Meta Ads After Andromeda: What Changed and How to Adapt

Meta rolled out Andromeda in 2023, and it fundamentally changed how their ad delivery system works. If your Meta Ads strategy has not evolved since then, you are leaving money on the table. Here is what actually changed, why it matters, and what you need to do about it.

What Andromeda Actually Is

Andromeda is Meta's upgraded ad retrieval system. Before Andromeda, Meta's system would pull from a relatively small pool of ads when deciding what to show a user. Andromeda expanded that pool by orders of magnitude. Where the old system might evaluate a few hundred ads for each impression, Andromeda evaluates tens of thousands.

The practical impact is that Meta's algorithm got significantly better at matching ads to users. It no longer needs narrow audience targeting to find the right people, because it can evaluate far more options in real time.

What This Means for Advertisers

Three big shifts happened as a result of Andromeda.

1. Targeting matters less, creative matters more

Before Andromeda, media buyers spent most of their time building audience segments, testing lookalikes, and refining targeting parameters. That work is now less impactful. Meta's algorithm is better at finding the right people than you are at defining them.

The lever that still matters is creative. The algorithm can find the right audience, but it can only deliver the ads you give it. If your creative does not stop the scroll and drive action, no amount of algorithmic sophistication will save you.

This is why we see broad targeting outperforming detailed interest-based targeting for most DTC brands. The algorithm works better with fewer constraints and more creative options to test.

2. Creative volume and variety are now competitive advantages

With Andromeda evaluating more ads per impression, the brands that give the algorithm more options tend to perform better. This does not mean producing more ads for the sake of it. It means running a systematic creative testing program that consistently feeds new concepts into the system.