I’ve been asked this over and over during audits lately. And I don’t think there’s one percentage that works for every account. What I can share is how we think about it.
First question: do you actually have a winning concept to iterate on?
To me, a winner isn’t an ad that had a good ROAS on $300 of spend. A winning concept is one that earns spend and continues to deliver as it spends.
If you have that, I’m very comfortable putting roughly 50% of the creative testing budget into iterations. On some accounts, 60%.
But iteration doesn’t mean duplicating the winner 15 times and changing the headline. It means understanding why it won and expanding from there:
- Same persona, new hook.
- Same objection, different creator.
- UGC into stills. Stills into carousels.
- Same message, different execution.
And ideally, a landing page that continues the exact story that got the click.
But we still need net-new concepts
Even your best winners fatigue. You keep serving variations of the same idea to the same audience long enough, and eventually you’ve squeezed everything you can out of it.
I’d like to see roughly 30% going toward genuinely new concepts, built around the personas, objections, and motivations we’re seeing in reviews, customer conversations, performance data, and previous winners.
And then, when the budget allows, I like keeping another 10% or so for true new production. Not another edit of something we already have. Go shoot something new. New creators, new environments, new raw material to work with.
This is actually one of the reasons we built our in-house studio. We can now go from an idea to shoot to finished assets in 1-2 days without turning every new idea into a $10K production.
Treat the 60/30/10 as guideposts, not a formula. The exact split isn’t the point. The point is knowing what you’re iterating, what you’re learning, and what you’re deliberately trying to discover next.
