Post-Andromeda Targeting: The Only Method That Works

If you are still running interest targeting (audience targeting based on declared interests) or lookalike audiences (LLA, audiences modeled on existing customers) on Meta and your ROAS is dropping, you are fighting the algorithm. The only post-Andromeda Meta ads targeting method that consistently works in 2026 is broad targeting paired with creative signals that explicitly call out your buyer through visuals, copy, and price. Meta’s Andromeda update, powered by the GEM (Generative Embeddings Model) AI, now distributes ads as a full-funnel sequence rather than optimizing a single placement, which means restricting audiences upfront actively hurts the algorithm’s ability to find your buyer.

This is the same method we use across the brands we manage at Skaleit, where we spend millions per month on Meta ads.

TL;DR

– Interest and lookalike targeting are obsolete post-Andromeda. Meta has even removed the ability to untick “audience suggestion” on some accounts, forcing advertisers into broad. – GEM + Andromeda = full-funnel ad distribution. The algorithm optimizes the entire user journey, not single ad views. – Creative is now the targeting layer. Visuals, copy, price, age callouts, and production quality tell Meta who to serve the ad to. – Real results: One client did $220K revenue on $60K spend (3.62x ROAS), another did $831K on $210K spend (3.95x ROAS), another did $125K on $37K spend (3.38x ROAS) using this exact method. – Quick audit: Drop your creative into ChatGPT and ask who it targets. If the answer is vague, Meta will be vague too.

Why Interest and Lookalike Targeting Stopped Working Post-Andromeda

Andromeda is Meta’s algorithm update that re-architected how ads get matched to users. The brain behind it is GEM, an AI model that delivers ads in sequences to optimize the entire user journey rather than a single impression. According to Meta Engineering’s own breakdown of Andromeda, the system relies on dense embeddings and contextual signals far more than declared audience attributes.

What this means in practice: every restriction you add at the ad set level (interest layer, lookalike seed, narrow age range) reduces the data Meta needs to find your buyer. You are blindfolding the algorithm at the exact moment it is most capable of seeing.

We have also noticed that on newer ad accounts, the interest field is locked as a “suggestion” only. You cannot untick the expansion box. Meta is literally taking the option away. That alone should tell you where the platform is heading.

The audience-pool diagram that explains it

Imagine your total addressable audience is 20 million people:

– Broad targeting: Meta sees all 20M and uses creative signals to find the ~500K to 1M who match. Maximum signal, maximum scale. – Interest targeting: You cap Meta at ~10M, but you do not actually know if your real buyers sit inside that 10M. You hope they do. – Lookalike targeting: Same constraint as interest, just modeled on a seed list.

In all three cases, your creative still only resonates with the same ~500K to 1M. The reachable audience does not change. What changes is how fast and accurately Meta finds them. Broad wins 99% of the time in the long run, even when interest looks better in the first 48 hours.

How GEM Reads Your Creative to Target Buyers

Creative signals targeting is the new mechanism. Meta’s AI reads the image, the copy, the price on screen, the people shown, and the implied user awareness stage, then matches the ad to users whose behavior patterns suggest they will convert.

This is why two ad sets with identical broad targeting can perform completely differently: the creative itself is the audience filter. Our internal data across $10M+ in managed spend shows that creative diversification, not audience layering, is what unlocks scale. We covered the structural side of this in our breakdown of Meta’s Andromeda update and what it means for ecommerce, which is the pillar for everything we publish on the topic.

The ChatGPT audit (do this in 30 seconds)

Upload your ad creative to ChatGPT and ask: “What audience is this ad targeting on Meta?” Whatever ChatGPT answers, Meta is interpreting your creative the same way, because Meta uses the same kind of multimodal AI. If ChatGPT cannot identify the age, income level, or pain point, neither can GEM.

The Wrong Way: A Generic Creative That Forces Meta to Guess

Let’s say you sell a luxury anti-wrinkle cream and your buyer is a 45+ woman with high purchase power who will spend $500 on skincare.

The wrong creative looks like this:

– Headline: “Anti-aging cream” – Visual: Product on a white background – No price shown – No human in the frame – No age reference

This tells Meta nothing. The messaging is too broad, the image only shows the product, and there is zero relevance signal for who should see it. Meta will serve it to 22-year-olds who never convert.

The Right Way: A Creative That Tells Meta Exactly Who to Target

The correct version of that same ad includes:

1. High production quality visual that communicates luxury at a glance.
2. The price visible on screen (e.g., “$170”) which pre-qualifies clicks and filters out low-intent traffic.
3. A human in the creative who matches the target demographic (a 45+ woman, not a 25-year-old).
4. Audience callouts in the copy: “Reduce 95% of wrinkles in 2 weeks for women in their 40s.”
5. A clear awareness-stage match: speak to problem-aware or solution-aware buyers if you want to scale beyond $30K/month.

Notice what we are doing: we are not asking Meta to find 45+ high-income women through ad set settings. We are showing 45+ high-income women in the ad itself, and letting GEM connect the dots. This is the same logic that powers the creative strategy that generated 7.02x ROAS for one of our ecommerce brands.

Why showing the price matters more than ever

For premium products, displaying the price in the creative does two things: it self-qualifies the click (no one accidentally clicks a $170 cream) and it gives Meta a hard purchase-power signal. We have seen CPA drop 20-30% on luxury skincare and supplement brands just by adding a visible price tag to top-of-funnel statics. Meta’s own creative best practices documentation backs this up, recommending price transparency for premium positioning.

Real Results From This Method (Last 30 Days)

These are pulled directly from accounts we manage as of November 2026:

| Brand | Ad Spend | Revenue | ROAS | |—|—|—|—| | Brand A | $60,000 | $220,000 | 3.62x | | Brand B | $37,000 | $125,000 | 3.38x | | Brand C | $210,000 | $831,000 | 3.95x |

Every one of these accounts runs 100% broad targeting. No interest layering. No lookalike audiences. The targeting work happens entirely inside the creative brief.

For brands struggling to break past $30K/month, the issue is almost always that they are stuck targeting the most aware audience (product-aware and solution-aware buyers only). To scale, you need to push creative angles that speak to problem-aware and unaware audiences. We unpack the mechanics in our guide on why you can’t scale Meta ads post-Andromeda and how to fix it.

How to Build a Creative-First Targeting System in 2026

Here is the framework we follow on every account:

1. Lock ad set targeting to broad. Country only. Age 18-65+ unless legally restricted. No interests. No lookalikes.
2. Define your buyer in a 1-page brief: age, gender, income bracket, awareness stage, pain point, objection, identity.
3. Translate every element of that brief into the creative. Age shown via casting, income shown via price and production quality, pain point shown via hook and copy.
4. Test 5-10 angles per week, not 5-10 audiences. Angle diversification beats audience diversification every single time post-Andromeda.
5. Use the ChatGPT audit before launching every new concept.
6. Read winners by hook + visual + awareness stage, not by ad set.

This is the system that scales. Audience targeting is a 2019 conversation. In 2026, the creative is the audience.

FAQ

About the Author

Antonio Ventre is the founder of Skaleit, a Meta ads agency for ecommerce brands. Skaleit manages over $10M in ad spend across fashion, skincare, supplement, and toy brands, and specializes in post-Andromeda scaling systems.

Want Skaleit to Build This System for Your Brand?

If you want our team to audit your account, rebuild your creative strategy around post-Andromeda targeting, and scale your Meta ads with the same system that drives 3.6-3.9x ROAS for our clients, book a call with Skaleit here.