“Creative Is the New Targeting” – But Only If You Structure It Right

Everyone running Meta ads has heard it by now. “Creative is the new targeting.” And the core idea is correct: in a broad-delivery world, the creative you run determines who responds, which determines who the algorithm optimizes toward.

But saying “creative is the new targeting” and actually structuring your account so creative functions as targeting are two different things. Most advertisers got the first part. They stopped stacking interests, retired their lookalike ladders, and leaned into broad. Then they dropped 15 creatives aimed at completely different buyers into one ad set and called it a strategy.

What that actually gives the algorithm is noise. Mixed signals, no coherent optimization direction, and a budget that drifts toward whoever is cheapest to reach.

TL;DR:

• “Creative is the new targeting” is correct, but only when each ad set is built around a single coherent signal (one persona or one theme).
• Advantage+ audiences removed the structural separation that options like interest and lookalike targeting used to provide. If you’re using Advantage+, you have to rebuild that separation through creative grouping.
• Mix signals inside one ad set and Meta optimizes toward the cheapest audience – not your highest-value buyer.
• More ad sets means you need more creative. Match your ad set count to what your budget can realistically fund through the learning phase.
• The structure scales with budget: fewer signals when you’re lean, more as you grow.

What Broad Delivery Actually Broke

Audience-level targeting did structural work that nobody talks about anymore. Before Advantage+ audiences became the default, those settings separated your buyers for you at the ad set level.

If you had an ad set built around the “Yoga” interest and another around “Marathon,” each one forced Meta to find people inside those pools. The creatives inside each ad set only had to resonate with one type of person, because the audience settings had already done the sorting.

Advantage+ audience settings removed that built-in separation. With Andromeda’s full rollout in 2025, Meta pushed even more weight onto creative signals and reduced the impact of audience-level inputs. The algorithm now decides who sees what based largely on what the creative tells it about who will engage.

That shift was the right call for the platform. The problem is that the industry told people what to stop doing without explaining what to replace it with. Interest stacking and lookalike ladders used to impose structural discipline on your account. When those went away, the discipline was supposed to come from creative. For most advertisers, it never did.

One Ad Set, One Signal

Each ad set should be built around a single coherent signal. The principle is straightforward, even if execution takes real work. That signal is either a specific persona, or a theme strong enough to function as its own targeting.

Every creative inside that ad set should speak to the same type of buyer, reinforce the same angle, and push the algorithm in one direction. The algorithm still finds whoever it wants within Meta’s full delivery pool. You’re not restricting reach. You’re giving the system something consistent to read so it can optimize toward a specific response pattern rather than averaging across contradictory ones.

There are two ways to build that signal, and which one you use depends on how your buyers cluster.

When buyers share a situation

Persona-based ad sets work when your buyers cluster around a shared lifestyle, situation, or mindset. The key word is mindset, not demographic. Demographics are what the person looks like on paper. Mindset is what makes them buy.

Say you sell a magnesium supplement. Consider two women who are both 32, both living in a mid-size city, both active, both earning roughly the same household income. On paper they look like the same customer. But one of them heard about magnesium from her naturopath after months of poor sleep. She sees a creator video where someone talks about their nighttime routine, feels an immediate “that’s me” recognition, and adds to cart within 30 seconds. The other found the same supplement through a fitness subreddit, cross-referenced three clinical studies on bioavailability, and won’t convert until she’s seen a comparison ad that breaks down dosage, form, and third-party testing.

Same age, same zip code, same lifestyle on the surface, same product. Same intent, even – both want magnesium for sleep. What differs is the mindset they bring to buying it. That difference is the persona. The first buyer needs a creator-led story that mirrors her situation. The second needs an information-dense breakdown that satisfies her research process. If both creatives live in the same ad set, the algorithm has no way to cleanly optimize toward either. It splits the difference, and your CPA suffers for it.

A persona-based ad set picks one of those mindsets and fills the ad set with creative variations that all speak to it. Different hooks, different formats, different visual styles, but the same core buyer.

When intent crosses demographics

Theme-based ad sets work when one hook cuts across demographics entirely and the shared trait is buying intent rather than lifestyle. Think about marketers who use AI tools. They could be 24-year-old freelancers running a one-person content shop or 52-year-old VPs of marketing at a mid-market SaaS company. B2B, B2C, agency-side, in-house. Nothing about their demographic profile overlaps. But they share the same buying intent: they want to move faster, produce more, and reduce the cost of creative output. A theme-based ad set groups creative around that shared intent and lets the algorithm find the buyers across whatever demographic lines they happen to sit on.

The choice between persona and theme isn’t permanent or exclusive. Most accounts will use both, depending on the product and the buyer landscape.

The rule is simple: if your buyers share a situation, use persona. If they share an intent that cuts across situations, use theme.

