How Many Creatives Should You Run in a Meta Advantage+ Campaign?

Updated September 2026

Post-Andromeda, the answer is 15-20 active ads with distinct hooks per ad set — not per campaign, per ad set. That number comes from Meta's own updated best practices after the Andromeda retrieval engine replaced audience-based delivery with creative-based delivery. Running fewer gives the algorithm too little signal diversity; running more than ~50 without a clear rotation strategy tends to dilute spend across too many low-impression variants to learn anything useful.

This matters because the creative pool is now doing work that targeting used to do. The more hook and format variation you give the algorithm, the wider the effective audience it can reach.

Why Andromeda Changed the Creative Math

Before Andromeda, you could run 3-5 ads in a tight ad set and compensate with detailed audience targeting. The system found your buyers; creative was the thing you showed them.

After Andromeda, the retrieval layer matches ads to users primarily on creative signals — visual content, hook type, format, and engagement patterns. Meta's Andromeda engine improved ad retrieval recall by 6% and ad quality by 8% in tested segments, but those gains only materialize when the creative pool is wide enough for the system to find differentiated signals. A narrow pool of near-identical ads looks like one ad to the retrieval model.

The practical consequence: creative volume and diversity is your targeting strategy now. A team that treated creative production as an afterthought in 2023 cannot compensate with audience precision in 2026.

The 15-20 Active Ads Benchmark — What It Actually Means

Post-Andromeda best practice is 15-20 active ads with distinct hooks per ad set. Two words in that sentence do most of the work: active and distinct.

Active means ads that are spending. Paused ads don't count. Ads stuck in review don't count. If you have 25 ads but 10 are paused or throttled below the minimum spend threshold, the algorithm is working from a pool of 15 anyway.

Distinct hooks means the first 2-3 seconds are genuinely different — different opening visual, different opening line, different problem framing. Swapping a color or resizing for a different placement does not produce a distinct hook. The retrieval model has seen your variations; it groups near-duplicates.

A practical way to structure a pool of 15-20:

This isn't a rigid taxonomy — it's a way of ensuring the hooks are genuinely different rather than superficially varied.

Advantage+ vs. Manual: Does Creative Volume Matter More on One?

Advantage+ automation now touches 82% of Meta advertisers and runs 62% of ecommerce ad spend on the platform. That scale reflects real performance: Meta cites a 22% ROAS lift for simplified Advantage+ campaign structures, and Advantage+ Shopping Campaigns deliver 17% lower CPA than manual campaigns on average.

The live post at Meta Advantage+ vs Manual Campaigns covers the structural tradeoffs in depth. The creative-volume angle is different: Advantage+ needs more creative because it distributes spend dynamically across a larger implicit audience pool.

In a manual campaign with tight audience targeting, the algorithm shows your ads to a constrained pool of people. Five ads can saturate that pool's attention without running into frequency issues. Advantage+ campaigns have no such constraint — the system is free to reach anyone on Meta who might convert. That means creative fatigue hits differently. The algorithm keeps finding new people, but if your creative pool is narrow, the same three hooks appear at the top of the rotation constantly. Frequency rises on your best performers, CTR drops, quality scores decline.

In manual campaigns, 5-8 ads is often sufficient. In Advantage+, you genuinely need 15-20.

How to Build and Rotate a 15-20 Ad Pool Without Burning Your Team

The bottleneck is almost never ideas. It's production velocity — how fast you can turn a brief into a finished ad that meets placement specs.

A sustainable process has three phases:

Phase 1: Weekly production cadence. Set a recurring goal of 3-5 net-new ads per week. This replaces the "monthly creative refresh" habit, which was adequate for manual campaigns but leaves Advantage+ campaigns running on stale creative for too long. At 3-5 new ads per week, your pool cycles completely every 3-6 weeks.

