When an algorithm goes from scanning thousands of ads to evaluating tens of millions of ad candidates in milliseconds, the game changes overnight. If you’ve felt like Meta’s ad auction has been behaving like a completely different beast lately, you aren't imagining things. The days of outsmarting the algorithm with lookalikes or granular audience targeting are officially dead.
Meta has quietly been rolling out "Andromeda", its massive new AI retrieval engine. Radically rewriting the rules of paid media.
Although it was rolled out in late 2025, this year has been focused on trial and error to try to understand what is best practice for what we deem as one of THE most significant evolutions in paid media.
In simple terms, Andromeda completely flips the old media buying playbook upside down. For years, the formula was simple: you manually defined an audience, and Meta served your ads to them.
Andromeda reverses this logic. Instead of matching ads to predefined audiences, the AI reads hyper-subtle, real-time behavioural signals, like the exact visual style a user is stopping on right now, and matches individuals to your creative. Meaning Meta now builds a full behavioural pattern of a single user and is now doing the targeting for you. Therefore, your ad creative itself has officially become your primary targeting tool.
This news is perfect for those of us who love the creativity that comes with digital advertising! Think of it this way: your ad creative is no longer just a billboard trying to catch someone's eye - it is the literal steering wheel of your entire campaign. By launching a mixed portfolio of static imagery, user-generated content (UGC), and founder stories, you are handing Andromeda the exact data signals it needs to self-optimise. Meta is pushing for creative diversification.
This structural shift isn’t just an engineering milestone - it’s a massive evolution in behavioural science.
For over a decade, digital tracking forced marketers to treat consumers like static data, locked into rigid demographic boxes (e.g., "Target females, aged 25–34"). But human psychology is fluid; our mindset shifts constantly based on stress levels, time of day, and immediate emotional triggers. Andromeda is the first ad infrastructure built to adapt to this psychological fluidity in real-time.
Instead of matching ads to static customer personas, the AI tracks immediate micro-behaviours - like the split-second speed of a thumb scroll or how long someone lingers on a specific visual aesthetic and matches the user's immediate psychological state to your creative.
As Meta's AI-native architecture now handles audience selection, delivery, and real-time optimisation, the algorithm no longer needs human instructions on who to target; it requires human strategy on what to say. The ultimate growth lever has officially shifted from media buying to the creative production pipeline.
Consequently, our role as Paid Media specialists is to guide our clients away from the obsession over old-school technical levers and lead them toward utilising these new features successfully. Winning in this era means shifting client conversations away from lookalike percentages and redirecting their energy into building a robust, structured "test-and-learn" creative pipeline.
Now that the rules have changed, how do you actually structure your ad account to give Andromeda exactly what it wants? To stop fighting the algorithm and start scaling, you need to transition your paid strategy into a creative-centric framework.
The number one killer of performance under Andromeda is account fragmentation. When you split your budget across lots of hyper-targeted ad sets, you are actively restricting the machine learning model from optimising efficiently.
Under the old playbook, you used different ad sets to target different customer segments. Under Andromeda, you use different creative formats inside the same campaign to achieve that exact same goal.
Because Andromeda favours historical data and existing social proof, mixing brand-new creatives into an active scaling campaign is a recipe for failure. The machine will naturally dump the majority of the budget into the proven winners, starved of the chance to gather data on your new concepts.
We have officially entered the AI-Era of digital performance marketing. The rollout of Andromeda signals a future where Meta Ads is no longer a manual optimisation environment. Meta has shifted to predictive, behavioural personalisation. Moving forward, the brands that scale seamlessly won't be the ones with the most complex account settings, but the ones with the most agile, psychology-driven creative pipelines. The algorithm has taken over the logistics of distribution so we can finally return to what makes advertising work: true human storytelling & creativity!
Creative is the New Targeting: Traditional targeting and lookalike audiences are dead! Under Meta's Andromeda AI retrieval engine, your ad creative itself serves as the primary steering wheel to attract and build your target audience.
Adapting to Psychological Fluidity: Rather than relying on static demographic boxes, Andromeda evaluates tens of millions of ad candidates in milliseconds by tracking fluid human behaviors, like thumb-scroll speed and visual aesthetic linger time.
Consolidate to Avoid Fragmentation: Splitting budgets across multiple hyper-targeted ad sets suffocates machine learning.
Diversify Your Formats: Reach different customer segments within the same campaign by utilising a portfolio of visual formats, including static imagery, UGC, carousels, and video.
Establish a Structured Testing Pipeline: Because Andromeda favours existing social proof and historical data, avoid mixing new creatives directly into active scaling campaigns. Launch a dedicated "test-and-learn" pipeline to successfully surface new winners.
Want to see how your current paid strategy stacks up against the Andromeda framework? Contact our team today for a comprehensive creative and account structure audit, and let's unlock your next phase of scale.
Contact Us
Mya is a Senior Paid Media Executive at Reflect Digital, specialising in crafting and optimising high-performing PPC and paid social campaigns. She blends data-driven insights with creative thinking to deliver results that truly make an impact, while staying ahead of trends in the ever-changing digital landscape. Collaboration is at the heart of her approach, working closely with teams and clients to ensure every campaign not only meets but exceeds expectations. Always hungry to learn, Mya is committed to growing her expertise while championing innovation, curiosity, and excellence.
More about Mya