TL;DR: Every Meta operator right now is being told to push 20, 30, 40 new creatives a month. That number is meaningless in isolation. Creative volume is a function of your monthly spend, not a fixed mandate. If you're a startup spending $10K/month running 50 new ads, you're not doing Andromeda well; you're wasting production budget on noise. Common Thread Co. refreshed the operating math last week. Under $10K/month you need 8-12 new distinct creatives; $10K – $50K/month you need 15-25; $50K+ you need 25-40+ with weekly refresh cycles. Baseline heuristic: roughly one new distinct creative per $1,000 of monthly spend. Inside that, split at least 40% video, 60% static. And most importantly, variety beats volume every time.
At Third Pillar Marketing Group, we work with growth-focused founders across health-tech, consumer hardware, spirits, and DTC. Every one of them is running Meta at some scale, and every one of them is being told the same thing this quarter: push more creative. Volume wins. Feed Andromeda. That advice is right in principle and dangerously incomplete in practice.
The "20-30 conceptually distinct creatives per ad set" mandate that's been circulating is a $10K – $50K/month tier benchmark. It got extracted from that context and turned into a universal target, which is how founders end up burning production budget on a volume they can't actually feed with signal. If your account can only afford $200/day, Andromeda cannot possibly learn from 40 new creatives in 30 days. Each one gets a handful of impressions. Nothing gets tested. Everything gets clustered. You lose money and you lose the ability to trust your results.
Common Thread Co. refreshed their Andromeda operating manual last week. The math they published is the cleanest version I've seen.
Under $10K/month in Meta spend: 8-12 new conceptually distinct creatives per month. That's roughly two to three per week. Enough to keep the learning phase fed without starving spend.
$10K – $50K/month: 15-25 new creatives per month. Four to six per week. This is where most of our clients live. It is also the tier the "20-30" number came from.
$50K+/month: 25-40+ new creatives per month, with weekly refresh cycles. This tier can absorb sprint launches and needs them.
The heuristic worth memorizing: roughly one new distinct creative per $1,000 of monthly spend. It tapers to one per $2,000 as accounts scale into six figures monthly, depending on catalog and performance. But the $1K baseline is the right anchor to start every conversation.
Inside your monthly volume, split at least 40% video and 60% static. Static is cheaper to produce and faster to iterate. Test your angles in static first. Let the winning static angles inform which videos you commission. This is the opposite of how most brands work. Most brands commission the hero video, then cut static from it. That approach starves you of angles, because you're testing the same story in different frames.
Andromeda uses a Creative Similarity Score. Once your ads cross roughly 60% similarity to each other, Andromeda collapses them into a single entity for competition purposes. It routes them all through the same retrieval node. You are, in effect, running one ad against itself while paying for four.
Meta does not surface that score to advertisers. Third parties like Triple Whale offer a version of it. But you don't need a tool. Common Thread Co. wrote the plain-language test that's stuck in my head all week:
Open your active ads. Try to describe every one of them with the same sentence structure. "[Hero] with [background] and [text angle]." If more than 60% of your ads fit the same sentence, you have an entity clustering problem. You do not have four creatives; you have one creative, expensively duplicated.
Real variety means different angles inside each format. For video, that means UGC, founder or educational, brand lifestyle, and text-over-b-roll style. For static, that means testimonial, comparison, bold headline, and lifestyle aesthetic. If your creative brief comes back with four crops of the same product shot and three colorways of the same headline, you have a production problem, not a testing program.
Common Thread Co. closes the article with a 5-step Andromeda audit. It maps cleanly to any account, with one important swap for service and healthcare founders.
Step 1: Audit your creative similarity score. Run the one-sentence test above. Do it monthly.
Step 2: If you're ecommerce, count your active Catalog Ads. Under 40 is the danger zone. Design Rules matter; the top-vs-bottom performer gap on this single feature is 103%. If you're service industry (Forta, agencies, telehealth, any lead-gen business): swap this step for platform lead form ads with a scheduling-tool hookup. Meta shipped Book Time integration inside Instant Forms last week. For service accounts, that is your Catalog Ads equivalent. The hard part is qualifying leads without collecting PHI, which is a real design constraint but not a reason to skip the lever.
Step 3: Consolidate your campaign structure. If you're running more than 3-4 Meta campaigns and none of them is hitting 20 conversions per week, you're fragmenting your signal. Andromeda needs a data-dense pool. Consolidate the best-performing ads into a single broad-targeted campaign.