Phase 2: Performance-informed replacement. Don't rotate for the sake of rotation. Look at hook rate (the percentage of impressions that watch past 3 seconds) and thumbstop rate. Ads below your account's median on both metrics for two consecutive weeks are candidates for replacement — not necessarily pause, but replacement in the queue. The live post on how to use ad performance data to make better creatives covers the specific metrics worth tracking.

Phase 3: Learning preservation. This is where most teams lose progress. Pausing and re-uploading a winning ad resets its delivery learning — see why re-uploading a winning ad resets its learning. Keep top performers in rotation even as you add new challengers. The goal is a living pool that grows with winners and prunes consistent underperformers, not a full swap every month.

The Production Capacity Problem

Maintaining 15-20 distinct active ads requires a production system, not a production moment. That's the real constraint for most teams.

More than 8 million advertisers are already using Meta's AI creative tools as of mid-2026, and 63% of enterprise brands using AI-generated creative report 45-58% ROAS improvements. The direction is clear: teams that can't produce creative at volume will be outcompeted by teams that can.

The tools worth knowing about for production scale:

For static and video volume: Creatify (creatify.ai) and Arcads (arcads.ai) are both built around high-volume output. Arcads produces more than 100,000 advertising assets per month across 6,000+ clients. They're production tools — you still need to write the briefs and review the output.

For closed-loop production (generate → launch → measure → regenerate): Playad (playad.ai) — disclosure: this is our product — is built around the full cycle. Its Ad Monitor feeds performance signals back into creative generation so the system learns which hooks and formats are working in your specific account. Plans start at $39/month. The relevant limitation: it's best suited for teams that want the full workflow automated, not teams that want a standalone creative generator they'll feed into their own campaign management.

For creative intelligence (knowing what to make): MagicBrief (magicbrief.com) and Foreplay (foreplay.co) are strong for competitive research and creative briefing. They don't generate; they inform generation. Useful as inputs to whatever production tool you choose.

Comparison at a Glance

FAQ

How often should I add new creatives to an Advantage+ campaign?

Add 3-5 new ads per week, not in monthly batches. Advantage+ campaigns distribute spend dynamically, so creative fatigue can hit your top performers faster than it would in a manually targeted campaign. A weekly cadence keeps the pool fresh without triggering full learning resets.

Will running too many creatives hurt performance?

Above roughly 50 active ads per ad set, spend tends to fragment across too many variants for any single ad to accumulate meaningful delivery data. The algorithm needs volume on each creative to learn — thinly spread impressions slow that process. The 15-20 range is a practical ceiling for most accounts spending under $50k/month on Meta.

Do I need different creatives for different placements (Feed vs. Reels vs. Stories)?

With Advantage+ Placements enabled, Meta automatically adapts creatives across placements, but the adaptations are limited. A horizontal video won't perform on Reels regardless of how the system crops it. For Reels and Stories specifically, vertical 9:16 creative outperforms adapted horizontal creative in most accounts. Treat vertical-native creative as a distinct hook category, not as a placement variant of existing assets.

Can I use the same creative pool for prospecting and retargeting in Advantage+?

Advantage+ Shopping Campaigns (ASC) blend prospecting and retargeting into a single pool by default. This is one of the documented criticisms of the format — reported ROAS can be inflated by high-converting existing customers who would have bought anyway. If you need separate measurement of prospecting vs. retargeting performance, you'll need to use the "existing customer budget cap" setting in ASC or maintain separate manual campaigns for retargeting. See the measure black-box campaigns post for measurement approaches that work when the algorithm won't tell you what it's doing.

What counts as a "distinct hook" for Advantage+ purposes?

The hook is the first 2-3 seconds of a video or the dominant visual of a static ad. Two ads have distinct hooks if a viewer who saw one would not recognize the other as "the same ad." Different voiceover on the same opening visual is not a distinct hook. Different problem framing, different opening visual, or different talent/character all qualify. When in doubt, ask: would this ad reach a different person than my other ads? If yes, it's distinct.