In our day-to-day agency work, capturing the attention of the right user at the right time has become one of our biggest challenges. That's why, when Meta introduced AndromedaWe knew we were in for a major change in the way we managed campaigns for our clients.
What is Andromeda de Meta?
Andromeda is the new AI-based ad system of Meta. Its aim is to optimise in real time which ad each user receives.
Unlike traditional engines, which rely on predefined rules and manual segmentations, Andromeda uses deep learning and high-performance data architecture to process millions of options in milliseconds and choose the most relevant one for each individual.
In our agency we see it as particularly useful because it is designed to work with:
- Massive user behaviour data.
- Multimodal creativity (text, image, video).
- Automated testing of thousands of creative variations.
Why it is a technological leap
Meta has not only adjusted its algorithm, it has also created a new infrastructure to support it. Among what impresses us most:
1. Hierarchical indexing architecture
It enables large-scale storage and retrieval of advertisements without sacrificing speed, which is essential for a network that processes more than 4 billion monthly active users.
2. Multimodal processing
Andromeda is able to analyse advertising content in text, image and video, understanding the context and intent behind each element.
3. Continuous learning
The system does not "reset" with each campaign, but rather constantly learns of previous interactions, optimising subsequent advertising decisions.
4. Extreme scalability
The engine is designed to process ads efficiently, even in global campaigns with millions of creative variations.
What it means for advertisers
The arrival of Andromeda implies a profound change in the way advertising campaigns are approached in Meta.
1. Less manual segmentation
Instead of narrowly defining audiences, the engine recommends broader campaigns for the AI to automatically find the users most likely to convert.
2. More focus on creativity
The key will be not so much manual targeting, but to offer multiple creatives (images, videos, text) so that the system can determine which works best for each micro-segment.
3. Real-time optimisation
Changes and improvements do not depend on weekly reviews, but are adjusted in milliseconds based on user behaviour.
How to leverage Andromeda in your strategy
If you want to get the most out of this engine, we recommend you follow these practices:
1. Use Advantage+
Andromeda works particularly well in conjunction with Meta Advantage+The system automates the placement and combination of creatives.
2. Multiply your creativity
Create at least 5-10 variations of ads per campaign: different texts, colours, formats and calls to action.
3. Provides a variety of formats
Includes short video, still images, carousel and text-overlay ads to cover all modes of consumption.
4. Monitor, but don't over-optimise.
Avoid changing creatives or audiences too early. Let the AI collect enough data to optimise.
Initial results of its implementation
According to internal data and beta advertiser testing, early results are very promising:
- +6 % to +8 % in advertisement quality (more relevance and engagement).
- +7 % in conversions average.
- +22 % in ROAS (Return on advertising investment).
- Reducing the time needed to optimise campaigns by 35 %.
In our agency it is still too early to see radical changesbut the trend is clear: intelligent automation outperforms manual adjustment.
Winners, losers and the shift in audience control with Andromeda
The arrival of Andromeda not only redefines the way Meta manages advertising, it also redistributes "power" between advertisers and the platform itself. Like all disruptive innovation, it brings clear winners... and some who will have to rethink their strategy.
Who wins
- Advertisers with a strong creative approachAndromeda: Those who produce multiple variations of high quality ads (video, image, text) see how Andromeda automatically finds the most effective combination for each user.
- Brands with medium to high budgetsThe volume of data accelerates AI learning and improves the accuracy of its recommendations.
- End-usersThey receive more relevant and personalised ads, with less saturation of irrelevant content.
Who stands to lose
- Small advertisers or very specific nichesIf the target audience is very small, the AI may take longer to find it, spending part of the budget on testing.
- Managers relying on ultra-detailed segmentationAutomation reduces the scope for manual "fine-tuning" of the targeting.
- Highly regulated sectorsIn industries such as health or finance, creative and targeting freedom may be limited by regulation.
Is there a loss of control over the audience?
Yes, but it is a change of control rather than an absolute loss. Andromeda reduces the advertiser's power to define exactly how they want to advertise. who you will see an advertisement, and puts the focus on what advertisement display and how to present it.
In practice:
- Less control over the exact audience.
- More influence on the quality and variety of creativity.
In this new scenario, the key is not so much to know the demographics of your audience down to the last millimetre, but to produce content that is sufficiently varied and attractive for the AI to serve it to the right people at the right time.
Timeline for implementation
- Beta phase (gradual implementation)Meta started to implement Andromeda in phases from January 2025. Some advertisers received the system already in January, others in March and April.
- Full global implementationThe beta launch began in January, but it was in June 2025 when full adoption was achieved across all advertising accounts.
Conclusion
All in all, although Andromeda's early results are promising, it is still too early to draw definitive conclusions.
In our experience, it is becoming increasingly difficult to run fully manual campaigns and to maintain detailed control, especially in small accounts or very specific niches.
Personally, I miss being able to segment with much more precision. and fine-tune every detail of the campaign, something that was previously more accessible.
In this new scenario, the key will be adapt, rely on intelligent automation and focus on creativity.We are learning to balance the power of AI with the strategic vision that only we, as an agency, can bring to the table.