The Ad Set Still Does

The ad set is where Meta commits to a signal.

Inside any given ad set, Meta is only optimizing toward people who respond to that ad set’s creatives. Mix two buyers into one ad set and Meta optimizes toward the cheaper one. Split them into two ad sets and each one gets its own optimization, even if budget moves between them.

Most accounts should run CBO. Audiences differ in size – a high-intent niche persona and a broad theme can’t be budget-balanced by hand, and Meta is better at finding where the volume sits than you are. The cost of that flexibility is that cheap signals can starve expensive ones. Say you’ve got a $24 moisturizer ad set and a $120 clinical serum ad set in the same campaign. Meta will favor the moisturizer all day – more abundant audience, cheaper to reach. Your serum ad set quietly suffocates.

The lever for that is minimum spend rules. Floor-protect the ad sets carrying signals you don’t want Meta to abandon – high-AOV products, B2B in a B2C-leaning account, or any persona where the audience pool is smaller but the conversion value is higher. Set the floor at whatever your ad set needs to clear the learning phase, and let Meta optimize from there.

ABO still has its place – small accounts that want predictable per-signal spend, or testing setups where you’re forcing equal-weighted reads. But for most scaled accounts, CBO plus minimum spends is the more honest structure: it lets the algorithm do the allocation work it’s good at, while you protect the signals you can’t afford to lose.

The Campaign Structure Didn’t Change

The campaign-level structure that works hasn’t actually changed. If you’re expecting a new framework, you can relax. One testing campaign for new creative. One scaling /evergreen campaign for proven winners. That’s still the foundation.

What changed is what happens inside those campaigns. Ad sets used to be organized by persona or theme through audience-level settings. You’d have an ad set for “wellness-focused buyers” built on a supplement-purchaser lookalike, and another for “performance athletes” targeting interests like “CrossFit” or “Cycling.” The audience setting did the organizational work.

Now the same organizational logic has to come from the creative itself. Your testing campaign still has multiple ad sets, but each one is organized around a creative signal rather than an audience setting. Same structure, different mechanism for separation. If you’ve been naming your ad sets by audience type all along, you can keep the names. The ad set labeled “overwhelmed new parents” just doesn’t have a parenting interest audience attached anymore. Instead, it has 10 to 20 creatives that all speak to overwhelmed new parents, and the algorithm figures out who those people are based on who engages.

More Signals Mean More Creative, and Budget Has to Match

This is where the math gets uncomfortable, and where most teams feel the gap between knowing what to do and actually doing it. In the interest-targeting era, you could run 3 to 5 creatives per ad set and get clean signal because the audience was already pre-filtered. Post-Andromeda, the practitioner consensus has shifted toward 10 to 20 distinct variations per ad set. Manual interest-based ad sets are an exception and can still run leaner, but for broad delivery, volume matters.

More personas and themes mean more ad sets. More ad sets mean more creative. This scales directly with budget, and the relationship between ad set count and budget is where most advertisers make their biggest structural mistake. The math is the same whether you’re running ABO or CBO with minimum spends – each ad set still needs enough daily spend to actually learn from.

On a small budget, the temptation is to build out every persona and theme you can think of. Four personas, three themes, seven ad sets. The problem is that each ad set needs enough daily budget to exit the learning phase and generate meaningful signal. Split a $200/day budget across seven ad sets and you’ve given each one roughly $28/day. Every ad set starves, nothing optimizes, and the whole account underperforms.

The better approach on a small budget is fewer ad sets built around your single strongest signal. One or two ad sets, fully funded, with enough creative volume to give the algorithm room to optimize. As budget grows, you expand by adding new persona or theme-based ad sets, each one funded at a level where it can actually learn. The structure follows budget, not the other way around.

On the other end, under-splitting on a large budget creates the chaos problem from the beginning of this article. Plenty of money, all of it dumped into a handful of ad sets with mixed signals. The budget is there, but the structure isn’t directing it.

The discipline here is matching your ad set count to what each one can realistically be funded to learn from. Get that ratio right and the rest of the structure follows.

Creative Is the Targeting. The Ad Set Is Where It Happens.

“Creative is the new targeting” was never wrong. What was missing is the second half of the sentence: creative is the new targeting when the ad set enforces the signal. Without that structure, you’re just hoping the algorithm figures out what you meant. The targeting happens at the ad set level, through the intentional grouping of creatives that speak to one buyer, one mindset, or one intent.

Most teams I talk to understand the framework. The bottleneck is almost always production. They know what to build, they just can’t produce enough of it fast enough to keep every ad set fed with 10 to 20 fresh variations on a regular cycle.

That’s the problem Billo is built to solve. Real creators, UGC-style video ads, at the volume you actually need to run this structure. If creative is your targeting, your creative pipeline is your targeting pipeline, and it needs to keep up.