Why audience targeting doesn't work the way it used to on social platforms

If your account performance has felt less predictable over the past year, and broad targeting has started beating your carefully built lookalikes, that's a real shift. The major platforms have changed how they decide who sees your ads, and most performance teams are still optimising for a system that no longer exists.

The old model is going

For roughly a decade, ad delivery worked from the audience down. You defined interests, demographics and lookalike seeds, and the platform used those settings to filter which ads reached which people.

That model is being replaced across the board. Major platforms' retrieval systems now work from the creative up, reading the ad itself first, then predicting who's likely to respond based on behavioural signals, rather than the audience box you ticked. Meta's own engineering team has confirmed this shift directly: its Andromeda retrieval system scans tens of millions of ad candidates per session and has delivered a 6% recall improvement and an 8% lift in ads quality since rollout. Every advertiser is increasingly operating inside systems built this way.

What drives performance now

Creative diversity is doing the job audience targeting used to do. These systems reward genuinely distinct executions over 10 versions of the same ad with different headlines: different hooks, different formats, different value propositions. 15 to 20 active, meaningfully different ads is a more realistic benchmark than the smaller sets most accounts still run.

Clean measurement is doing the rest. The better a platform understands what happened after the click, the more accurately it can predict outcomes. Patchy conversion data costs you attribution, and it costs you retrieval too.

How AdStudio helps you keep pace

Meeting this appetite for volume and diversity in-house is hard, because creative capacity, not targeting settings, is usually the real bottleneck. AdStudio is Brave Bison's AI-enabled performance creative solution, built specifically to fix that.

Our team of creatives, strategists and technologists turns your existing assets, organic content and UGC into fresh, channel-native ads at the pace today's platforms demand, all grounded in AudienceGPT so the creative diversity you're scaling is built on real audience data, not guesswork.

It pays for itself quickly too, through a competitive, volume-based pricing model that typically pays back within six months. Seasalt, working with our AdStudio team, saw a 69% increase in CTR and a 32% improvement in ROAS while keeping its brand identity intact. Trafalgar saw a 15% increase in Meta ROAS after AdStudio delivered around 500 creative assets over six months.

The bottom line

Targeting didn't disappear. It moved into your creative and your measurement instead of your audience settings. Accounts that have adapted are seeing broad targeting outperform the old playbook. Accounts that haven't are usually stuck on volume rather than strategy.

Want the volume and diversity without the in-house bottleneck? Get in touch with our AdStudio team.