With over 1 Billion active users on Facebook worldwide, one of the hardest things to get right with your campaigns is audience targeting.
Luckily, Facebook offers a fantastic targeting option for performance advertisers called Lookalike Audiences. If you’re reading this blog, you’re probably already familiar with the basic concept:
You choose a source audience on Facebook (e.g. an uploaded list of buyers or leads, a website custom audience, fans of your page) and Facebook identifies the common qualities of the users in that source audience and gives you a new audience in the country of your choosing, that “looks a like” to your source. 1% being the most similar to your source, 10% being the least similar.
Lookalikes can be a powerful tool in the hands of an experienced advertiser. But if you’re not the most seasoned campaign manager, you may be wondering how to best leverage these audiences in your campaigns…
For the newbies out there, you’re in luck, what follows is our internal 7-step checklist for Lookalike Audience Success we use for our own Facebook campaigns, and our clients…
Step 1. Test the Highest Quality Sources First
If a lookaliike audience gives you a cohort of users similar to your source, it stands to reason an audience similar to your high price point buyers will perform better than your audience similar to your page fans.
Our testing over millions of dollars in spend and scores of different clients bears this out.
When testing your lookalike sources, you’ll want to start from your highest value users first, and make sure there is enough of a distinction between each source.
For established businesses, start with:
- High value buyers (users who’ve bought a high price point)
- High Volume buyers (users who’ve purchased multiple times)
- Buyers of a specific product
- Here’s a list of possible seed audiences for all you lead gen advertisers out there:
- High value buyers
- High volume buyers
- Buyers by specific product
- Buyers by channel (Facebook vs Google vs Amazon)
- All buyers
- Leads for specific funnel (upload and custom audiences)
- Leads by specific sources (upload and custom audiences)
- Leads from the last 30,60, and 90 days (upload and custom audiences)
- Sales page visitors
- Order form visitors
- Site visitors from the last 7,14 and 30 days
- Video engagement audiences for your high converting ads
- Phone lead list
- Phone + email lead list combined
- Purchase event lookalike audience
- Add to cart event lookalike audience
- Mobile buyers
- iOS buyers
- Android buyers
- Desktop buyers
- Mobile, desktop, Android, iOS leads
Step 2. Keep in Mind Your Ideal Seed Audience Size
We’ve found that the ideal seed audience size for lookalikes is somewhere between 1,000 and 10,000. Any smaller than 1,000, Facebook may not have enough info to find the right users who are similar to your source. Any larger than 10k, the similarities between users in your seed audience become more general and the similar audience becomes more general, and less valuable in turn.
Step 3. Always Start With 1% Lookalike Audiences
To limit your risk while testing LLAs (lookalike audiences) you’ll always want to start with audiences of 1% similarity.
We’ve found that in a majority of cases, 1% LLAs provide the most efficient spend. While you’re testing your source audiences, you want to limit risk as much as possible.
Step 4. Use the Audience Overlap Tool Determine If There Is Significant Overlap with Your LLAs
Before you go about launching 10 different 1% LLAs, you’ll want to check the overlap between each using the Audience Overlap tool in the Audiences tab of your account. Select both the lookalike you want to check for overlap, click the “actions” dropdown, then “show audience overlap”
We’ve found that if your LLAs have more than 30% overlap that it’s a good idea to include them in a single ad set, or exclude your highest value LLAs from your lower ones.
Step 5. Match Your LLA to Your Placement and Device Targeting
Did you know that you can make a custom audience of everyone who has bought on an iOS device? Or an Android device? Or on Desktop?
One effective strategy I’ve found for LLAs is pairing the source audience with the placement or device targeting. If you’ve found that iOS devices have a high ROAS in our Facebook Audience Audit dashboard, then you can make a source audience of iOS purchasers from your pixel by using the “advanced options””
For the best results, you can target your iOS purchaser LLA in ad set only targeting iOS mobile traffic…
Step 6. Test Higher Percentage LLAs For Sources That Work With 1%
After you’ve tested through your seed audiences and found your winning 1% LLAs, it may be time to expand on those winners with higher percentages.
Previously, when testing higher percentage lookalikes, we would use a “nested” lookalike strategy to limit overlap and avoid targeting the same people with too many ads.
Now, you can the “advanced options” feature when creating your LLAs, and saved a step when targeting your high percentage LLAs.
Step 7. Layer Interests & 5 or 10% Lookalikes for Added Scale
For advertisers with more niche offers, you may find that lookalikes beyond 1% don’t perform as well because the audiences are too broad…
One way we’ve found to get extra scale with your LLAs and avoid the audiences being too untargeted is to layer interests and behaviors with your 5% or 10% LLAs.
In other words, in one ad set, you’ll want to include your high percentage lookalike, and also target some highly relevant interests or behaviors, ending up with an audience that is “Lookalike users who are interested in X”
To find your highly relevant interests, open up audience insights, and add either:
a) your custom audience of buyers
b) your custom audience of leads
c) your competitors interests
Go to the page likes tab, and sort by relevance.
Now you have your interests to target in conjunction with your lookalikes.
Want a Quick Reference Checklist for These 7 Steps?
Download our 7-Step Checklist to Lookalike Audience Success and keep it handy as you’re testing lookalike audiences. Click on the button below to get your copy.