Finding the right custom audience to create a lookalike audience off of can be the difference between meeting and missing your CPA goals. One of the first tests I recommend running, when starting a new account, is an A/B test using different lookalike audiences. Assuming you have quite a few customers, I recommend the following lookalike audiences:
1. All of your email subscribers. This is most likely going to be the largest audience. It provides a ton of data for Facebook to find a lookalike audience off of.
2. All purchasers. If your conversion is a purchase, create a lookalike audience off of everyone who has purchased. This will allow Facebook to create a lookalike audience of people who look like your purchasers. This audience can sometimes be significantly different than the audience created off of all email subscribers.
3. Top LTV customers. This is similar to all purchasers, but if that audience is larger than 15k, we may want to tailor it down by creating an audience off of only your top customers.
Once the three audiences are built, it is important that they are actually different from each other and that there is minimal overlap. Most 1% audiences will have roughly 2 million users in them. Create audience 1 and exclude audience 2. If the audience reach goes below 500,000 users, this means that there is quite a bit of overlap between the two audiences, and you should probably just pick either of the audiences. Then do the same with audience 1 and audience 3, and finally audience 2 and audience 3.
If all the audiences are different, you will want to create 3 separate ad sets, going to each of the lookalike 1% audiences and excluding the others. It’s important to run the different ad sets at the same time with the same amount of spend. From here, it’s up to you to determine which lookalike audience is the strongest!