As you may know, I've been testing the new ABC LAL audience in collaboration with Facebook. This uses the Facebook API to create a LAL using a campaign or ad set as the seed, as opposed to another audience. This gives us more signals to use. Phase one results below:
1/9
1/9
First and foremost, I am very surprised by how strong this performed, despite being told that others were seeing great results so far (I'm always cautiously optimistic when being a first-mover on a new FB product).
2/9
2/9
I am also very impressed with the consistency coming out of it. Day by day there were no real massive jumps in performance, which is a breath of fresh air.
More info on my build-out below for anyone interested:
3/9
More info on my build-out below for anyone interested:
3/9




4/9

5/9




6/9
While I've not included the source ad set performance in the data above, I can share that the ABC LAL ad set performed 2x as strong as the source ad set. This was a huge win for adding scale to the account.
7/9
7/9
Additionally, this is great horizontal scale because ABC LALs do not cannibalize their source audiences, and on average only see <10% overlap with other ABC audiences.
8/9
8/9
That said, based on what I've discovered above, the next phase of this test is to add a new ABC LAL ad set using a lower-funnel campaign as a seed considering that source size doesn't matter much. Phase two will be an ATC campaign ABC LAL. More info to come as always! 
9/9

9/9