Meta has fundamentally changed how advertising works on its platform, and most marketers are still playing by the old rules.
After years perfecting audience targeting at Airwallex—building detailed interest stacks, lookalike audiences, behaviour targeting—I was sceptical when Meta shifted to a creative-first approach. It felt reckless to strip back everything that had worked.
Then we tested it.
We ran a dental marketing campaign with one simple hook: “If you’re a dental clinic doing over 70k per month, then this is for you.”
No audience layering. No interest targeting. Just broad campaigns with specific creative.
Our cost per qualified lead dropped 60%.
What Actually Changed
Meta’s shift to AI-driven creative analysis has fundamentally rewired how the platform matches ads to people.
The old model prioritised who you targeted. The new model prioritises what your creative signals.
Meta’s AI now analyses every element of your creative—visuals, copy, soundtrack, context—to determine user intent and match ads accordingly. The platform can pattern-match creative elements across billions of users in real-time.
This means your creative does the targeting work.
The Binary Qualifier Framework
Here’s what I learnt from that dental campaign.
I tested three hooks:
“If you’re a dental clinic doing over 70k per month, then this is for you.”
“Struggling to fill your appointment book consistently? Here’s what’s missing.”
“Most dental clinics waste thousands on marketing that doesn’t work. Here’s why.”
The revenue qualifier destroyed the others. It got 40% more plays past the 6-second mark.
The difference? Specificity that creates self-selection.
Dentists running smaller practices scrolled past. The ones hitting that revenue threshold stopped because we were speaking directly to their situation.
The pain point hooks performed okay. But they were too broad. Every dental clinic thinks they have marketing problems.
The revenue hook cut through because it was binary. You either are or you aren’t a dental clinic doing 70k per month.
Meta’s AI learnt who engaged with that specific language and found more of them.
What Makes A Strong Signal
I look for specificity that forces self-selection.
The signal needs to be concrete enough that the right person thinks “that’s me” and everyone else scrolls past.
Revenue numbers work. “Doing over 70k per month” is binary.
Time in business works. “Been operating for 5+ years” is binary.
Specific problems at scale work. “Managing a team of 15+ across three locations” is binary.
Regulatory situations work. “Navigating GDPR compliance for client data” is binary.
Tech stack mentions work. “Currently using Salesforce but struggling with HubSpot integration” is binary.
What doesn’t work? Emotional states or generic pain points.
“Feeling overwhelmed with marketing?” is useless. Everyone feels that way.
“Spending over £10k monthly on Google Ads with declining returns?” That’s a signal.
The test I use: could this apply to half your target market or just a specific segment?
If it’s half, it’s too vague.
The Broad Campaign Paradox
Most marketers panic when I suggest broad targeting.
They think specificity in messaging means shrinking their audience.
The maths tells a different story.
Would you rather reach 100,000 people where 50 might be interested, or reach 10,000 people where 500 are interested?
Specificity doesn’t shrink your audience. It focuses your spend.
You’re not losing reach. You’re losing waste.
That dental campaign reached fewer people overall. But our cost per qualified lead dropped 60% because we weren’t paying to show ads to single-chair practices or start-ups doing 20k per month.
According to a 2025 AppsFlyer report, 70-80% of Meta ad performance now stems from creative strength, not targeting precision.
The platform’s AI is brilliant at scale. Once it learns who engages with “dental clinics doing over 70k per month,” it finds more of them across the entire platform.
You’re not limiting reach. You’re giving the AI crystal clear instructions.
How We Structured The Campaign
We stripped back everything we would have normally done.
No detailed interest stacks. No lookalike audiences. No behaviour targeting.
Instead, we ran broad campaigns and let the creative do the heavy lifting.
The hook “if you’re a dental clinic doing over 70k per month” was front and centre in the creative itself.
We built multiple variations around that same intent signal. Rather than creating 15 different ad sets with narrow targeting parameters, we had one broad campaign with diverse creative formats all signalling the same audience through the messaging.
Meta’s AI picked up on who was engaging with that specific language and found more of them.
It was simpler to manage and performed better.
The Eight Creative Formats That Work
Meta’s AI responds particularly well to specific creative formats.
User-generated content and testimonials. Real people speaking directly to camera about specific outcomes.
Product features with clear visuals. Show exactly what you’re offering with no ambiguity.
Comparative content. “Us versus them” messaging that highlights specific differences.
Founder stories. Personal narratives that signal expertise and authority.
Face-to-camera explainers. Direct address that creates immediate connection.
Carousels showcasing range. Multiple products or services in one scrollable format.
Statistics for credibility. Hard numbers that prove specific claims.
Before and after transformations. Visual proof of specific outcomes.
The key is mapping these formats to different buyer journey stages.
Early awareness needs broad hooks with specific qualifiers. “If you’re X, this is for you.”
Mid-funnel needs proof. Testimonials, case studies, comparative content.
Late-funnel needs clarity. Product features, pricing, specific next steps.
The Resource Reallocation
This shift changes where you invest time and money.
Previously, you’d spend hours building complex audience segments. Testing different combinations of interests, behaviours, demographics.
Now, that time goes into creative development.
You need diverse asset libraries that signal different intents. Multiple hooks testing different qualifiers. Various formats addressing different journey stages.
The campaign structure becomes simpler. The creative library becomes richer.
Early adopters of Meta’s Advantage+ creative features saw a 22% increase in ROAS according to Meta’s internal data.
The platform rewards creative diversity now more than targeting precision.
Building Your Intent Signal Library
Start with an audit of existing assets.
Look at your current creative. What signals does it send?
If your messaging could apply to half your target market, it’s too vague.
Identify the binary qualifiers in your audience. Revenue thresholds. Team sizes. Tech stacks. Regulatory situations. Years in business.
Build hooks around each qualifier.
“If you’re a fintech company processing over $1M monthly in transactions…”
“If you’re managing a dental practice with 3+ locations…”
“If you’re spending over $50k annually on Google Ads…”
Each hook becomes a creative variation. Same core message, different entry point.
Test which qualifiers drive engagement past the 6-second mark. That’s your signal that Meta’s AI is finding the right people.
The Campaign Structure That Works
One broad campaign.
Multiple creative variations, each with specific intent signals.
Let Meta’s AI do the matching.
Meta recommends 8-15 creative variations per ad set. Each variation should signal a different qualifier or address a different journey stage.
Track which creative drives qualified leads, not just clicks.
The AI learns from engagement patterns. When someone watches past 6 seconds, clicks through, and converts, that’s a strong signal.
The platform finds more people who match that engagement pattern.
What This Means For You
Stop building complex targeting stacks.
Start building diverse creative libraries.
Your hooks need binary qualifiers. Revenue numbers. Team sizes. Specific problems that only happen at scale.
Your campaign structure needs to be broad. One campaign with multiple creative variations beats multiple narrow campaigns.
Your resource allocation needs to shift. Less time on audience research. More time on creative development.
The marketers who win in the Andromeda era will be the ones who understand this fundamental truth: specificity in creative creates focus in spend.
You’re not shrinking your audience by being specific.
You’re giving Meta’s AI the clearest possible instruction about who to find.
And when the instruction is crystal clear, the AI does remarkable work.
