How Meta’s Andromeda Algorithm Changed Creative Strategy Forever
Something shifted in late 2024 and most advertisers felt it before they understood it.
Campaigns that had been running profitably for months started declining. Funnels that had converted reliably stopped performing. A simple creative refresh, the kind that used to restart a flagging campaign, either didn’t work at all or worked briefly before performance dropped again. Ad fatigue arrived faster than it ever had before.
If you experienced any of that, you weren’t imagining it. And it wasn’t bad luck.
It was Andromeda.
What Andromeda Actually Is
Andromeda is Meta’s ads retrieval system, announced by Meta’s own engineering team in late 2024. It’s the mechanism that decides which ad gets shown to which person, and it works fundamentally differently from what came before.
The old system leaned heavily on audience signals. You told Meta who to target, and it found those people. Your targeting choices did a significant portion of the work. Creative mattered, but a well-targeted ad with mediocre creative could still produce reasonable results.
Andromeda shifted that relationship. The system now reads the creative itself, extracting signals from your visuals, copy, audio, and engagement patterns, and uses those signals to find the right audience. You’re no longer primarily telling Meta who to reach. Your creative is telling Meta who to find.
This sounds like a subtle technical change. The practical implications are anything but subtle.
What Changed in Practice
The shift showed up in account behaviour before most people had a name for it. Here’s what it looked like on the ground:
Campaigns fatigued faster. Creative that used to run for months started declining in weeks. The same ad, shown to the same broad audience, exhausted its effectiveness much more quickly than before.
Simple refreshes stopped working. The old playbook for a declining campaign was straightforward: duplicate it, let the algorithm reset, no creative updates needed. You could usually do that a few times before it stopped working.
That approach either stopped working entirely or produced a brief bump before performance dropped again. The algorithm needed something genuinely different, not a slightly adjusted version of the same thing or simple restart.
Funnels that had worked reliably started underperforming. This was the most disorienting part for many advertisers. Nothing had changed in the funnel or ad account. The targeting was the same, the budget was the same, the creative was the same, the offer was the same. But results declined anyway. Because the system had changed how it was reading everything.
Variations of the same concept stopped testing as separate ads. This is the piece most people still don’t understand. If you produce five ads that are slight variations of the same concept, same hook, same visual style, same core message, Meta’s algorithm reads them as essentially the same ad and treats them accordingly. You think you’re testing five different things. You’re actually testing one thing five times.
The Misconception That’s Costing Advertisers the Most
When Andromeda’s implications started becoming clear, the industry’s response was largely: produce more creative. More volume, more variations, more AI-generated copy, more assets in the account.
That response misunderstands what the algorithm actually rewards.
Andromeda doesn’t reward volume. It rewards diversity. Those are not the same thing.
Volume means more ads. Diversity means more genuinely different ads, different hooks, different visual styles, different angles, different personas, different proof formats, different emotional entry points.
If you produce 40 ads that are slight variations of the same concept, Meta reads weak, repetitive signals and treats your creative pool as one undifferentiated asset. You’ve spent the time and budget producing 40 ads and gotten the distribution benefit of roughly one.
If you produce 10 ads that are genuinely, strategically different from each other, Meta reads 10 distinct signals and can match each one to the specific people most likely to respond to it. You’ve gotten ten times the targeting surface area from a quarter of the creative volume.
The goal was never more ads. It was always more distinct creative thinking. Andromeda just made the cost of ignoring that distinction much higher.
What “Genuinely Different” Actually Means
This is where the practical work happens, and where most creative strategy falls short.
Genuinely different doesn’t just mean a new headline on the same visual. It doesn’t just mean the same video with a different first three seconds. It doesn’t just mean running your existing creative in three different aspect ratios.
You can do those things and it might help some, but genuinely different means different at the strategic level. Different persona. Different angle. Different proof format. Different emotional entry point. Different visual world entirely.
Here’s how this plays out in practice.
Think in Personas, Not Variations
Every business has more than one type of ideal client, even if you think you only have one. The coach who helps women build online businesses has at least three distinct personas inside that broad description: the corporate professional who wants to leave her job, the stay-at-home parent returning to work, and the existing entrepreneur who’s plateaued. Each of those people has different fears, different objections, different language, and different proof requirements.
When you build creative around a single generic “ideal client,” you produce ads that speak adequately to everyone and compellingly to no one. When you build creative around specific personas, each ad speaks directly to one person’s exact situation. That specificity is what Andromeda reads as a strong signal.
In my own client accounts, the shift from “variations of a theme” to “genuinely persona-driven creative” produced two measurable outcomes: cheaper leads and slower creative fatigue. The ads lasted longer because they were reaching different people with different messages rather than exhausting the same audience with the same message repeatedly.