Step 4: Build a creative cadence, not a creative sprint. Ad fatigue window is 2-4 weeks in Andromeda. Minimum 2-3 new conceptually distinct creatives entering the pipeline per week at moderate spend, more at higher spend. This is an operating system, not a campaign launch.
Step 5: Change your success metric. Stop optimizing ad-level ROAS. Start optimizing account-level aMER and measure creative against incremental ROAS (iROAS), not platform attribution. Some of your best creatives are building the funnel, not closing it. Kill them without understanding what they're doing, and you break your top-of-funnel.
Want to right-size your creative volume and audit your diversity this week? Kick off with this prompt in Claude.
You are my fractional CMO. I need help right-sizing my Meta creative
volume and auditing my current active ads for real diversity, using
the Common Thread Co. Andromeda operating manual (spend-relative
volume tiers + one-sentence similarity test + 5-step audit).
Context I'll give you:
- Business: [describe your business, category, offer]
- Monthly Meta spend: [amount]
- Category: [ecommerce / service / lead-gen / other]
- Current monthly creative production output: [# new distinct
concepts, not variants]
- Current format split: [% video / % static]
- Number of active ad sets: [count]
- Number of active campaigns: [count]
- Current best-performing conversion event: [event name + weekly count]
- Do you have Book Time / scheduling integration inside your lead
form today: [yes/no]
Run this audit and give me a concrete action list, not a summary.
Ask me one clarifying question at a time when you need more info.
1. VOLUME RIGHT-SIZING.
Using my monthly spend, place me in the correct CTC tier (Under $10K
= 8-12 creatives/mo; $10K-$50K = 15-25; $50K+ = 25-40+). Compare
against my actual production output. Tell me if I'm over-producing
(wasting budget on noise), under-producing (starving Andromeda), or
right-sized. Recommend a specific monthly volume target.
2. FORMAT SPLIT AUDIT.
Compare my current video/static split against the 40/60 minimum.
Tell me the exact number of new videos and new statics per month
I should be commissioning based on my right-sized volume target
from Step 1.
3. SIMILARITY SCORE HEURISTIC.
Give me a copy-paste template of the one-sentence test:
"[hero] with [background] and [text angle]." Then ask me to describe
my current active ads using that structure. Score my diversity: what
percentage of ads fit the same sentence? Flag which ads should be
retired or reworked.
4. STEP 2 SUBSTITUTION FOR SERVICE ACCOUNTS.
If I'm ecommerce, count my active Catalog Ads and check whether
Design Rules are enabled. If I'm service or lead-gen, tell me whether
I have platform lead form ads with a Book Time / Calendly hookup
active today, and if not, that's my #1 priority for the quarter.
5. CAMPAIGN CONSOLIDATION.
Given my campaign count and per-campaign conversion volume, tell me
whether to consolidate. If yes, name the consolidation target
structure. If no, tell me what threshold triggers consolidation later.
6. CADENCE AND MEASUREMENT.
Design a weekly production and testing cadence for the volume from
Step 1. Tell me what to measure: aMER at what interval, iROAS
methodology given my conversion volume, and what NOT to use
(ad-level ROAS as primary decision metric).
Output format: numbered action list with specific numbers, not
ranges. No preamble. Direct recommendations.
If you want a version of this prompt already loaded with your account structure, spend history, and current creative library, that's the kind of starting artifact we build for our fractional-CMO clients.
The gap between the operators winning at Meta right now and the ones burning production budget is not talent or spend. It is context. Knowing that 20-30 creatives per ad set is a $10K – $50K tier number and not a universal mandate. Knowing that Meta doesn't publish the Similarity Score. Knowing that the fastest CPL win for service accounts this quarter is a lead form with Book Time, not another round of creative variants. That kind of context sits on top of a full-time job of watching platforms, translating updates, and holding accounts accountable to the plays that emerge.
Most founders don't need a full-time marketing hire. They need a leader who can watch the platforms, run the plays, and manage the vendors already spending against those platforms.
That's what we do at Third Pillar. If your Meta creative production feels like it's either not enough or too much, we should talk.
Ready to talk about whether your creative volume is right for your spend? Book a call.
Product and operations are handled. If marketing is the missing piece, let's talk.