Three Advanced Lookalike Strategies to Increase Facebook Performance

By: Tom Murray | Managing Director

Lookalike audiences have been the bread and butter targeting option that put Facebook on the map years ago by being able to take your own data and finding similar customers at scale with relative ease. It became so widespread that almost every digital platform has followed suit with their own form of lookalikes, calling them everything from spend-a-likes to act-a-likes.

Most marketers have probably taken advantage of lookalike audiences at one point or another, and have used the standard options that Facebook provides. As a refresher on the capabilities of lookalike targeting:

1. You can create lookalike audiences in increments of 1%, from 0-1%, all the way to 9-10%.
2. You can create larger segments, such as a 3% bucket, which combines the 0-1%, 1-2%, and 2-3% lookalikes.
3. You can create lookalikes on various sources, such as (but not limited to):
   1. Website visitors
   2. Email customer lists
   3. Facebook and Instagram page engagers
   4. Video views
   5. Mobile app audiences
4. You can create lookalikes in specific countries, regions, or even worldwide.

Now that we got the standard options out of the way, here are three better ways to test out lookalike audiences that are much more advanced than the above.

1. Higher Percentage Lookalikes
   
   As stated above, lookalikes typically go up to 10%. However, taking a deeper look into the Facebook Developer API, lookalike audiences can actually go up all the way to 20%, in the same 1% increments as 0-10%. This provides you a way to target even more broadly, while keeping some level of similarity to a target audience. Facebook has been preaching going broader with targeting over the past year or two, and many can’t get behind targeting the entire 200 Million+ users in the US. Using up to a 20% lookalike can help you still target a mass audience of around 40 Million users, while allowing Facebook to fish in a more specific pond compared to the giant ocean of audience pools.
   
   
2. Lookalikes from Previous Converters
   
   Taking a look at the same Developer API doc, there are ways to create lookalikes audiences from your previous converters. Now you might be thinking, “We can already do that using email lists and pixels,” which is correct. However, this option allows you to create converters from SPECIFIC campaigns and ad sets of your choosing. This can help you focus on a very specific seed audience of people that you are looking to reach. You will need to build a tool via the API using the below code, or you can reach out to us, since we have already built the tool! Keep in mind you will want at least 100-200 conversions at these levels before building the lookalikes.
   
   
3. “Lookalike Stacking”
   
   Lookalike stacking is when you take a large lookalike audience (say 5%) and then overlay another targeting option such as interests on top of this, cutting the audience down. It is not recommended to do this with small lookalike groups like 1%, because you are taking an already small audience (2 Million) and cutting it down likely into the low hundred thousands of users. You should aim to stack and still end up with at least a few million in your pool to assure you of scale, as well as limiting the CPM increase of targeting a small audience.
   
   You can test the same audience with various stacks to compare and find pockets of performance. This is a way to help qualify lookalikes a little better, but as we said, should only be used with higher percentage lookalikes for it to work optimally.