06 Aug Meta Andromeda Ads: Why a Hybrid Structure Works Better
Andromeda has been live for a while now — Meta rolled it out around August or September 2025, and it’s had real time to prove itself. I’ve implemented it across nearly every one of my client accounts. The honest result: not every one of them has actually seen great success with it.
What Andromeda promises
Andromeda is Meta’s rebuilt ad delivery engine — a new matching system Meta describes as smarter about reading audience intent. The pitch is straightforward: instead of running a few tightly targeted ads, you load a single ad set with a large pool of creative — Meta suggested numbers as high as 20-30 ads — covering different personas. Middle-aged, older audiences, teenagers, male, female, parents, whatever segments matter to your business. The algorithm is supposed to figure out which version to show to which person, based on who they are and where they are in their journey.
Meta ads has always been intent-based advertising. What Andromeda changes is how much of that matching work you’re handing over to the algorithm versus doing yourself through manual ad set structure.
Two problems that show up in practice
After running this for months across a real client base, two patterns kept surfacing.
Budget concentrates on two or three ads. This isn’t unique to Andromeda — it’s a familiar problem any time you run many ads in one ad set. Out of ten ads, two or three will typically take the lion’s share of spend. That’s fine if it’s performing well and staying consistent. It’s the second problem that makes it a real issue.
The winning ad doesn’t stay funded. An ad performs well this week — strong CTR, solid cost per result — and next week, Meta pulls back budget from it to test other ads still sitting in the pool. That inconsistency is what creates real volatility. You can’t build a scaling plan around an ad that might get starved of budget the following week for reasons that have nothing to do with its own performance.
To be fair, this isn’t a verdict that Andromeda doesn’t work. It’s closer to a 50/50 split across the client base — some accounts genuinely do well fully on Andromeda, others don’t, and the difference isn’t always obvious upfront.
The hybrid structure that’s actually stabilizing results
Here’s what a hybrid approach looks like in practice, using a $100/day budget and 10 ads as an example:
- Run an Andromeda-style ad set with all 10 ads, spending roughly $50-60 of the daily budget there.
- With the remaining budget, run separate single-ad campaigns — starting with maybe 5 of those same ads — testing each at a modest daily spend (around $10 each) on a weekly cadence.
The logic: if Andromeda is genuinely stable for this account, you gradually shift more budget toward it over time. If it’s not, the single-ad structure gives you a fallback that’s already proven itself.
Single-ad campaigns tend to be more stable specifically because there’s no pool of competing creative for the algorithm to reshuffle budget across. If an ad is running at a certain CTR, CPM, and cost per result, it tends to keep running at roughly that level. That stability is what lets you scale with actual confidence, instead of scaling into a structure that might reshuffle itself next week.
How the testing cycle works
Test single ads for about two weeks. If an ad isn’t working by then, retire it and bring in a new one to test. Keep cycling until you’ve built up a bench of proven winners — some accounts end up with 4-5 solid single-ad campaigns running well, others build up to 10-15 over time. Those get scaled directly. Meanwhile, Andromeda keeps running in parallel, and if it continues to perform, you lean into it further; if it doesn’t, you keep shifting weight toward the single-ad side.
Don’t get attached to a framework — get attached to results
The honest position here isn’t “Andromeda is bad” or “single-ad structures are always better.” It’s that neither framework deserves loyalty for its own sake. The job isn’t to prove that one structure is superior — it’s to find whichever structure gets a specific account the best result, and stick with that. This is the same reasoning that applied years ago when advertisers were arguing over detailed targeting versus broad targeting: the goal was never to win the argument, it was to see what actually worked and follow that.
Testing this properly does take real strategy work and close attention to the key metrics week over week — it’s not something that gets resolved by looking at an account once and issuing a verdict.
A quick self-check for your own account
- Are 2-3 ads eating most of your budget while the rest barely get tested?
- Did last week’s top performer lose budget this week without any change on your end? That’s the inconsistency pattern to watch for.
- Have you tried a parallel single-ad structure alongside your Andromeda ad set, or is Andromeda the only structure you’re running?
- Are you scaling based on a few weeks of consistent single-ad data, or reacting to daily fluctuation inside a large creative pool?
This breakdown is written by Jason Gan, a Meta Certified Professional and Badged Meta Business Partner who has personally audited over 1,000 advertiser accounts since 2010. You can see real account breakdowns on the Jason Gan YouTube channel.
Frequently Asked Questions
What is Andromeda and how does it change Meta ad delivery? Andromeda is Meta’s rebuilt ad delivery engine, designed to match ads to audience intent more precisely by learning from a larger pool of creative in a single ad set — potentially 20-30 ads covering different personas, ages, and genders. The promise is that Meta’s algorithm figures out which version to show whom, based on where each person is in their journey.
Should I stop using Andromeda-style ad sets with lots of creative in one ad set? Not necessarily, but don’t rely on it alone. Roughly half the accounts running this structure see genuine benefit from it, while the other half experience budget concentrating on two or three ads and shifting inconsistently week to week. A hybrid approach — testing Andromeda alongside a separate single-ad structure — lets you see which one actually performs better for your specific account.
How do I know if Andromeda is working for my account? Watch for two signs of trouble: a small handful of ads taking most of the budget while the rest barely get tested, and a top-performing ad losing budget the following week even though nothing about its performance changed. If you see either pattern consistently over a few weeks, that’s a signal Andromeda isn’t stabilizing for your account.
What’s a hybrid Andromeda and single-ad structure? It’s splitting your budget between an Andromeda-style ad set with many creatives, and a separate set of single-ad campaigns tested individually with smaller budgets. Winners from the single-ad tests graduate into their own scaled campaigns once they show a few weeks of consistent performance, while underperformers get retired and replaced.
Get 1-on-1 Meta Ads Coaching — $400/month, No Minimum →
Weekly 30-minute strategy sessions, ongoing account access, and email support in between — built for exactly this kind of week-to-week structural testing and adjustment.
If you’d rather start with a single diagnostic look at whether Andromeda or a hybrid structure fits your account before committing to ongoing coaching, a Meta Ads Audit is the lower-commitment first step.