Meta's advertising ecosystem has just taken another significant step. On 31st March 2026, its engineering blog introduced the Adaptive Ranking Model, a redesign of its ranking system intended to work with LLM-scale models and respond in milliseconds.

Beyond the technical, there's an interesting piece of data: according to Meta, this system has already been implemented on Instagram and, in its early tests, achieved a +3% in conversions and a +5% in CTR.

Put simply: Meta is getting better at deciding which ad to show, to whom, and when.

If you work with paid campaigns, whether within a company or from a marketing agency in Barcelona, this isn't a technical curiosity. It's a pretty clear signal of where things are headed: less reliance on manual adjustments and more emphasis on creativity, quality data, and actual business signals.

To make it easy to understand, think of two phases.

First, Meta decides which adverts “make the cut”. This is where everything related to systems like Andromeda comes into play, filtering among thousands of options in record time.

Next comes the important part: deciding which advert is actually shown.

The ranking phase

This is where the Adaptive Ranking Model comes in. It's the system that decides which ad wins priority.

According to Meta, this new architecture allows for working with much more complex models, with levels approaching a trillion parameters, without losing speed.

The key is to adjust the calculation level according to each user and context.

Practical translation

Previously, a large part of the result depended on how you set up campaigns.

Now, it depends more and more on how you feed the system:

• What data do you give it
• What creativity do you use?
• What business signals do you receive

Why this matters to businesses and marketing agencies

In markets like Barcelona, where competition on Meta Ads is high, any improvement in the ranking system directly affects costs and results.

If Meta distributes impressions better based on intent, creativity and performance, brands that work well on those variables will have an advantage.

Change of focus

This implies a quite clear twist:

• Less obsession with segmentation tricks
• More focus on truly distinct messages
• Greater importance of well-measured data
• More need for simple structures

From the agency's perspective, we're seeing that technically optimised accounts with no strong creative element are increasingly underperforming compared to simpler but better-thought-out accounts.

The big idea: Meta better understands intent

One of the most relevant changes is how Meta interprets user behaviour.

The new system can analyse longer sequences of interaction: what content someone consumes, how they navigate, whether they are browsing or ready to buy.

What does this mean

It's no longer just about striking at the right time. It's about understanding where the person is at.

And that completely changes how you should approach your campaigns.

Think about the journey, not just the impact

It's starting to make more sense to work with pieces for different phases:

• Discovery
• Consideration
• Test
• Conversion
• Trust

One single creative trying to do everything... it works worse and worse.

What truly changes in creativity?

Here is one of the key points.

With more advanced systems, making many almost identical versions loses its purpose.

If the message is the same, the learning is limited.

What starts to work

What really adds value is creating distinct concepts:

• A piece focused on the problem
• Another one in the solution
• Another on social proof
• Another one on offer
• Another one on objections

The greater the real difference, the easier it will be for the system to detect what works for each user.

What we see in practice

In the projects we manage, when we move from “aesthetic variations” to “strategic variations”, performance improves quite clearly.

It's not magic. It's giving the algorithm useful material.

The data matters more, not less

A common error is thinking that with more automation, tracking loses importance.

Actually, the opposite is true.

Meta continues to bet heavily on Conversions API, which allows sending more reliable data from the server side.

Current context

Furthermore, with privacy restrictions, such as Apple's, relying solely on cookies or weak signals is becoming increasingly unreliable.

Without solid data, the system has less context.

And without context, it optimises worse.

What does this imply

• Measure events well
• Avoid duplicates
• Align metrics with real business

What should companies do

This is where it all comes down to earth.

1. Review which conversion you are optimising
   Don't optimise for vanity metrics. The system will pursue exactly what you tell it to.
2. Simplify campaigns
   Less fragmentation usually aids learning.
3. Create creativity with real variety
   No more same versions. More distinct approaches.
4. Improve data signals
   Pixel, Conversions API and well-defined events.
5. Thinking in sequences
   An advertisement no longer lives alone. It is part of a journey.

If you need support, you can work with a complete strategy from our agency: Contact | Karmina – Marketing Agency in Barcelona

What does this mean for a company in Barcelona

For local businesses, e-commerce, or startups, the message is quite clear: Meta Ads is looking less and less like “pressing buttons” and more like building a system.

In a city like Barcelona, this is noticeable quickly.

• Brands that operate on intuition have a harder time.
• Those who work well with creativity and data have an advantage

That's why, today, the value of a digital marketing agency in Barcelona is no longer just in launching campaigns. It's in aligning strategy, creativity, and measurement.

Conclusion

Adaptive Ranking Model is not a new visible feature. It's a background change.

Meta officially launched it on 31 March 2026, and initial data points to real performance improvements.

But the important thing is not the system. It's how you adapt to it.

In 2026, winning on Meta Ads will depend increasingly on three things:

• Truly distinct creativity
• Well-measured data
• Campaigns that allow the system to learn

That's where you start to notice the difference between investing... and scaling.

Does your company want to improve results in Meta Ads without relying on blind testing?

At Karmina, we help you build campaigns with better strategy, better data, and creativity designed to convert.