Meta ads targeting is dead. Creative is the targeting now

Targeting is dead, and creative is the targeting now. You used to use targeting settings to tell Meta which audience to go after; it now matches on the creative itself, which is why it pushes everyone toward Advantage+ audiences and demotes what you typed to a suggestion. So you reach a different audience by making creative that speaks to that persona, and by spending enough to give the algorithm something to optimize on. Restrictions just make it less efficient or point it the wrong way.

Meta now treats the targeting you type in as a suggestion.

That is not a quirk of the interface. That is the platform telling you it can find these people better than you can, and it is right.

If I open your Meta account and the first thing I see is a stack of interest audiences and lookalike sets, I already know what happened. Someone sold you on the old way.

That is where paid accounts used to be won. It is not where they are won anymore. The lever moved, and most teams are still pulling the old one.

The algorithm targets better than you do

The algorithm is good enough to find your people without you telling it who they are. Broad targeting plus a strong conversion signal now beats hand-picked audiences almost every time, because the platform watches thousands of signals on every user and finds the likely converters far better than your guess about which interests matter.

Meta knows this, and it has stopped being subtle about it. That is what Advantage+ audiences are: you can still type an audience in, and the platform treats what you typed as a suggestion rather than a constraint. Read that as the tell it is. The company with the data is telling you its algorithm can find these people for you.

When you box the algorithm into a narrow interest audience, you are handing it a worse version of the job it already does automatically, and starving it of the scale it needs to learn. Give it too many restrictions and you do one of two things: make it less efficient, or put it on the wrong path entirely. So the manual audience-building most teams still obsess over is not just low-value. It actively gets in the way.

The creative is the targeting

Here is the shift that matters, and it is an inversion rather than a downgrade. You used to use targeting settings to tell the algorithm which audience to go after. Now the creative does the finding.

Which means the lever is still there. It just moved. If you want to reach a different audience, you make creative that speaks to that persona, and the ad goes and finds them. A specific ad speaks to a specific person, and the people it resonates with are exactly the audience the algorithm learns to chase. The hook, the visual, the first three seconds: that is your audience selection now.

So the question stops being which audience do I pick, and becomes which personas am I actually speaking to, and how many distinct angles can I get in front of the machine.

That reframes the whole job. You are not an account manager tuning audiences. You are running a creative operation.

The other half of the reframe is spend, and it gets skipped. The algorithm needs enough data to optimize on. An account that never gets out of the learning phase is not going to find anyone for you, however good the creative is, and no amount of creative volume compensates for a budget too thin to produce a signal.

What actually moves the needle: three things

Once you go broad, the account is won on three things. Most teams only do the first one.

1. Creative volume and velocity

Not one polished ad, and not fifteen variants of the same angle. Dozens of hypotheses running at once, across different personas, angles and hooks. Different people, different problems, different hooks. More variants means faster learning means better results, and the team shipping thirty concepts a week beats the team shipping three and arguing about audiences.

2. Clean data signals

This is the one most teams get wrong from the start. Your front-end and back-end events have to fire correctly and pass as many user properties back to the platform as possible. Good creative with broken data is a leaky bucket: the algorithm can't optimize toward outcomes it never sees.

3. End-to-end automation

This is where the best programs pull away from everyone else. Performance data feeds automatically into analysis. Winners and losers get identified without manual reporting. Results feed into new creative briefs, new variants get generated and launched, and the loop repeats as fast as possible, with AI at the center of it. One person running this system beats a full agency running interest audiences, every time.

What to actually do

Go broad. Give the algorithm room, a clean conversion signal to optimize against, and enough budget to get out of the learning phase. Then put your real energy where it now belongs: a creative pipeline that produces a high volume of distinct angles aimed at distinct personas, a testing rhythm that cuts losers and scales winners quickly, and AI in the loop to make both faster. Stop grading yourself on how clever your audiences are. Start grading yourself on how much good creative you ship and how fast you learn from it.

The takeaway

Creative is the targeting now. That is the whole thing.

You used to tell the platform who to go and find. Now you make something that speaks to a particular person and the platform goes and finds them, which is why it demotes what you typed to a suggestion and pushes you toward Advantage+. Give it a broad target, a clean signal and enough spend to learn on, and move your effort to creative aimed persona by persona. Restrictions do not make it more accurate. They make it less efficient, or they put it on the wrong path.

Common questions

Should I still build custom audiences and lookalikes on Meta?

Mostly no. Broad targeting with a clean conversion signal now beats hand-picked audiences, because the algorithm optimizes on far more signal than you can. Put that effort into creative instead.

If targeting is dead, what actually decides paid performance?

The creative, because the creative is now doing the targeting: you reach a different audience by making something that speaks to that persona. Behind that, three things. Creative volume and velocity across distinct personas, clean data signals so the algorithm can optimize toward real outcomes, and end-to-end automation that turns results into the next round of creative. You also have to spend enough for the algorithm to have data to optimize on. Great creative on broken data, or on a budget too thin to learn from, is a leaky bucket.

How much creative do I need to ship?

As much as you can produce and honestly read. Volume gives the algorithm angles to work with; velocity lets you cut losers and scale winners fast. An AI-assisted pipeline is how you do both faster than competitors.