Meta’s Andromeda Update Just Made Creativity King

Posted on by Jeff Cassman

Over on LinkedIn, my son Michael asked a great question:

Here’s a detailed answer from ChatGPT for advertising geeks:

✅ What is Andromeda (in plain English)

  • Meta describes Andromeda as a next-gen ad retrieval engine — essentially the system that selects which ad (out of many) gets shown to which user. The MTM Agency+3Engineering at Meta+3Foxwell Digital+3

  • It’s built on large-scale ML and high-throughput hardware (e.g., NVIDIA Grace Hopper superchips) enabling Meta to process vast volumes of creative/ad candidates, make decisions in real-time, and serve more personalized ads. Engineering at Meta+2Foxwell Digital+2

  • The upshot: rather than manually finely segmenting audiences + creating a few ad variations, the system now thrives on lots of creative assets + broad targeting + algorithmic optimization. For example: “Andromeda can now scan thousands or even tens of thousands of ads in milliseconds.” Smart Marketer+2The MTM Agency+2

🔍 What’s changed / What the media-buyers are seeing

Yes — you’re correct on many of the bullets. Here’s a more structured list of what’s changed, and some extra ones you might not have seen.

Major shifts

  1. Audience targeting (narrow interest sets) is de-emphasized

    • The old playbook: build many granular audiences, niche interests, multiple ad sets, test who responds.

    • Now: the update prioritizes broad targeting (or even “everyone” within a country), letting Meta’s algorithm find the pockets of responsiveness. The MTM Agency+1

    • Hard truth: If your agency/funnel is still heavily using ultra-segmented interest sets + many ad sets, you’re fighting the platform. See: “targeting is dead.” nicreated.com.au+1

  2. Campaign/ad-set structure simplified & consolidated

    • Fewer campaigns, fewer ad sets → more budget/data in fewer buckets gives more signal to the algorithm. Five Nine Strategy+1

    • For example: one campaign + one ad set + many creatives beats the old approach of many ad sets each with limited creatives. The MTM Agency+1

    • Hard truth: This takes away some of the “manual control” media buyers were used to. For buyers used to micro-managing audience splits, that’s a loss of control.

  3. Creative volume, diversity, differentiation now the key lever

    • Rather than tweaking headlines/CTA on 1 creative (the old “3:2:2” rule etc.), you need many distinct creative concepts (big changes in hook, imagery, message, format), not slight variations. Mindbees

    • The reason: Andromeda now can match which creative to which user rather than just “which audience gets this ad”. The creative becomes the targeting. nicreated.com.au+1

    • Example: Running 25 creatives in one ad set versus 5 creatives across 5 ad sets produced a test result of +17% conversions and –16% cost per conversion. Five Nine Strategy

  4. Faster refreshes, ongoing creative iteration

    • Since the system rewards fresh and diverse signals, creative fatigue (or creative that becomes too similar) hits harder. So you’ll need to iterate often (weekly, bi-weekly). The MTM Agency+1

    • Hard truth: This demands creative bandwidth — more assets, more ideas, faster turnaround — which raises demands (and cost) for copywriters/designers.

  5. Measurement & signal quality become even more critical

    • The system depends on good data (conversion events, pixel accuracy, user behavior signals) to optimize. If your tracking/analytics are weak, you’ll get weaker outcomes. IMM+1

    • Hard truth: Where previous campaigns could prosper despite some sloppy tagging/data, now the algorithm punishes weak signal more.

Additional things you should be aware of

  • The machine doesn’t just pick the best audience; it picks the best creative for each individual. So your old “test creative A vs B in the same ad set” still matters, but it’s more about concept variation than minor tweaks. Socium Media

  • Exclusion logic becomes more important: You’ll want to exclude existing customers or low value audiences so the algorithm doesn’t waste budget optimizing what you don’t want. (Seen in some case studies) Smart Marketer+1

  • The “buy button / campaign objective” still matters — you need to pick your objective cleanly (leads vs purchases) because Andromeda optimizes to that end. If you feed it fuzzy goals, it still will “do its job,” but maybe in a way that doesn’t serve your business. The MTM Agency

  • Because targeting is broader and more volume-based, creative messaging must cover various audience states (awareness levels, pain points, personas) — earlier narrow campaigns let you craft tightly for one persona; now you may need more persona-based variation.

  • From a structural point: Budgets may need to be larger to “feed” the algorithm enough signal quickly. Smaller budgets might struggle because the system can’t learn as fast. This is more of a “scale advantage” for bigger spenders.

🎯 Why is Meta incentivizing this? Who benefits?

You asked why Meta is incentivizing the heavier investment in creative and volume — good question.

Why Meta is doing this

  • From Meta’s engineering blog, Andromeda is designed to “process an enormous number of ad candidates … enabling a faster and more efficient delivery of relevant ads.” Engineering at Meta

  • More creatives + broader placements mean Meta can serve more ads with better match quality — good for user experience (less irrelevant ads) and good for Meta’s business (more efficient ad spend, less wastage).

  • It also supports more automation, meaning less manual media-buyer work and more reliance on Meta’s internal optimization. That increases Meta’s leverage as platform.

  • More creative variation also means more ad inventory (since each advertiser is uploading more ads) → Meta can monetise more placements and better fill.

  • It shifts the competitive advantage away from manual targeting tactics (which many advertisers replicated) to creative production ability — making the barrier to entry higher for small players with weaker creative & production capability.

