Your New Statics Are the Same Ad

Here’s a test worth running on your own account. Pull up the last ten statics you launched and sort them by concept, not by design. Most brands get to one. Maybe two.

Different headline, different background, different product shot, same persona, same angle, same offer. That’s one ad, ten times. And it matters more than it used to, because Meta’s delivery system has started making the same judgement.

What a concept actually is

A concept is three decisions.

1. Persona. Who are we talking to?

2. Angle. What are we saying to them?

3. Offer. What are we putting in front of them?

Format is the fourth layer: the mechanism you deliver it through. A listicle, a before-and-after, a screenshot of a text message, a founder holding the product. A format on its own isn’t a concept. It’s the vehicle the concept travels in.

Change one of the first three and you have a new concept. Change nothing but the design and you have a variation.

Take a magnesium sleep supplement. Persona: women in their forties who wake at 3am. Angle: it’s cortisol, not caffeine. Offer: a 30-night trial. Swap the persona to shift workers and you have a new concept. Swap the angle to “you’re taking the wrong form of magnesium” and you have a new concept. Change the pouch from green to navy and you have the same ad in a different jumper.

Why the background colour test is a waste of money

Suppose blue genuinely converts 2% better than red for your brand. Great. Now try to prove it.

To detect a 2% relative lift on a 2% conversion rate with any confidence, you need somewhere around two million sessions per variant. At a £40 order value, that is millions of pounds of spend to learn which colour to use. Nobody has that budget for that question, which means the test will be called early on noise, and the “winner” will be whichever one got lucky in week one.

A test worth running passes three checks:

• Measurable. The change is big enough that the result won’t be noise.

• Fundable. You’d happily put real budget behind the answer.

• Useful. The answer changes how you brief the next batch.

“Use blue” fails all three. “The 40+ persona responds to the longevity angle and ignores weight loss entirely” passes all three, and you’ll still be using it in six months.

What Meta changed

Since Andromeda, Meta’s creative-led delivery system, the creative does the targeting. Each ad finds its own audience based on what’s in it. Which means two ads that look almost identical find the same audience, and Meta has decided to stop pretending otherwise: near-duplicates get grouped, counted as one ad, and served to the same people.

The reason, as far as anyone outside Meta can tell, is the AI image flood. When generative tools got good, some accounts started dropping thousands of near-identical variants a day to game the platform’s preference for high creative velocity. Grouping lookalikes was the fix.

The consequence for the rest of us is simple. A slight variation used to be weak. Now it’s invisible.

Variations are still worth doing

None of this means stop varying. A concept that’s working should be squeezed, because it’s the cheapest volume you will ever get into an account. The point is that a variation has to be big enough to register: a different opening line, a different visual metaphor, a different proof point. A hue shift is a rounding error.

And when a format lands, resist the reflex to make ten more versions of it. Rotate an element of the concept instead and keep the format fixed. Problem-agitation worked on the greens powder? Run the same format, same angle, same persona, on the protein. Same format, same offer, but now you’re talking to men instead of women. That’s how you get scale out of a winner without Meta quietly folding it back into one ad.

The same logic works across brands. A format and angle that’s already working for someone else is a head start. Keep the structure, rebuild it around your product, your customer and your offer, and you have a new concept for your account with some of the risk already taken out.

The brief

Put four lines at the top of every static brief: persona, angle, offer, format. If the first three match an ad already live in the account, it’s a variation. Label it as one and set your expectations accordingly.

For a batch of ten statics a month, a sensible default is three or four concepts with two or three variations each. One concept in ten outfits is the default most accounts are running, and it’s why “we tested loads of statics and none of them worked” usually means “we tested one static, repeatedly”.

The recap

• A concept is persona, angle and offer. Format is the vehicle it’s delivered in.

• Change one of the three and it’s a new test. Change only the design and it’s a variation.

• Tiny variations can’t be measured, aren’t worth funding, and tell you nothing about the next brief.

• Meta now groups near-identical ads and serves them to the same people. Slight variation adds nothing.

• Scale a winning format by rotating persona, offer or angle, not by making ten more versions of it.

• A format already working for another brand, rebuilt around yours, is a new concept with a head start.

• Four lines at the top of every brief. If three of them match a live ad, say so.

If you’d like a monthly batch of statics modelled on what’s already working for other brands, rebuilt around yours, get in touch.