Your Creative Is Your Targeting Now

You spent years learning to layer interest targeting, narrow lookalikes, and pick the perfect audience. Quietly, Meta has been making that work matter less, and last October it finished the job. The buyers crushing it right now noticed, and they stopped touching the targeting settings.

There was no email titled "Targeting Is Now Dead." There was no announcement on stage. But anyone reading their Ads Manager honestly over the last year has felt it. The lever you used to pull to find your audience has been quietly disconnected. Something else is doing the targeting now, and it is sitting right in front of you.

What Andromeda Actually Did

Andromeda is Meta's deep-learning ad retrieval system. The name barely escaped engineering blog posts. The impact is everywhere. Meta started rolling it out at the end of 2024 and finished the global rollout last October, which means every Meta advertiser on the planet is now sitting on top of it.

What it actually does is read your creative. The whole creative. Every frame of a video, every visual element on a static, the words on screen, the tone of the script, the audio, and the landing page it points to. Then it decides who to show the ad to. Your campaign settings still exist. They are just not really the decision anymore. They are more like polite suggestions Andromeda chooses to mostly ignore.

Picture a billboard for a second. The old Meta was a billboard rented for a specific commuter route you picked. The new Meta is a billboard that reads itself, decides who would actually buy what is on it, and follows those people around all day. You used to choose the route. Now the billboard chooses the route, and the only thing it really cares about is what you put on it.

Why Interest Targeting Stopped Mattering

The numbers tell the story honestly. Meta's own benchmark puts Advantage+ campaigns at roughly twenty-two percent higher return on ad spend than traditional manually-built ones, and about fourteen percent lower cost per lead. The bigger number is the adoption curve.

Eighty-two percent is not early adoption. That is the industry quietly agreeing the old way lost. The reason is simple. When Andromeda gets to decide who sees the ad, it almost always finds buyers your carefully stacked interest list would have missed.

In the old model, you narrowed your audience to people interested in, say, yoga, and Meta showed your ad mostly to those people. Now Andromeda is happy to show your yoga ad to someone who has never expressed an interest in yoga, because something about the creative tells it that this specific human is likely to convert anyway. Most of the time it is right. Your interest layer was a guess. Andromeda is a much better guesser.

Your Creative Is Doing the Targeting

If the algorithm is choosing the audience by reading your creative, then your creative is the targeting. That sentence is doing a lot of work, so let it sit for a second.

Andromeda is not reading your mind. It is reading the specific signals you put in front of it. The thing that decides who Andromeda finds for you is a stack of inputs you control, every one of them inside the creative.

• 01
  The first three seconds

  The hook tells Meta what kind of human reacts to this ad. A founder-led opener pulls in a different audience than a UGC review opener, even with identical offers behind them.

• 02
  The visual content

  Vision models read every frame. Who is in it, what they are wearing, the setting, the on-screen text. The face on screen helps Andromeda picture who else looks like a buyer.

• 03
  The script tone and pain points

  The words you use to describe the problem narrow the audience for you. An anxious tone finds anxious buyers. A permission-seeking tone finds undecided ones.

• 04
  The landing page

  Andromeda crawls it to verify the offer matches. Cohesion between ad and page is part of how it qualifies the audience.

Notice what is not on that list. Age, gender, interests, behaviors, lookalike percentages. Those still exist in the interface. They are not where the decision is being made.

The New Job of the Media Buyer

This is where the practical pivot lives. The job has not gone away. It has moved.

The old job was an audience-builder job. You spent your morning building audiences, layering interests, excluding past buyers, testing lookalike percentages, and arguing about whether 1% or 3% would win. Some of that work is still useful at the edges, but it is no longer where wins come from.

The new job is a creative-systems job. You spend your morning on Personas, on hooks, on emotional angles, on the next batch of scripts. You brief the creative team like a director, not a clerk. You watch front-end creative signals like hook rate, click-through rate, and frequency on each individual ad, instead of obsessing over which audience is winning. You let Meta do the matching. Your job is to give it something worth matching. Once you accept this, your creative refresh cadence becomes the actual lever for the whole account.

Every senior media buyer I have talked to recently has independently arrived at the same line, just phrased differently. The smarter the buyer, the more they sound like a creative strategist.

What to Stop, What to Start

Here is the practical translation if you have read this far and want a quick reset for your week.

Stop layering five interests on every ad set. Stop building forty-seven custom audiences nobody can keep track of. Stop hand-tuning 1% versus 3% lookalikes like it is 2019. Stop blaming the audience when a campaign tanks. The audience is no longer where the leverage is.

Start treating creative production like a real production line. Start briefing your team with full-funnel angles, not vague vibes. Start watching hook rate per ad, not ROAS per audience. Start refreshing creative on a schedule, before the current winner fatigues, because in this new world the whole account is riding on the creative pipeline. If you have not already, this is also the moment to put AI agents to work on the parts of the day that used to be manual.

The buyers who adapt to this stop fighting the algorithm and start feeding it. The buyers who do not are about to look very expensive.