Why Andromeda Means The Old Meta Playbook Is Dead

For years, paid social performance on Meta platforms lived and died by targeting. Interest stacks, behaviour layering, ever more precise audience definitions. Creative mattered, of course, but largely as a supporting act. If targeting was tight enough, even fairly average ads could still perform.

That balance has now shifted.

With the rollout of Meta Andromeda, Meta has fundamentally changed how ads are selected, matched and delivered. The platform now places far greater weight on creative signals when determining who sees what, using machine learning to analyse visuals, formats, pacing, language and structure at scale.

In practical terms, this means the creative itself is doing far more of the work that targeting used to do.

What Andromeda Is Actually Doing

Rather than relying primarily on audience definitions, Meta’s delivery system is increasingly interpreting creative content to determine relevance. Video construction, hook style, visual energy, messaging angle and format choice all feed into how ads are distributed across users.

If two ads are conceptually similar, even if the copy or colours differ slightly, Meta is likely to treat them as near-identical signals. That limits exploration and restricts where spend goes. On the other hand, distinct creative approaches give the system more room to learn and match ads to different motivations and behaviours.

This is why broad targeting structures have become more effective over the past year. The system needs freedom, and Andromeda rewards advertisers who give it meaningful creative variation rather than tight audience constraints.

Why Most Creative Testing Is Still Falling Short

Where many brands struggle is not in understanding this shift conceptually, but in how cautiously they respond to it.

Minor adjustments are often mistaken for proper testing. A rewritten headline. A different CTA. A colour change. These tweaks feel productive, but from the system’s perspective they rarely represent a new idea. The algorithm sees variations of the same thing and learns very little as a result.

Under Andromeda, creative testing needs to move beyond refinement and into contrast. The question is no longer which version of this ad works best. It is whether the platform has been given enough genuinely different signals to discover new pockets of performance at all.

Here's an example that would no longer work under the update: identical ads, testing nothing more than the sub-heading. Once a strong creative performer, now one ad would just be unshown entirely!

What Effective Creative Testing Looks Like Now

Proper testing today requires more volume, more intentional diversity and a faster learning cycle.

Instead of polishing one core idea, brands should be launching multiple distinct concepts at once. Different messaging angles. Different emotional drivers. Different formats and energy levels. Each creative should exist for a reason, aimed at a particular motivation or moment in the customer journey.

This does not mean lowering standards or flooding campaigns with weak work. It means accepting that variation, not perfection, is now the primary input the system needs to optimise effectively.

In this environment, creative diversity is not a luxury. It is the mechanism by which Meta finds performance.

An example of a campaign famjam previously ran that would be more in line with how the latest update would expect you to be producing creatives - multiple images, multiple copy variants, and multiple variations on layout.

Structural Simplicity Supports Better Learning

Campaign complexity often works against this new reality. Multiple ad sets segmented by interest or micro audience can restrict delivery and slow learning.

Simpler structures tend to perform better. One campaign per objective. Broad targeting. Fewer ad sets. A wider creative mix running simultaneously. This allows Meta to allocate spend based on creative performance rather than audience constraints and gives Andromeda the signal depth it needs to do its job properly.

The optimisation no longer happens between audiences. It happens within creative.

How We Approach This at famjam

This shift has not changed how we work as much as it has confirmed why our approach has always worked.

famjam has always been a creative-led agency, particularly when it comes to Meta ads. Long before Andromeda, we favoured ideas, angles and format diversity over hyper-specific audience builds. We have never believed performance comes from cosmetic tweaks or endlessly refining the same ad. It comes from giving the platform something genuinely different to learn from.

Creative is not layered on after campaign setup. It is the starting point. We build around motivations, use cases and moments, not just products or offers. We focus on contrast between concepts rather than small visual or copy changes, and we treat testing as an ongoing system rather than a fixed phase.

Performance is judged on outcomes, not surface-level engagement, and learning speed matters just as much as efficiency.

Andromeda does not demand a new philosophy. It rewards the one we have been using all along.

Closing Thought

Meta Andromeda has made one thing very clear. Audience-first advertising is no longer the dominant model. Creative-led performance is.

Brands that cling to old structures and cautious testing will feel like results have quietly slipped away. Those willing to embrace wider experimentation, stronger ideas and faster iteration will find Meta still works extremely well.

The advantage now belongs to the bold testers, not the careful tweakers.