By EGC Group Meta’s Andromeda ad retrieval system has fundamentally changed what running a campaign means. Instead of optimizing a fixed ad set against defined segments, it incentivizes uploading a greater diversity of creative and showing it to a broad audience. It then lets the algorithm decide which messages, formats, and angles resonate with which users. What truly works best is not minor variations of the same concept but several fully distinct concepts and stories. The punchline, from a recent Morning Brew deep-dive on Meta’s AI push: A recent client brief called for 1,000 unique creative assets under a single campaign concept. Read That Again: “One Campaign. One Concept. One Thousand Assets.” This is not an edge case. This is where Meta performance is going, and the brands treating it as a production anomaly are the ones who will wake up in Q3 wondering why their CPAs are climbing while those of their competitors are not. Here is what is genuinely happening under the hood: What works best inside Andromeda is not minor variations of the same ad. It has several fully distinct concepts and stories. Once you game that out across four different target personas with five new concepts for each, a single campaign requires work that used to take a full quarter to produce. The creative brief did not get bigger. The platform’s appetite did. Giving More Control to Meta For marketers, the proliferation of AI features has meant ceding more control to Meta’s systems where targeting, budgeting, and even the creative process are concerned. Often, marketers are opted into new AI features whether they want them or not, constantly playing Whac-A-Mole to figure out what feature has been turned on without notice. This brings us to the real structural problem that’s created; one that goes well beyond production volume. The old agency model was built on separation: Media team buys the placement, creative team makes the ad, and account team reports the results. Those three conversations happened sequentially, in different rooms, often on different timelines. That model made sense when campaigns were static and a television spot ran for six weeks. It does not make sense when the platform is running continuous creative tests, reallocating budget in real time, and the performance signal you need to act on has a 48-hour shelf life. Opportunities to Deliver Above and Beyond A rebundling is happening. AI is forcing reappraisal of the separation of creative, media, and production, and there are real opportunities for the truly integrated agency to deliver exceptional value across every layer. This is exactly what EGC’s Performance Media Engine was built for. Not as a tagline, but as a structural answer to the Andromeda problem. When creative, media, and measurement operate as one connected system (not three departments in sequence), you can accomplish what Meta’s algorithm truly rewards: Testing meaningfully distinct concepts at volume, reading the signal fast, and reallocating before the budget disappears into a format that stopped working two weeks ago. The Rules of How to Win Have Changed The brands winning on Meta right now are not the ones with the biggest budgets. The winners are brands whose creative pipelines are connected to their media data in real time, and whose agency is structured to act on that data (not report on it after the flight ends). As one agency leader stated at this week’s indie agency summit in Chicago: “When every agency has access to the same AI, the same data, the same playbooks, capability stops being an advantage.” Correct. This means that integration is the advantage. The question for every brand running serious Meta spend right now is simple: Is your agency built for the platform you are actually on, or the one from three years ago? EGC Group’s Performance Media Engine connects creative production, media buying, and real-time measurement in one system. Talk to us about what that looks like for your Q2.
