We sat down with a marketing scientist from Meta, and he told us your video ad creative, not your audience settings, is the thing deciding whether your ads work now.

For years, winning on Meta was a targeting game. You found the right audience, layered on the right interests, and narrowed it down until you were sure you were only paying to reach the exact person who'd buy. The creative mattered, but the real skill was supposedly in the settings.

We had Tommy Changaris on the show, a Marketing Science Partner at Meta, and he said the quiet part out loud. That game is over. Targeting is automated. Bidding is automated. His number: about half of auction outcomes are attributable to the creative itself. So the biggest lever you have left is the thing most brands still treat as the last step.

Creative is the new targeting.

You don't tell the algorithm who to find anymore. You put that in the creative.

Here's the shift: in the Advantage+ era, you're not supposed to hand the algorithm a tiny predefined audience and tell it exactly who to serve. You make an ad that speaks to a specific person, and you let the system go find that person, plus everyone like them you never would have thought to target.

Tommy's example was a beach blanket brand. The old way: guess your buyer is young and single, make one ad, aim it at 25-year-olds who like the beach. The new way: make one concept of single people playing frisbee, and another of a family building a sandcastle, and let the algorithm route each one to the people it fits. You stopped describing your audience in the settings. You started putting them in the creative, the casting, the location, the props, the dialogue.

Which means the creative actually has to be different. And that's where most ad budgets quietly bleed: brands spend real money producing "variations" that the algorithm reads as the exact same ad.

Swapping the CTA is not a test

Ask most brands for "creative diversity" and you get fifty versions of the same ad. Different font. Different button. "Learn More" becomes "Shop Now." The opening line changes. To the algorithm, that's the same ad. It can't use it to reach a different kind of person, because nothing different is actually happening. Tommy's version: it isn't materially or structurally changed enough for the system to treat it as a different ad, so it never gets served to a different person.

Real diversity is the structural stuff: the hook, the motivator, the format, the casting. Same beach blanket, but one ad is about how easily it packs, one is about how good it looks, one is about how fast the sand shakes off. Those reach different people. A new button color does not.

We've been saying this from the production side for years, and honestly it was nice to hear a data scientist at Meta say it back to us. A brand asks for "a couple variations" and what they mean is the same spot with the text swapped in the first three seconds. We used to just make it, because it's what they asked for. Now we gently suggest another approach: let’s play with the concept, the casting, or the motivator in pre-production. Because If it isn't structurally different, the algorithm won't treat it as different ad, and it won't do the job.

Here's the catch: creative is now your bottleneck

This is the part every growth lead should sit up for. If creative is the lever, then your ceiling isn't your ad budget or how clever your targeting is. It's how much genuinely different creative you can get into the market before your current winner burns out.

And it will burn out. Tommy calls creative fatigue a lagging indicator, which is a polite way of saying: by the time your ROAS visibly drops, you've been leaving money on the table for weeks. The brands that lose find one winner, run it into the ground, and only then scramble to make the next thing. The brands that win already have the next batch ready before the current one dies. His rule of thumb: refresh the concepts every two to four weeks.

Now wait. We can already hear the panic about how anyone is supposed to pay for all this. Hang on, we're getting there.

Because the real question stops being "who do we target" and becomes "how do we keep making genuinely different creative, forever, without it eating us alive." That's not a targeting problem anymore. It's a production problem. And you cannot solve it with the production model most brands still run on: one big hero video, once a quarter or year.

"That sounds expensive." It's the opposite.

The first thing a brand says is… that it sounds expensive. It's the opposite. The expensive option is the one big shoot that makes one ad, because that one ad is a single bet, and when it doesn't perform, you spent the whole bag finding that out.

This isn't about spending more. It's about sizing the production to the goal. If the goal is volume and variety, you build the shoot around volume and variety. You plan multiple concepts into a single day. You pull stills off the video shoot and turn them into static ads, because a video is just a lot of frames and some of those frames are a great photo you already paid for. You keep the raw footage instead of tossing it, so you can cut new variations fast from the excess the second the data tells you what's working.

Think of it like buying dinner. The old way, we hand you a finished pizza. One ad, done, hope you're hungry for exactly that. The new way, we pre-chop all the ingredients into little bowls, and you can make ten different meals out of one trip to the store. One shoot day, sized right, gives you several video ads, cutdowns, extra clips, iPhone content, and a set of stills. That's not one asset. That's a real stack of tests out of a single production. We built our whole production model on this: months of content from a few shoot days, because the point was never the shoot. It was the system. And the system builds on itself, growing your ability to make new content over time.

Tommy gave it a name we're going to steal. The brands that win aren't the ones with the cleverest audience settings. They're the ones with the best creative operating system.

The brief has to carry all of this now

Last episode we said your video brief is lying to you, because it describes what to make instead of who it's for and what it has to do. This is the same fight, one floor up. If the creative is the targeting, then the customer, the motivators, the strategy, all of it has to be built into the concepts from the start, not bolted onto the ad account later. The insight that used to live in your targeting settings now lives in the creative. That's a pre-production decision, not a media-buying one. So put it in the brief.

The takeaway

Stop trying to out-target the algorithm. It's better at that than you are now. Your job is to feed it a steady stream of genuinely different creative, and to build the production system that can keep making it.

Creative is the new targeting. So the smartest thing you can build isn't a better audience. It's the production system that keeps the creative coming.

‍

Full episode with Tommy Changaris on YouTube, and everywhere you get podcasts: https://share.transistor.fm/s/aa9642fd