Should you run manual placements on Facebook Ads?
Originally published July 2019. Some details have changed since: Facebook is now Meta, and Meta now calls its automatic placements Advantage+ placements. This is one account’s 2019 test, not a rule for every account.
In one account where Instagram feed clearly beat Facebook news feed, we tested running Instagram only against automatic placements. Automatic won, and Instagram on its own did worse than it did as part of the automatic mix. Our hypothesis: the algorithm picks the best users from a bigger pool of placements.
What was the test?
On that account, Instagram feed outperformed Facebook news feed while the algorithm was optimising across all placements (Facebook, Instagram and Messenger). That raised the tempting question of whether we could roughly double our return on ad spend by targeting only Instagram.
Facebook was strongly recommending automatic placements, so we ran a split test using Facebook’s split test feature rather than assume. The original screenshots show the reported results below.
What happened?
In the initial placement breakdown, reported purchase ROAS was 0.96 for Instagram feed and 0.54 for Facebook feed. In the later split test, it was 1.23 for automatic placements versus 0.27 for Instagram-only, on $680.71 versus $471.10 in spend.
The screenshots do not establish statistical significance or show the attribution window. The initial breakdown and later test are separate comparisons, not a single before-and-after control.
Why do we think automatic placements won?
Our hypothesis was that a wider placement pool gave the delivery system more opportunities to find valuable users. The result does not prove that mechanism. The old post cited a Facebook IQ paper; we could not recover it, so it is not offered as supporting evidence here.
What has changed since?
Meta now calls this Advantage+ placements, and the placement options and available formats are different. Our takeaway is to test a restriction rather than assume that a strong placement breakdown means the same placement will win alone. Placements also affect which creative formats you need; see how to adapt one Meta ad creative across feed, stories and reels and why Meta spends more budget on one ad.
Frequently asked questions
Should I use manual placements on Meta? In our 2019 split test, manual Instagram-only placement did worse than automatic. Test it yourself in your account before you narrow placements.
What are automatic placements called now? Advantage+ placements.
Why would automatic placements beat a better placement? Our hypothesis is that the algorithm has more users to pick from, so it finds the best ones across placements rather than only within one.
Published
