Meta Andromeda Explained: What CPG Brands Need to Change About Creative
Many CPG brands are seeing the same pattern on Meta. Audience targeting feels less predictable, winning ads fatigue faster, and creative teams are being asked to produce more content across more formats with less time.
This is not just another platform update. It reflects a structural change in how Meta decides which ads people see.
That change is being driven in part by Andromeda, Meta’s AI-powered ad retrieval system. For CPG brands, the implication is significant: creative is no longer just the message delivered to an audience. It is increasingly one of the signals Meta uses to find that audience.
What is Meta Andromeda?
Andromeda is Meta’s ad retrieval system. Its job is to evaluate a massive pool of eligible ads and narrow it to a smaller set of relevant candidates. Meta’s ranking systems then decide which ad is most likely to perform for a specific person.
The important part is how Andromeda makes that first decision. It interprets the creative itself, including the imagery, product, language, hook, format, and broader meaning of the ad.
In the old model, marketers defined the audience first and then built creative for that audience. In the Andromeda model, the creative itself helps Meta determine who may be interested.
Put simply, creative is becoming part of the targeting system.
How the delivery model is changing
The Old Model
The Andromeda-Era Model
The difference is not cosmetic. In the old model, creative supported targeting. In the Andromeda model, creative helps perform the targeting.
Why this matters for CPG brands
CPG products often connect to multiple needs, routines, and occasions. A snack can be positioned as a lunchbox solution, a post-workout option, an afternoon energy boost, or a better-for-you indulgence. Each of those concepts gives Meta different signals.
The mistake is treating creative variation as a design exercise. Five ads with the same product shot and five different background colors may look like five assets to the marketing team. To Meta, they may communicate nearly the same idea. Meaningful variation comes from changing the underlying consumer problem, use occasion, product benefit, proof point, or emotional outcome.
Cosmetic variation versus strategic variation
Imagine a better-for-you snack brand launching a new protein bar. Cosmetic variation might include a blue background, a yellow background, a tighter crop, a shorter headline, or a different call to action. Those versions can still be useful. They may improve a promising execution or help adapt it for different placements. But they are still versions of the same idea.
Strategic variation would look different. One concept might focus on the afternoon energy crash. Another might position the bar as a post-workout option. Another could emphasize taste. Another might highlight convenience for busy mornings. Another could use an ingredient comparison. Another could feature a creator-led reaction. The product stays the same. The reason to care changes. That is what meaningful creative diversity looks like.
Only about 5% of ads become meaningful winners
One of the most important findings in the research is that only a small percentage of ads become meaningful winners. Motion’s 2026 Creative Benchmarks study analyzed 578,750 creatives across 6,015 accounts and $1.3 billion in spend. It found that only about 5% of creatives qualified as winners. The rate ranged from roughly 3.8% for lower-spend accounts to about 8.2% for brands spending more than $1 million per month.
That number should change how brands think about creative. If only about 5% of ads become major performers, the job is not to produce one perfect ad. The job is to build a system capable of finding the few ideas that deserve to scale.
This is not a creative crisis. It is the math of the platform. The mistake is assuming every asset should work, then slowing down production when most do not. A low hit rate is part of the search process. The value comes from what the team learns and how quickly those learnings shape the next round.
The creative power law
This is why creative velocity matters.
If your organization can only produce a handful of new concepts each quarter, it is not really running a testing system. It is buying lottery tickets.
100 ads launched
Approximately 5 become meaningful performers
Their hooks, occasions, and benefits shape the next round
Why the production model has to change
Traditional production was designed around campaigns. A brand developed a concept, planned a shoot, produced a set of hero assets, and adapted them across channels. Andromeda rewards a more continuous model.
Creative needs to be developed, launched, evaluated, refreshed, and replaced on an ongoing basis. A concept that performs well this month may fatigue next month. A weak concept may still reveal a strong hook. A creator video may uncover a new use occasion worth exploring. The creative process has to become a feedback loop.
The Andromeda creative loop
The goal is not endless content. The goal is faster learning.
Where AI generation fits
Generative AI can help close the production gap. It can accelerate concept visualization, storyboarding, image generation, hook development, localization, resizing, and versioning. For CPG brands, that can reduce the need to organize a full production for every creative variation. One approved concept can be adapted into multiple environments, formats, and audience contexts.
But AI is not a vending machine. You cannot type one prompt and expect finished, brand-ready content. Even experienced prompt designers usually need several rounds of generation, selection, correction, compositing, and retouching. The strongest outputs often require another pass in Photoshop, After Effects, Premiere, or a similar tool. The real learning curve is understanding how to combine AI generation, traditional design, brand governance, and performance data into a repeatable workflow.
AI changes production. It does not remove it.
The wrong expectation is that AI replaces the creative process. A better model is human-first and AI-assisted. Humans define the consumer insight, concept, brand codes, and quality standard. AI helps expand the idea into more executions. Designers and editors make the output accurate, consistent, and ready for market.
AI is excellent at producing more. It is less reliable at deciding what is worth producing. The broader research supports this distinction. AI-assisted generation is becoming necessary for volume, adaptation, localization, and variation. It is not a substitute for strategy, brand-building, or creative judgment.
Brand consistency becomes more valuable
As content becomes easier to generate, generic content becomes easier to generate too. That creates a risk for CPG brands. More assets may increase testing capacity while weakening recognition. The goal should be controlled variety. The hook, occasion, format, and creator can change, but the packaging, visual identity, tone, and brand point of view should remain recognizable.
The brands that win will not be the ones that generate the most content. They will be the ones that create the most useful variation without losing what makes the brand distinct.
Is Andromeda exposing a creative problem?
Brands struggling on Meta should ask a few direct questions. Are we producing genuinely different concepts, or mostly versions of the same ad? Can we refresh creative every few weeks? Are we measuring performance by concept, hook, benefit, and occasion? Can our AI workflow protect packaging and brand consistency? Are performance insights shaping the next creative brief?
When the answer to several of these questions is no, the problem may not be campaign setup. It may be the creative operating system behind the campaigns.
The brands that learn faster will win
Andromeda is not a reason to abandon strategy in favor of volume. It is a reason to connect strategy, production, media, and learning more tightly. The strongest brands will produce enough distinct creative to find the small percentage of winners, understand why those ideas worked, and apply those insights faster than competitors.
AI can help make that possible. But it still needs a strong concept, disciplined brand standards, and an experienced human hand. The prompt is not the product. The system around it is.
Is your creative operation ready for Andromeda?
85SIXTY helps CPG brands build high-throughput creative systems for Meta, combining strategy, AI-assisted production, brand governance, creative testing, and ongoing refresh programs.
Talk to 85SIXTY about where your current creative process may be limiting Meta performance.
Leave a comment:
You must be logged in to post a comment.