<aside> 💡

TL;DR

You can’t just rely on Meta’s broad audiences. We’ve seen data-enriched audiences from Proxima drop CPAs by meaningful levels at scale. In turn, this has enabled us to crank up ad spend much more efficiently than what our typical broad audiences allowed us to.

So yes, the good ol’ days of audience targeting are back. And you should consider experimenting with Proxima’s AI Audiences if you’re spending +$50k per month.

</aside>

Case Studies

Hype10 use cases for Proxima

The Explanation

The primary reason we started testing Proxima’s audiences was that rolling reach was diminishing for many of our brands. We needed to find niche audiences that were overlooked by the pixel learnings.

Meta's targeting capabilities are not even close to what they used to be. Between privacy changes, soaring CACs, poor match quality, and limited targeting options, prospecting at scale is more challenging than ever.

We invest heavily in creative, but that’s not the only lever to pull.

There are tactics beyond creative testing that are delivering serious results for Hype10’s roster of high-growth DTC brands.

What’s working incredibly well right now?

→ Secondary scaling audiences

Broad targeting is good but limited in many ways:

• It’s heavily reliant on creative inputs (frequency, volume, new angles, etc)
• You end up targeting the same pools of people over and over again. To the system, they’re the most qualified to buy, but it’ll keep saturating that pool w/o expanding to net new ones. We refer to this as “local maximums”.
• When looking to scale, we need to open the aperture. So we’ve looked to secondary audience sources (and datasets) to crack that open.