Variants Aren’t Dead, But They Can’t Be Your Strategy Anymore

There’s a new wave of advice sweeping paid social: variants are dead.

The story goes that Meta’s evolution (Entity IDs, creative grouping, Andromeda) has rendered creative variants obsolete, that the only way to scale now is with genuinely different concepts, and that hook testing or refining winning ads is wasted effort.

It’s a neat narrative, but it’s wrong.

Understanding the nuance is what separates brands that scale profitably from those stuck spending <£50k/month because they’ve hit a creative ceiling.

The brands that win aren’t the ones throwing hundreds of clones into Meta. They’re the ones balancing conceptual diversity with thoughtful refinement, understanding where to push new ideas and where to optimise existing ones.

— Daniel Watts, Co-Founder, BARK

What Actually Changed (And Why It Matters)

Meta’s rollout of Entity IDs and creative grouping under Andromeda is the biggest shift in creative delivery since the pixel. The platform now fingerprints ads far more aggressively: if two creatives look or feel too similar, Meta treats them as one, same learning, same predicted performance, same audience delivery.

This is why lazy variants no longer work.

Previously, near-duplicates, swapping a headline, changing a background colour, could scale independently. Not anymore.

Conceptual diversity now separates winners from losers. Different formats, angles, people, settings, energy. If your creative library blurs together, Meta sees it the same way and you’re stuck delivering to the same pocket of users while CAC climbs.

Concepts matter more than ever, but the “variants are dead” narrative misses the point.

Why “Concepts Only” Is Also Wrong

Since Andromeda’s global rollout, some advice has swung the other way: forget variants entirely. Every ad should be a new concept. No hook tests. No refinements. Pure conceptual diversity.

It sounds sophisticated but it also leaves performance on the table.

Concepts don’t arrive fully formed as winners, they often need refinement. The hook (the first three seconds), how you frame the problem and how the ad delivers the message all determine whether a concept converts at scale.

A concept without variants is like going fishing with only one type of bait. Different concepts are like dropping lines in different parts of the ocean but the variants are the types of bait you use. If you only used one type of bait in a location and didn’t catch anything you might wrongly assume that there are no fish there.

If you’re not testing any hook variants, you’re depriving your carefully-developed concept of it’s best chance of succeeding.

— Alex Smith, Head of Growth, BARK

Where Variants Still Win

1. Testing “Hooks” on Video and UGC Ads

The hook (the first 3 seconds of a video ad) is the highest impact variable you can test in a video ad. The same concept, creator, and script can perform completely differently depending on the opening.

Example: one version opens with a creator holding the product up and asking a surprising question that addresses a common pain point, immediately grabbing attention. Another version opens with the same creator simply explaining the benefit in a calm, straightforward way. Same concept. Same story. Only the hook differs.

The question-led version becomes a long-term top performer. The straightforward version barely registers in testing. That difference? A variant, not a new concept and it’s the difference between a winning ad and one that fails to gain traction.

2. Refinement of Winning Concepts

When a concept shows promise, give it a chance to scale. Test different framings, benefit hierarchies, or delivery approaches.

Launching only one version risks writing off potential hero ads prematurely. Winning brands test distinct ideas, then give each enough variants to succeed through thoughtful refinement.

Even small changes, a different visual cue, phrasing, or energy, can uncover the version of a concept that truly connects.

3. Low-Lift Optimisation for Static Creative

Headline swaps or background tweaks matter less than they used to but they’re extremely low effort. If you can generate 10 headline variants in minutes, do it. The lift may be small, but the cost is negligible.

These micro-variants aren’t your strategy; they’re tactical optimisation on top of genuine conceptual diversity.

The Framework That Works in 2026

1. Launch as many genuinely distinct concepts as your capacity allows. Different formats, angles, people, settings, problems solved. These drive reach and learning.
2. Give each concept at least three variants. For video and UGC, focus on hooks. For statics, try headline or framing variations. This ensures each idea has a fair chance to perform.

Think of it this way: each concept gives the potential to reach a distinct pocket of your addressable market. Within that pocket, variants give it the best chance to convert.

30 concepts × 3 variants = 30 potential audience segments reached. 10 concepts × 9 variants = only 10 potential segments.

Both of these outputs deliver 90 ads, the first has the potential to reach 3x as many people. Volume matters, but it’s the genuinely distinct ideas that unlock new reach and learnings, not the total number of ads.

What to Ignore (And What to Watch Out For)

Be wary of tools promising “hundreds of ads in minutes.” Most just repurpose variant spam for Andromeda. Minor edits, different backgrounds, cropped images, rewritten headlines, don’t create true diversity.

Meta sees these as the same ad: same learning, same predicted performance, same audience. They don’t unlock new reach; they crowd your account. Even 200 versions may only count as a handful.

Conversely, don’t swing to “only concepts, no variants.” Skip variants, especially hooks, and you’re doing 90% of the work but missing the 10% that turns a good concept into a winner.

The Bottom Line

Variants aren’t dead, but they can’t be your entire strategy.

Conceptual diversity is the foundation, it unlocks reach, learning, and scale. Variants are the mechanism that refines and pressure-tests those concepts to find their highest-performing expression.

Brands that win in 2026 will:

• Launch genuinely distinct concepts.
• Give each concept thoughtful variants.
• Focus on hooks, framing, and execution, not just cosmetic changes.

The goal was never volume for volume’s sake. It’s connecting with customers effectively. Tactics have gotten smarter.

Creative diversity wins:

• Different formats (video vs static vs carousel)
• Different angles (core messaging pillars)
• Different people, settings, and energy

Meta’s Creative Similarity and Creative Fatigue metrics show when advertisers are fooling themselves, thinking they’re testing ideas when they’re really not.

Squint at your creatives: if they all look similar, Meta sees them the same way. Volume isn’t the goal, difference is.

BARK builds creative strategies that scale for modern performance brands. Want to see where your creative approach is leaving money on the table? Get in touch.