Facebook is an incredibly powerful platform, where marketers can seek and engage new prospects across a variety of business goals (purchases, lead generation, awareness, etc.). As I’ve said before, the channel is an indispensable asset for generating prospective student leads to the degree programs we market.
Beyond the vanilla demographic targeting options available (Interests, Fields of Study, remarketing, etc.), wrapped within custom audiences, there lies an enormous bastion of advertising potential. This is the second part of my series dedicated to Facebook’s Custom Audiences, and today’s post examines Lookalike Audiences – a feature that is undoubtedly one of the greatest tools the channel holds in its arsenal. Here’s a bit of info:

Custom Audiences Beget Lookalike Audiences

Although powered by data from Facebook’s trusted third-party partners Acxiom, Datalogix, and Epsilon, Lookalike Audiences must have a source to mimic in order to inform the amalgamation of a viable target audience. If third-party data partners are the fuel, then Custom Audiences are the engine of this vehicle. Traditionally, Lookalike Audiences are cloned from one of these 3 sources: Custom Audiences (Customer Lists: email addresses, phone numbers or mobile advertiser IDs, Website Traffic, App Activity, or Engagements), Conversion Tracking Pixels, and Facebook Pages; of these options, I have found Email Lists and Conversion Pixels both to be viable resources for spawning a killer Lookalike Audience.

The more data, the better…to a point

Screen Shot 2016-07-21 at 2.44.56 PMAccording to Facebook, “The optimal source [for creating a lookalike audience] is 10,000 to 50,000 people”; however, it also notes “Lookalike Audience performance declines when it’s based on a source of over 50,000 people.” If acquiring an email list of this magnitude is not feasible – don’t fret! “The most important aspect of a source is that it is made up of high value customers” – and I’ve seen exceptional results from lists as small as 1,000 high-quality users.

Lookalike Audiences are HUGE

That list size of 1,000 to 5,000 users? It’s about to seem really, really tiny. Get ready to reach a vast amount of users across Facebook’s placement options (Desktop, Mobile, Instagram, etc.) – and that’s still while only leveraging the highest degree of similarity Lookalikes can be set at. If Customer Lists are David, then Lookalikes are at least ten Goliaths.
In the penultimate step of Lookalike creation (which we’ll cover next), remember, “when choosing the size of your audience… smaller audiences more closely match your source audience. A larger audience increases your potential reach, but reduces the level of similarity to your source audience.”
So, let’s walk through creating a Lookalike Audience:
Custom Audiences HEMJ 1
With a Custom Audience (or Conversion pixel) already in mind, navigate to the ‘audience’ section of your business manager and select Create Audience > Lookalike Audience. A light-box will appear, prompting you to ‘create a lookalike audience’:
Custom Audiences HEMJ 2
Click on the Source entry field and you will find a myriad of your saved Audiences available to choose from. Next, select a Country to target.
Custom Audiences HEMJ 3Finally, select an Audience Size; this is where you will decide the degree of similarity between your source audience and the resulting Lookalike. As illustrated in the graphic below, “Audience size ranges from 1% to 10% of the total population in the country you choose, with 1% being those who most closely match your source.”
I prefer to begin with a one percent Lookalike (the highest degree of similarity). When performance of that segment starts to ebb, I then begin to spiral out towards three percent — although I haven’t dared go beyond this degree of dissimilarity, my Paid Search teammates have gauged success in as large as a four percent Lookalike audience.
Quick tip: it’s a good idea in managing concurrent Lookalike segments to sidestep intra-account competition by excluding Lookalikes of the same kind (source) from their peers (e.g. excluding the one percent Pixel Lookalike Audience from your two percent Pixel Lookalike Audience); this can be managed on the ad set level. Furthermore, if you proceed beyond two percent in defining your Lookalike’s size, it may also be a good idea to narrow your audience by implementing cross-targeting (see “How does detailed targeting work” under Facebook’s Targeting Basics); here I like to start with Facebook’s old glory, Interests.
Have you tried Lookalike Audiences? Let us know your thoughts!
Andrew croppedA graduate of the University of California, Andrew is our analytics and paid search team lead. He is both Google Analytics and AdWords certified. With an ROI-focused and problem-solving approach, he researches, plans, and manages our clients’ PPC campaigns.
 
Next time on Facebook Ads Tips & Tricks: Boosted Posts vs. Newsfeed Ads – which is the right fit?
 
 

Leave a Reply