Think in Visual Worlds, Not Templates
The second dimension of genuine creative diversity is visual. Most businesses find a template that works and stick to it. Same layout, same colour palette, same general aesthetic, different text or image swapped in. It feels efficient. It’s actually limiting.
Andromeda reads visual signals as part of its creative matching process. If all your ads look broadly similar, the algorithm reads similar signals across your entire creative pool. You’re not giving it much to work with.
Testing truly different visual styles, a polished branded graphic versus a raw talking-head video versus a text-heavy static versus a UGC-style piece shot on a phone, gives the algorithm genuinely different signals to match to genuinely different people. Some of your ideal clients respond to polished and professional. Others respond to authentic and unproduced. You don’t know which until you test formats that are actually different from each other.
The combination of persona-driven copy and visually distinct formats is what produces a creative pool that Andromeda can actually work with. Each ad is a different signal. Each signal finds a different subset of your ideal audience. The whole becomes greater than the sum of its parts.
The Strategic Shift This Requires
Andromeda doesn’t just change how you produce creative. It changes how you think about creative strategy from the ground up.
Before: Build one strong campaign with a proven creative concept, optimise it, scale it, refresh it when it fatigues.
After: Build a diverse creative ecosystem where multiple distinct concepts run simultaneously, each speaking to a different persona or angle, with a continuous pipeline of genuinely new creative thinking rather than variations of existing winners.
This requires more strategic thinking upfront, not more production volume. You need to know your personas well enough to build truly different creative for each one. You need to be willing to test visual styles that feel uncomfortable or unfamiliar. You need to accept that the creative that worked last quarter may not work this quarter, not because you did something wrong, but because the audience has seen it and the algorithm has exhausted its signal.
The businesses that are winning on Meta right now are the ones that treated Andromeda as a signal to think harder, not produce faster.
What This Means for Your Creative Process
If you’re currently running Meta ads, here are the practical changes to make based on how Andromeda actually works:
Audit your creative pool for genuine diversity. Look at everything currently running. If you can see a clear template or pattern across most of your ads, you don’t have a diverse creative pool. You have one concept in multiple formats. That’s not what the algorithm needs.
Map your personas before your next creative round. Before briefing any new creative, write out at least three distinct personas within your ideal client profile. Give each one a name, a specific situation, a primary fear, and a primary desire. Then build at least one ad concept specifically for each persona rather than one generic concept for all of them.
Test visual styles, not just messages. Your next creative test should include at least one format that looks genuinely different from what you’re currently running. If everything you run is polished and branded, test something raw and unproduced. If everything is static, test video. The goal is to give the algorithm new signals, not refined versions of existing ones.
Extend your creative planning horizon. Because Andromeda rewards diversity and penalises repetition, you need a longer creative pipeline than you did before. Thinking one creative round ahead is no longer enough. Map out your next three rounds of creative concepts before you launch the current one, so you’re never scrambling for new ideas when performance drops.
Stop measuring creative success only by ROAS. A creative that produces a strong ROAS in week one and collapses in week three is not a winner. A creative that produces a slightly lower ROAS but maintains performance for eight weeks is far more valuable. Factor longevity into how you evaluate creative performance, because longevity is a direct signal of how well the creative is working with the algorithm rather than against it.
The Bigger Picture
Andromeda didn’t make Meta ads harder. It made lazy creative strategy more expensive.
The fundamentals that have always driven strong Meta ad performance, deep buyer understanding, specific messaging, genuine creative diversity, strategic thinking before production, are now the only things that reliably work. The shortcuts that used to paper over weak strategy, narrow targeting, lookalike audiences, minor creative refreshes, no longer do.
For coaches, consultants, and service businesses, this is actually good news. You were never going to win a volume war against DTC brands with dedicated creative teams and six-figure monthly production budgets. But you can win a strategy war. Because strategy requires thinking, not headcount.
The businesses that understand their buyers deeply enough to build genuinely persona-driven creative, and that are willing to test visual worlds rather than just message variations, are the ones Andromeda rewards. That’s always been the right way to run Meta ads. The algorithm just made it the only way.
For the full framework on how to build a creative strategy that works with Andromeda rather than against it, start with the pillar guide: Meta Ads Creative Strategy in 2026: The Complete Guide for Coaches, Consultants, and Service Businesses.
If you want to understand exactly how creative strategy differs from the media buying side of your account, What Is Meta Ads Creative Strategy? And Why It’s Different From Media Buying covers that in full.
And if you’d like to talk through what this means specifically for your ad account, book a free 20-minute Clarity Call and we’ll look at what’s running, what the algorithm is reading, and where the real opportunity is.
Julie Lowe is a Meta ads creative strategist and founder of Socially Aligned. She has 26 years of marketing experience and specialises in helping coaches, consultants, and service businesses build Meta ad creative systems that generate quality leads without requiring massive budgets or endless content production.