Most brands are still running the 2023 playbook: find a winning ad, spin up 50 near-identical variations, ship them, and let Meta pick the best performer.

Here's the problem: in late 2024, Meta completely rewrote how ads get selected. The algorithm doesn't work that way anymore.

Right now, you've probably got 40+ ads running that the system treats as one single ad. Those ads are competing with each other for the same budget, the same learning curve, and the same tiny slice of your audience. Instead of opening new doors, you're fracturing your spend across duplicate concepts.

But here's what's interesting: brands that fixed this problem saw 20-40% efficiency gains just by reducing their ad count and increasing their concept count. They're shipping fewer ads that actually work, rather than shipping more ads that compete with themselves.

So the fix isn't more ads. It's fewer, structurally different ones.

Every ad you upload gets scanned and fingerprinted based on what's shown, who's speaking, what the text says, and the tone of the messaging.

Here's the trap: multiple different files get assigned the same entity ID, a digital fingerprint that the system reads as the same core concept and design style.

You can change the background, swap out the headline, and adjust the fonts, but you'll still get crushed into the same fingerprint because the system reads your core message, not surface-level tweaks.

Let's look at the real cost here. Say you're launching 8 baby food ads where the same mom holds a baby in the same bright kitchen with the same "natural ingredients" message. You're swapping backgrounds, changing fonts, tweaking the copy words.

To you, that's 8 different tests. To Meta's system, that's 1 concept running 8 times, all competing for the same budget and the same learning phase.

Your Creative Similarity Report will flag this since anything above 60% triggers suppression. Most brands ignore it until they're hemorrhaging efficiency.

But here's what nobody talks about: the algorithm actually deprioritizes duplicate concepts in favor of newer ones. So those other 7 ads aren't just inefficient right now. They're actively being pushed down in favor of genuinely different creatives.

Most accounts run 150+ ads but only have 20-30 actual concepts. You're evaluating 20 different concepts while simultaneously bleeding budget on 130 internal competitors all fighting for the same audience pocket.

Here's how it works: each new concept gets its own learning phase, its own audience distribution, and its own delivery path. But when multiple ads collapse into the same fingerprint, you get one learning phase wasted across duplicate files.

This is why headline rewrites feel like a safe, smart move but often perform worse. You're not actually building new concepts. What you're doing is fragmenting a single concept's efficiency across multiple files.

So here's what you need to do right now: open your Creative Similarity Report in Meta Ads Manager. Anything above 60% means you're hemorrhaging efficiency. Your total ad count doesn't matter anymore. Your concept count is literally everything.

New IDs aren't created by changing colors or rewriting copy. They're created when you build ads that actually solve different problems for your customers.

Instead of thinking about it as "variations," start thinking about it as "different jobs." Because here's the thing: each job unlocks a different audience pocket that you can't reach with your current winners.

So what does this look like in practice? Here are the different jobs ads can do:

Each one of these is a structurally different creative choice, and each one triggers a different concept ID.

This is where Format, Performer, Environment, and Angle actually start to matter:

If your entire account uses the same format and angle, you're basically running one concept in different outfits. So start asking yourself this simple question: can you describe each ad's job in a single sentence? If the answer's no, then it's a variation, not a concept.

This is the framework we run across our client portfolio, which includes Bespoke Post, Cuts Clothing, Topo Designs, MìLà, and Newton Baby.

1. Count concepts, not ads. Before you brief anything, tag your live library across Format, Performer, Environment, and Angle, and count the distinct combinations. If the whole account is one concept in ten hooks, that's the problem, and no number of new hooks fixes it.

2. Build a four-by-four matrix. Pick four formats, four performer archetypes, four environments, and four angles. Every cell is a structurally distinct concept, and every distinct concept is a new thing for the matchmaker to place. That's sixteen real entry points from a single planning cycle. Aim for five fully different concepts a week.

3. Map concepts to buyers Take your top three or four customer types and give each its own concept, hook, and proof point. You're not describing your product more times. You're describing a different person's problem each time, which is exactly the signal Andromeda routes on.

4. Use creators to multiply differences cheaply. One creator shoot can produce three to five structurally different concepts in a day, across environments and angles, while studio production usually gives you one concept in three coats of paint. The math favors creators every time you need a difference, not polish.

5. Refresh on the pocket, not the calendar. When CPMr climbs, you've saturated a pocket, not just fatigued an ad. Ship a genuinely new concept to open a new one instead of re-cutting the tired winner. The system is the asset. The individual ads are just what it outputs this week.

Do this right now, it'll take 10 minutes and could change everything.

If 70% or more of your ads share the same format and angle, you're basically running one concept, not 50. Every impression on those duplicate 49 is wasted reach that's not opening new customer pockets.

Two metrics will tell you immediately whether you're creating genuine new concepts:

Hook rate or 3-second views divided by impressions: When you're hitting new audiences with new concepts, this should clear 30%. If you're stuck somewhere between 8 and 12%, that's a sign you're recycling the same concept to the same people over and over.

Cost per thousand reach, known as CPMr: When CPMr climbs, you've saturated that audience pocket. When it stays flat or drops, that means your creative is actually reaching fresh, qualified people.

Here's the counterintuitive part that most brands get wrong: a new concept that acquires customers at 15% higher CPA but is actually hitting completely new segments is worth significantly more than your best performing ad re-converting repeat buyers you would've gotten anyway.

That's not inefficiency you're looking at, that's legitimate reach expansion.

And honestly, most brands are optimizing the wrong metric entirely. They chase ROAS on the winner instead of finding new winners. The difference in scale between those two approaches? We're talking 2 to 3 times growth over six months.

1. Kill the ad-count metric. Instead of shipping 50 ads that collapse into 5 concepts, ship 5 ads that stay 5 concepts. Start measuring concepts per month instead of total ads shipped.

2. Count distinct jobs before you even brief anything out. Ask yourself: "Does this solve a different problem than my existing ads?" If the answer's no, kill it. Also, if Format, Performer, Environment, or Angle stays the same across three different ads, then you're making variations, not concepts.

3. Build a quick matrix. Four formats times four performers times four environments times four angles gives you 256 possible combinations. You're not gonna fill all of them, but pick 5 or 6 genuine concepts per week. Don't try to maximize. Just diversify.

4. Map each concept to a specific buyer type, not just demographics. The customer who buys for health reasons gets one concept. The customer who buys for status gets a completely different one. Different problems solved equals different audience pockets.

5. Use creators to multiply your concept variety. One creator shoot can produce 3 to 5 structurally different concepts in a single day across different environments and angles. Studio production usually gives you 1 concept in 3 different coats of paint. If you need difference rather than polish, creators scale way faster.

6. Refresh based on saturation, not based on your calendar. When CPMr climbs, you've exhausted that pocket, not just fatigued that specific ad. Ship a genuinely new concept to open a new pocket instead of re-editing the tired winner.

We're doing free creative concept audits this month for brands that're spending $50K or more on Meta monthly.

We'll pull your active library, show you exactly which ads are collapsing into duplicate fingerprints, quantify your actual wasted budget on variations, and then send you a one-page brief on the concepts you're missing.

Here's what we usually find: typically 60 to 70% of your ads are duplicates. And that's not really a testing problem at that point. That's a 25 to 35% efficiency leak that's just sitting there.

Twelve audits are available right now. First come, first served.

You'll walk away with:

Three slots open this month. First come, first served.

Until next time
✌️, Conrad