Shifting the Focus: How Meta’s AI Advancements are Changing How Meta Advertisers Should Prioritize Their Efforts
Traditionally, people have seen the pillars of a successful Facebook (Meta) ads campaign built around the following:
Targeting: Precise audience segmentation to reach the right people based on demographics, interests, and behaviors.
Ad Creative & Format: High-quality visuals (e.g. carousel, video, image) and compelling ad copy that resonate with the target audience and capture their attention.
Ad Placement: Proper selection of where your ads will appear across Facebook, Instagram, Messenger, and the Audience Network.
Budget and Bidding: Effective budget allocation and choosing the right bidding strategy to maximize ROI. Lowest cost, Cost Caps, Bid Caps, Target ROAS, etc.
Visualization of Meta’s Lattice
Two of those pillars, Targeting & Ad Placement, have been extensively targeted by Meta as two elements of their advertising platform that can benefit the most from their AI, and thus no longer need customization from advertisers.
Meta rolled out its suite of automated tools under “Meta Advantage.” This system was designed to target audiences and placements based on the likelihood of the subsequent ad placements achieving the advertisers goals.
A further development occurred in 2023, when Meta rolled out it’s “Meta Lattice” deployment, which Meta calls “a new model architecture that learns to predict an ad’s performance across a variety of datasets and optimization goals that were previously supported by numerous smaller, siloed models.”
What’s most important, is Meta’s indication that their AI models “progressed to building hundreds of deep neural network models with trillions of parameters.”
Meta’s Advantage+ Suite of Products in Action
The main difficulty in manually trying to optimize audiences and placements, is trying to compete with an algorithm that factors in TRILLIONS of parameters/data points that only Meta has access to. Accelerated by Apple’s 2021 iOS update that significantly hampered Meta’s ability to track user data outside of the Meta app, Meta now is leveraging it’s vast user data of interactions people make within its app or web UI to create customer profiles and serve them ads accordingly.
When competing against a machine with vast computer power crunching trillions of data points in real-time, the human Meta advertiser making significant placement or audience decisions seems like a losing battle. And realistically, it is a losing battle.
With time constraints a Meta advertiser has optimizing an account, it’s much better used on an element that Meta’s AI is unable to compete in: content creation. By submitting to AI’s significant advantage in making placement and targeting decisions, optimization and creation of creative content decisions can now lead to significant impacts on performance in ways that Meta’s continual automation expansion is unable to compete with.
