How many ads does a launch need?

For a typical launch, we aim for somewhere between 12 and 27 versions of an ad. Not hundreds, and definitely not one.

Back when linear TV was how people saw ads, one spot was enough. You made a single commercial and placed it on different channels. That was the whole plan.

Today there are far more places to run ads, and each platform needs more to work with. That doesn't mean you need tons of unique creative. It means you need variations of a strong idea.

Why one hero ad fails on its own

A single hero ad is one guess. You're betting that this exact cut, with this exact opening, is the thing that will work. And making ads is expensive, before you even count the media spend.

Here's the good news. Once the upfront production is done, making variations and testing them costs very little. So instead of putting one guess out there, you can test a handful of hypotheses at once.

Give them a week or two. Then move the spend to the versions that are actually working. You'll get better results than you would by betting everything on one film.

A new ad versus a variation

Take a hero film. From that one shoot, you might make:

• A long version and a short version
• Different aspect ratios for different placements
• A version with a different hook at the start
• A version with a different end card

Each of those is a variation of the same idea. The platforms need that range to see which one connects with the audience best.

The math we use

Here's how we usually structure a launch:

• Three core concepts, so there's always an alternative to compare against
• Two to three aspect ratios for each
• Two to three lengths for each

That works out to roughly 12 to 27 versions. It's enough to learn which concept works best, at which length, in which format.

One note on counting. On Meta, the different aspect ratios of one cut usually sit inside a single ad, and the platform picks the right one for each placement. So the number of ads you upload can be smaller. The number of versions you need doesn't change.

Why more isn't always better

AI makes it much easier to produce versions. You can edit faster and resize in minutes. It opens up a lot of possibilities.

But if you flood the system with hundreds of versions, you dilute your spend. Most of those ads won't get enough budget for anyone to really see them. Worse, you'll have so many variables in play that you won't know why something worked.

The goal isn't the most ads. It's enough ads to learn something.

Keeping it all on brand

With a dozen or more versions running, the risk is drift. The fix is a central concept. Every version should ladder back to your positioning and the reason your brand exists. If an ad has nothing to do with that, there's no point putting it out there.