Who benefits

  • Advertisers/agencies with strong creative production capability: If you can rapidly generate many high-quality, conceptually different creatives, you’ll win.

  • Advertisers with budgets large enough to feed the algorithm: Because it’s volume + data density, bigger budgets and brands with more resources will see more benefit.

  • Meta (the platform itself): They benefit by being able to charge more, show more ads, more effective combinations, and deliver better ROAS (which keeps advertisers spending).

  • Creative teams/copywriters/designers: Indirectly benefit if they adapt — the demand for creative variation goes up; clients will need more creative concepts, more storytelling, richer angles.

Who is at risk / loses advantage

  • Smaller advertisers or those with limited creative capacity will struggle, because they might still rely on the older playbook (few creatives, narrow audiences) and get left behind.

  • Media buyers whose competitive edge was in micro-targeting / audience segmentation will see that strategy de-prioritised — so their skillset needs to shift more toward creative strategy + concept development.

  • Brands who have weak offers, weak landing pages, or low conversion funnels — Meta’s algorithm can only amplify a good campaign, not save a bad one. One write-up says: “Andromeda amplifies what already works, rather than fixing what does not.” The MTM Agency

✍️ Implications for you (as a copywriter) and what to do

Since you’re a copywriter working in a media-buying ecosystem (or supporting one), here are the strategic challenges and action points you should consider.

Strategic implications

  • You’ll need to shift from writing one great ad version to helping build a creative library — many variations, different angles, hooks, formats.

  • Messaging needs to target different personas/awareness stages simultaneously. E.g., unaware, problem-aware, solution-aware — each needs its own voice. This aligns with what you’ve seen about “Catch, Connect, Convert” buckets.

  • Copywriting needs to be more modular: You’ll write headlines, sub-heads, body copy, variations, but also map different emotional triggers, story styles, formats (UGC, founder story, testimonial, product demo, etc).

  • Because creative is now the lever, your role expands: Not just “write copy that converts,” but “help generate creative concepts that the algorithm can test,” which means thinking more like a strategist/creative manager than just “here’s the ad text.”

  • Timeliness matters more: Creative needs frequent refreshing, new angles, and you’ll likely iterate more often. Budgets may go toward testing heavier volumes.

  • Landing pages and offer clarity become more important. The algorithm can’t fix a weak offer. So your copywriting around offer clarity, value proposition, urgency should be strong.

Action steps you should push / ask your media buyers for

  • Creative concept planning sessions: Sit with media buyers/strategists early and map out persona × desire × awareness combinations. Decide a library of 8-15+ distinct ad concepts (or more) per campaign.

  • Ensure each creative is genuinely different: Not just tweaking CTA or image; but different story, tone, hook, format. For example: one story about transformation, another about founder vision, another about “vs competitors,” another about fear of missing out, etc.

  • Ask for performance data by creative concept: Which hooks/angles work best? Feed that back into copy production quickly. Having many creatives means you need to organise tracking and naming conventions so you can isolate what concept is resonating.

  • Align with landing page and funnel copy: If you produce ad copy but then the landing page doesn’t carry the same tone or message, the algorithm gets less signal. Make sure copywriting continuity exists.

  • Be ready to refresh: Set a schedule (e.g., weekly or bi-weekly) to add new creatives, retire under-performers, iterate. That means you may need retainer or production planning for ongoing deliverables, not just “one & done” ads.

  • Focus on big problems/benefits rather than micro-tweaks: Since the system rewards creative breadth more than little tweaks, your focus should escalate toward big emotional/strategic messages rather than tiny copy variations.

  • Support “Catch, Connect, Convert” structure: Yes — you’re right to focus on that three-bucket model. Think:

    • Catch (awareness): broader messaging, hook, grabbing attention.

    • Connect (consideration): deeper storytelling, credibility, differentiation.

    • Convert (decision): offer, proof, urgency, clear next step.
      And for each bucket, multiple creative concepts.

Hard truths / Things to challenge

  • Just because you can throw 50+ creatives into a campaign doesn’t mean they’ll all perform or matter. Quality still matters. If you flood with low-quality creatives, the algorithm may penalise performance or you’ll waste budget. One article said “simply producing more of the same will not improve performance.” The MTM Agency

  • The barrier for entry rises: Smaller budgets or low creative output brands may see less competitive performance. That means you might have to counsel your clients on increased investment in creative production if they want to keep up.

  • The “creative is pivotal” doesn’t mean targeting/audience becomes irrelevant — you still need good fundamentals (good landing pages, good offer, good measurement). Several sources stress that weak fundamentals = poor results, no matter how good the engine is. The MTM Agency

  • As a copywriter, you may still find clients/teams resistant to this shift (they want small tests, incremental tweaks, narrow audiences). Part of your strategic role is educating them on why “old playbook = risk” in this era.

🧮 Putting it all together: Why your original list was pretty much right

You listed:

  • 3-bucket structure (Catch, Connect, Convert) → yes, aligns with awareness/consideration/decision messaging.

  • Incentivizing massive volume of creatives → yes, agreed.

  • Discouraging subtle alterations of creative in favor of dramatic differentiation → yes.

  • Incentivizing much heavier investment in diversity & volume of creative → yes.

So you’re on point. The benefit is that the algorithm now rewards that approach because it has the architecture to match many creatives to many user-signals, so the “who” (audience) becomes less manual and the “what” (creative) becomes the new frontier. The winners benefit by being able to scale faster, find high-value pockets the algorithm discovers, and move ahead of advertisers still using the old model.