🌌 Meta Andromeda is one of the most talked about topics in DTC right now, and for good reason. It’s the biggest backend change to Meta’s ad auction system since iOS 14.
I’ve been digging in this year… reading, testing, and trying to understand the why behind it. The rollout still seems to be happening in waves, so the impact definitely varies brand to brand (and I’m still learning too! 😅 ).
First and foremost Andromeda wasn’t a philosophical pivot. It was a technical necessity.
Meta built it to handle the explosion of ad volume and generative AI creative scale. Without a new retrieval and pre-filtering system, the auction would slow and the user feed experience would degrade.
So Meta prioritized user experience (ad quality, relevance, and speed) because declining engagement leads to weaker auctions and long-term revenue decline. On social platforms, engagement is the economy.
And spoiler alert: that bet is paying off! Meta reports a modest single-digit increase in ad quality among users since rolling out Andromeda. That translates into billions in additional LTV across the platform.
The reason Meta is pushing creative diversity is because the sheer volume of ads now requires clustering to keep the system efficient. It’s nearly impossible otherwise. Ads only get prioritized if they generate net-new signal.
Each new AI or personalization layer adds compute cost, forcing Meta to balance personalization with ROI.
💡 Understanding these fundamentals is key to knowing what’s really going on in your Meta program. Now:
– Pre-filtering BEFORE auction: Ads are now ranked by predicted relevance and novelty. The first 24–48 hours act as a retrieval test. The model has entered your ad into the race but is still deciding if it should scale. Don’t pause ads too early!
– The BIG ONE: Creative diversity is now a requirement, not a best practice. 50 similar ads might spend like 2–4 used to. This isn’t just a platform update; it’s a resourcing shift. Creative volume and range now define performance. Meta’s own creative frameworks and best practices are built around this new protocol. My biggest concern here is household names with tight brand guidelines will struggle to maintain reach without loosening creative constraints, while performance-driven brands without a clear identity will burn through iterations without memory. This era will favor moving fast AND staying emotionally distinct.
– Emotion > aesthetics: optimizes for what feels relevant, not just what looks good. Meta matches creative to identity and behavior because those signals drive stronger engagement. « Omg I relate » will always hit harder than « that ad looks pretty ».
– Continuous learning = faster ad burnout: the model now updates constantly, prioritizing recency and newness over history. This is why even profitable ads are declining faster. In the Andromeda era, creative pipelines will beat evergreen scaling.
Feel free to leave your thoughts below. Lots to unpack here!