help individuals develop the basic skills required to make the best possible use of their platform. Nonetheless, if you want to become a true Facebook advertising whiz, you should familiarize yourself with several of the key tools that are available to you. This article is the first in a series that will cover the more advanced tools that Facebook offers.


In early 2013, Facebook rolled out Lookalike Audiences, perhaps the most powerful tool for marketers since Google AdWords. Lookalike audiences (aka “LALs”) are based on Facebook’s mountains of behavioral and demographic data, and represent the slices of Facebook’s population that most closely resemble an initial seed audience. The most basic seed audience for any brand is typically their list of fans. Using LALs, a business can find the Facebook users that most closely resemble their existing fans, knowing that these individuals are more likely to take some specified action than a random global user.


A more sophisticated seed audience might be a CRM e-mail list of current customers or e-newsletter recipients. Alternatively, it could be one of the Website Custom Audiences (“WCAs”) that Facebook helps marketers build by placing “Pixels” on certain websites. For example, if you have an “add to cart” (we’ll abbreviate this “ATC”) page on your website, Facebook can accumulate a list of all users that visited that specific URL. Name it “WCA ATC” and create some lookalikes (“WCA ATC LAL”, perhaps?) based on those website visitors that have shown an intent to purchase. Going a step further, you can create a Conversion Pixel for your e-commerce site, and track everyone who has purchased a product. You can even “train” pixels to only track checkouts of a certain volume; imagine how valuable a “Checkout Whale LAL” might perform.


Regardless of the size of your seed audience (we recommend a minimum seed size of 1,000 users), Facebook allows you to select the size (and scope) of your LAL. With over 200MM monthly active users (MAUs) in the US and Canada, a 1% LAL (meaning the one percent of the broader audience that most closely resembles the seed audience) would be approximately two million people! For a 5% LAL, it would be closer to 10 million. Facebook is doing the targeting for you, using their exabytes of proprietary data.


It’s important to remember that, in general, as your target audience size increases, the cost per impression will decrease, since there is more ad inventory available. Unfortunately, because broader targeting usually means less precise audiences, the quality may decrease as well. So the key question becomes: Do you buy the low quality stuff or the high quality stuff? These days, it’s all about return on ad spend (ROAS), so testing different LAL sizes is a critical exercise for the sophisticated Facebook advertiser. Based on ROAS performance, you may find that the “cheap stuff” provides a better bang for your buck! (Note: We’ll plan to look more closely at ROAS performance in a future blog post.)


To take it a step further, here at Ampush we can build “nested” LALs, with which we segment the different LAL slices. Because Facebook allows advertisers to exclude certain audiences from targeting (at Ampush, we call these “holdout audiences”), we can create a targeting group for which we target a 2% LAL from a seed group, but exclude the 1% LAL for the same seed group. (Editors note: Joyce Tabujara, one of Ampush’s talented in-house graphic designers, created this post’s graphic to help illustrate how one might think about nested LALs.)


Ampush tested nested lookalikes for one of our e-commerce clients, and the process was extremely helpful in driving efficient performance. This nested approach allows us to identify precise segments of a group within a broader LAL set, and isolate the best ROAS for our clients.


Isn’t this fun? And we’re just scratching the surface of what Facebook’s powerful engine can do. Stay tuned for the next installment of “Beyond the Basics,” and happy targeting!




Will Morel is the NYC-based Director of Media Analytics at Ampush a tech-enabled marketing services company that helps businesses achieve their goals on native, in-feed social platforms. A lifelong New Yorker, he cares about people, metrics and outcomes. He would love to talk to you about anything you deem worthy of a conversation, and would be happy to connect with you on LinkedIn.