Meta is constantly changing. In Higher Education, universities have had to adjust to new restrictions on targeting people under 18. This started in 2021 when many interest and audience targeting options were removed, and it continued in 2023 with even more limitations.

These changes have directly impacted how campaigns are planned and implemented, but while we’ve focused on what we can no longer do, something much more important has been developing in the background.

That something is Andromeda.

Andromeda: What is it?

Don’t worry, this isn’t a new platform that you need to plan for or activate.

Andromeda is Meta’s AI system that determines how ads are delivered. It has been quietly evolving for some time, but by 2026, it has fundamentally changed how campaigns operate.

It reflects the direction Meta has been pursuing for years:

• Broader audiences
• Less manual targeting
• Greater reliance on automation

Essentially, moving from telling the platform what to do to providing it with the tools and insights to figure it out for itself.

In simple terms: You no longer control who sees your ads, you influence it by what you give Meta to work with.

What’s Important Now? Feeding the Machine

If Andromeda is the engine, your job is to properly fuel it.

In 2026, there are three things matter more than anything else: creative, signals & feedback loops.

We’ve moved away from controlling delivery and towards influencing it.

For anyone sitting just outside of paid media, or even those close to it but not in the detail day to day, it’s very easy to default back to the old mental model. The one that says performance comes from tightening audiences, refining targeting, and making incremental adjustments to campaigns.

That way of thinking has worked for a long time. It’s what a lot of teams have been taught, and it’s why many stakeholders still expect to see clear segments, defined audiences, visible manual control over how ads are served.

It feels logical and it feels safe, but the challenge is that the platforms no longer operate in a way that rewards that approach.

Not just updated best practices documents but tangible experience tells us that you can build the most tightly defined audience possible, but it doesn’t carry the same weight it once did. In some cases, it can even limit performance.

Creative: Your New Targeting Tool

In 2026, creative equals targeting.

With Meta stripping back audience controls, it now relies on your creative work to determine who should see your ads.

That means…

• Ads that mention parents or supporting your child are more likely to reach parents.

• Ads that focus on accommodation, nightlife, or student life are more likely to attract younger viewers.
• Ads that showcase career progression or flexibility are more likely to appeal to older or postgraduate audiences.

The algorithm is interpreting context, not just demographics.

In practice, your messaging defines your audience more than any targeting settings ever could. If your ads all look the same, Andromeda treats them the same, which limits delivery and increases costs.

Signals: Quality In = Quality Out

At Cubed, this is exactly where our paid media teams are focusing.

We’re spending less time refining audience segments and more time strengthening the signals we feed into platforms. That means working closely with CRM teams to ensure lead quality is clearly defined and consistently fed back, so optimisation is based on what actually drives value, not just volume.

This has always been important, but you can no longer hide behind tight targeting. The quality of your signals directly impacts the quality of your outcomes.

Old approach:
“Send users to a course page and let Meta figure it out.”

New approach:
“Specify what meaningful engagement looks like for Meta and optimise for that.”

Instead of optimising for volume, you’re optimising for intent. Andromeda learns from these behaviours and finds more people who match, but that cuts both ways.

If you feed it the wrong signals, it won’t just underperform, it will scale the wrong behaviour.

First-Party Data: Your Competitive Edge

Meta’s algorithm relies on high-quality, deterministic data, and nothing is better than what you already have.

First-party data ranges in value from highest-quality CRM data (like email addresses and applicant lists), through high- and mid-intent actions (such as open day registrations and prospectus downloads), all the way down to lower-value signals like basic social engagement.

The higher quality data you have, the better your audience will be.

What Can You Do to Improve?

Your campaigns in 2026 operate under completely different rules than five years ago.

And they will continue to evolve.

What this means for you is being open to change and recognising that older strategies, like tight targeting and heavy manual optimisation, are becoming less effective. Success now comes from focusing on inputs rather than controls.

You won’t win by endlessly tweaking audiences, but by improving what you provide to the platform.

That means maximising signal quality through better tracking, more considered event setups, and stronger alignment between marketing and your CRM.

It also means investing properly in creative. Not just increasing volume, but developing meaningful variations that genuinely reflect the motivations of your audience.

Final Thoughts

Meta isn’t the only platform rewriting the rules & what works this year won’t necessarily work next.

The teams that succeed will be the ones that bring data, creative, and marketing closer together, adapting quickly and building campaigns around stronger inputs, not legacy tactics.

If you’re rethinking your approach or want a fresh perspective on how to improve performance, let’s have a conversation.

Josh

Paid Media Director