Google and Meta’s New Advertising Algorithms

From audience targeting to value proposition.

Digital advertising rested on a simple assumption for a long time:
Reach the right person, and the message already works.

Today, that assumption no longer holds.

Ads no longer play to the audience, but to intent.
Google and Meta’s new advertising algorithms position the audience not at the centre of the decision mechanism, but merely as a data layer.

Rather than defining the user by fixed interests;
it bases itself on their real-time behaviour, their response to content and the relationship they form with the message.

This shift is not just a new algorithm update; it is a structural break in the fundamental logic of advertising.

What are the algorithms doing differently?

Meta: the Andromeda approach

The new system Meta calls Andromeda is radically changing how ads are pulled into the candidate pool before the auction.
Rather than defining the user by fixed interests, this structure:

• Their real-time behaviour
• Their content consumption pattern
• The micro-interactions they form with visuals and text, evaluating them together to select the most fitting ad message.

As a result:

• The same ad takes on different meaning for different users
• Creative quality and message fit take precedence over targeting settings
• “Narrow audience” strategies can slow down learning

This is why Andromeda has been interpreted across the industry as “the manual targeting era is closing.”

Google: AI Max, Demand Gen and Performance Max

On Google’s side, the transformation is progressing more quietly but more systematically.

With AI Max for Search:

• Search campaigns automatically open up to broader queries
• Text assets are dynamically restructured
• The single “correct keyword” approach is losing its meaning

Demand Gen and Performance Max, meanwhile:

• Evaluate image, video and text assets together
• Automatically optimise the format and message match according to user intent
• Reward signal quality rather than campaign structure

Google is now optimising meaningful content, not the campaign.

Why don’t static visuals and one-size-fits-all messages work?

The new algorithms assume the following:
The user may be the same person, but their intent can change at any moment.

A static visual, or copy offering the same promise to everyone:

• Cannot respond to different intents
• Does not generate enough signal for the algorithm
• Restricts the learning process

This is why most performance drops stem not from budget, but from message architecture.

The core concept of the new era: value proposition

The value proposition, not the product or the campaign, is now at the centre of the ad.

A strong value proposition gives a clear answer to the following questions:

• Which problem does it solve?
• What concrete benefit does it give the user?
• Why does it matter now?
• Why is this brand trustworthy?

Algorithms learn by testing these answers across different creative variations.

What do successful brands do differently?

1. They distil the value proposition into a single sentence

Then they multiply this sentence from different angles:

• Economic benefit
• Time saved
• Risk reduction
• Proof of performance
• Ease of use

2. They produce a creative portfolio instead of a single creative

• Multiple opening messages
• Different formats (UGC, demo, explainer, proof-focused)
• Different narrative tones

But they all convey the same core value.

3. They simplify their campaign structures

Few but clear campaign structures:

• Faster learning
• Cleaner signal
• More consistent results

This approach supports the algorithm’s power, it doesn’t clash with it.

What has changed for SEO and AEO?

The line between ad copy and content is increasingly blurring.

In the age of AEO (Answer Engine Optimization):

• Copy should be written like an “answer”
• Headings should be clear and contextual
• The landing page should match the ad’s promise exactly

Because both Google and Meta use page content and message consistency as an active signal.

The Globalmeta perspective

Today, performance marketing:

• Not by spending more budget
• Not by producing more formats

is won with a better thought-out message architecture.

As Globalmeta, we define this shift as follows:

The era of ad optimisation is over. The era of message engineering has begun.

Brands now need to ask themselves this question:
“Why does this ad deserve to be shown?”

Brands that can answer this question clearly don’t fight the algorithms;
the algorithms work for them.