Making more creatives on Meta: volume needs diversity and specificity

I call this method creative fishing. It starts from a sound intuition, and yet it delivers disappointing results, to the point that some brands end up wondering whether Meta is right for them.

This article explains why volume, on its own, covers your market poorly, and what you need to add to it: diversity and specificity.

In short

• Since Andromeda, each creative goes after its own pocket of the audience within your target market. Your goal becomes covering as much of that market as possible.
• Volume helps, provided each creative goes after a different pocket. Close variations of the same creative stay in the same pocket.
• Meta may also group creatives that are too similar and deliver them to the same audience (a mechanism it does not officially document): you then pay for two productions to get a single surface.
• A creative that is too broad reaches a lot of people and converts few. A specific creative reaches a smaller audience and converts far better.
• The method that works combines three requirements: volume, diversity, specificity. It means steering by concept, starting from your personas, their problems and the angle that speaks to them.

Why the volume idea is appealing

The idea rests on solid ground. With Andromeda, targeting is automated: each creative is delivered to an audience the algorithm picks based on its content.

In that system, every new creative is a new chance to reach people you were not reaching yet. The more you make, the more ground you cover. It is mechanical.

Meta itself expects the number of creatives in its recommendation systems to rise sharply, driven by generative AI tools. According to its engineering blog, more than a million advertisers created more than 15 million ads in one month with these tools, and Andromeda was built to handle that volume (Meta Engineering, December 2024).

So far, the reasoning holds. It gets shaky when you look more closely at where each creative goes.

The mechanism: each creative opens its own pocket of the audience

Picture your target market as a large rectangle. It holds everyone who might buy your product.

Each creative you run automatically reaches part of that rectangle. Your creative C1 reaches an audience A1. Your creative C2 reaches an audience A2, which necessarily overlaps A1 in part. For a matcha brand, for example, people interested in matcha can be interested in other things too. Your creative C3 reaches an audience A3, smaller than the first two.

Your creatives open pockets of the audience within your market this way. The goal of the game on Meta becomes covering as much of that rectangle as possible.

From there, one question comes up: which creatives should you make to cover the market? Volume is part of the answer. Three reasons explain why it is not enough on its own.

First reason: close variations stay in the same pocket

The more creatives you have, the more surface you cover. On one condition: those creatives have to be truly different.

If you mostly make variations of your creative C1, which you could call C1 prime, C1 double prime and so on, each one will go after an audience very close to A1. Your volume goes up; the surface you cover barely moves.

The risk is real with creative fishing: by copying whatever is trending, you easily end up with creatives that look alike and land in the same areas of the market.

So on top of volume, you need diversity: creatives different enough to go after audiences in the corners of your target market, where your first creatives do not go.

Second reason: Meta groups creatives that are too similar

A second mechanism reinforces this point. Besides assigning an audience to each creative based on its content, Meta adds a safeguard in its delivery systems, Andromeda and GEM, which Meta describes as a foundation model that improves its other recommendation models (Meta Engineering, November 2025): the entity ID.

The entity ID is an identifier Meta automatically gives each of your creatives. Two creatives can receive the same one.

To be precise: to date, no official Meta documentation describes the entity ID. What Meta does state, in its presentation of Andromeda, is that greater ad diversity can improve the user experience and advertisers’ results (Meta Engineering, December 2024).

Take two simple shapes. A creative shaped like a circle and a creative shaped like a cross go after different audiences, in different places in your market. A triangle pointing up and a triangle pointing down, on the other hand, are too close. Meta can spot that they are two triangles, group them under the same entity ID and deliver them to the same audience.

The result: you paid to produce two creatives to reach exactly the same audience. When the goal is to cover as much surface as possible, paying twice for the same creative surface is waste.

Third reason: a creative that is too broad reaches many and converts few

The third reason is the least intuitive. People talk about it less than volume and diversity, and yet it weighs most on conversion: specificity.

Back to the rectangle. Creative C3 reaches a smaller audience than C1 and C2. To cover the market as fast as possible, you might be tempted to make only creatives that reach gigantic audiences.

The problem lies in your end goal: you want the person to buy, beyond simply getting your ads delivered. A creative that is too broad reaches a lot of people, but its conversion rate will be low.

Take the example of software built for paid media teams.

A first creative would say: “Software for digital advertising in the age of AI.” It would reach a huge audience, part of it even spilling outside the target market. But it would convert very little. Landing two orders with a message like that would already be lucky.

A second creative would say: “I replace the Looker Studio agency media buyers use for their Monday morning presentation.” Presenting the numbers to the client every week is a chore every agency media buyer knows. This creative would reach a much smaller audience. Yet by my estimate, it would sign ten times more customers.

The reason comes down to three precise choices. It speaks to a very precise persona, the agency media buyer. It names a very precise problem, the weekly presentation of numbers to the client. It carries a very precise creative angle, replacing a tool they already use. On its creative surface, it therefore catches many more conversions.

One nuance matters here. A specific creative, on its own, covers a small part of the market. Specificity only works together with the other two requirements: you need many specific creatives, each different from the others, to cover the whole rectangle.

What to do instead: a creative system

The three reasons lead to the same rule. The goal of the game on Meta after Andromeda comes down to three words: volume, diversity, specificity.

In other words, the aim is to cover your entire target market with many small, highly diverse audiences, so you reach every part of it, and to convert as much as possible within each one.

Applying this rule takes what I call a creative system, the opposite of creative fishing:

1. 1You steer your account by concept. Your decisions are made at the level of the creative concept, which brings together a persona, their problem, an angle and a product.
2. 2You know your personas and their problems inside out. That is what makes your creatives specific.
3. 3You know which creative angle to use for each one. That is what makes your creatives different from one another, and lands them in distinct pockets.
4. 4You know which product or service to put in front of each persona. The creative promises what answers their problem.

In this system, every test has a purpose: every test increases your creative surface.

What this method does not guarantee: growth without plateaus. Even when you already steer by concept, the account can seem to hit a ceiling. That ceiling can break, and the brand can keep growing on Meta, as long as it keeps gaining surface in its market.

The counterintuitive move: go after small audiences

When an account wants to reach more people, the reflex is to broaden the message so it speaks to as many as possible. The logic of the rectangle leads the other way: tighten each message around a precise situation, and multiply those messages. To cover a large market, a multitude of small audiences that convert well beats a few huge ones.

The more useful question to ask yourself

The question that moves an account forward: which part of my market will each of my creatives go after, and why would it convert there?

To answer it, ask yourself three questions about your active creatives:

1. 1For each one, can you name the persona, their problem and the angle you chose?
2. 2How many of them are variations of the same idea, and how many truly target a different part of your market?
3. 3Do your broadest creatives convert as well as your most precise ones?

If you answer no to any of them, start there.

Volume fills your account. Diversity and specificity cover your market.