Andromeda moved the deciding lever from audience targeting to creative. If your spend has stalled, the account is probably optimising the wrong thing. Here is what changed, and what to do about it.
Andromeda is the name for Meta's AI-driven ads system. In plain terms, it automates audience targeting and expands who sees an ad. You hand over broad inputs and a budget. The system decides the distribution, matching each impression to the person most likely to act.
That is a real shift. For years the craft of Meta buying was audience work: interest stacks, lookalikes, detailed targeting, careful segmentation. Andromeda absorbs most of that. It reads a far wider pool of signals than any buyer can, and it finds the audience itself.
So the question changes. It is no longer "who do I target?" It is "what do I give the system to work with?" The answer is creative. Andromeda decides who sees an ad, but it can only choose from the ads you make. Your creative is now the main thing you still control.
Most accounts that stalled after Andromeda stalled for the same reason. They kept pulling the old levers.
The habits that used to drive growth were audience and targeting changes: new lookalikes, tighter segments, more ad sets. Andromeda has taken that work over, so those tweaks no longer move much. You can spend a week restructuring campaigns and see nothing, because the system was already doing that job.
Meanwhile the lever that does move, creative, gets the least attention. An account running a handful of ads starves the system. It has almost nothing to match against a broad audience, so it saturates quickly, costs rise, and scaling stops. The ads did not get worse. The account simply ran out of inputs.
The honest read on a stalled account is usually not a targeting problem or a bidding problem. It is a creative supply problem.
Andromeda rewards range. To match a broad, shifting audience it needs a steady supply of genuinely different ads, not minor variants of one idea. Ten recolours of the same video do not count as ten creatives. The system sees one concept.
Diversity means different angles, different hooks, different formats and different people. UGC next to high-quality video next to statics next to motion. One concept that leads with price next to one that leads with a problem next to one that leads with a result. The more real variety you supply, the more room the system has to find a match you would never have guessed.
Volume matters because most ideas fail, and that is fine. Testing is how you find the few that scale. If you can only make four ads a month, you cannot test properly, and you hand the advantage to whoever can.
This is why production capacity, not media strategy, is now the thing that caps growth for most brands.
A practical list. None of it is exotic. The discipline is in doing it every week, not once.
Growth now comes from the rate at which you find new winning creative. Everything above exists to raise that rate.
Less than it did, and not in the old way. Detailed manual targeting mostly duplicates work the system now does better, and often fights it.
Targeting still matters as a signal rather than a cage. A clear offer, the right exclusions and clean conversion data point the system in the right direction. After that, the creative does the targeting. An ad that opens by naming the viewer's problem tells Andromeda who the ad is for, more precisely than an interest stack ever did. You steer with creative now, not with audiences.
The bottleneck is now creative volume and quality. That is a production problem, and most in-house teams are not staffed to solve it. Making a hundred genuinely different, well-made ads a month is hard, and few brands can do it without the work slipping or the quality dropping.
This is the gap a specialist agency fills. The job is a pipeline: a high volume of diverse, tested assets, plus the media buying to read results and reinvest. It is the model we built Multiply on.
We scaled monthly spend from $200k to $900k while CAC fell. Meta ad disapprovals dropped 95%, which is what made the volume possible.
Kalshi hit #1 free app on the US App Store. To fuel the surge we produced more than 100 ads in a single week.
We were the creative partner as Gamma scaled. Creative output rose and CAC fell 60%.
None of this replaces a good product or a real offer. But in a world where Meta decides the distribution, the brand with the better and deeper supply of creative wins. That is the honest state of the channel after Andromeda.
Andromeda is the name for Meta's AI-driven ads system. It automates audience targeting and expands who sees an ad, matching each impression to the person most likely to act. You hand over broad inputs and the system decides the distribution.
Most accounts stalled because they kept optimising the wrong lever. The old growth tactics were audience and targeting changes. Andromeda took that work over, so those tweaks no longer move much. Performance now hinges on the creative you feed the system, and accounts running a handful of ads starve it.
Far more than most teams produce. The system needs a steady supply of genuinely different ads, not minor variants of one idea, so it has room to find matches across a broad audience. The practical bar is volume and real variety every week, across formats, hooks and angles. We ship more than 100 tested assets a month per account for this reason.
Consolidate the account so the system has room to learn, then make creative the engine. Produce a high volume of diverse concepts, test them honestly, kill the losers fast and reinvest in the angles that work. Growth comes from the rate at which you find new winning creative, not from manual targeting.
Less than it did, and not in the old way. Detailed manual targeting mostly duplicates work the system now does better. Targeting still matters as a signal rather than a cage: a clear offer, the right exclusions and clean conversion data point the system in the right direction. The creative then tells it who the ad is for.
The bottleneck is now creative volume and quality, which is a production problem most in-house teams cannot staff for. A specialist agency supplies the pipeline: a high volume of diverse, tested assets, plus the media buying to read results and reinvest. We have advised on more than $500M in ad spend and test more than 15,000 creatives a year.
We make the ads and run the media that scales them. Creative that performs.
Book a call →