Detection: Active prospecting running on too few creative concepts

Key: creative_volume_floor Severity: Medium Confidence: 80% Shipped: v1.2, May 2026 Runtime: Requires populated creative IDs (Meta OAuth path). Self-skips on CSV exports because the parser sets creative_id="unknown".

What this detection looks for

We fire one account-level finding when all three of these are true:

The account has at least one active ad with positive spend AND a real creative ID (not "unknown")
Audit-window spend projects to $2,000/month or more
The count of distinct creative IDs across active, spending ads is less than 6

Why this matters

Pilothouse's "3-3-3" framework, Sam Tomlinson's 2026 audit, and Common Thread Collective's playbooks all converge on roughly six active concepts as the floor below which fatigue is guaranteed inside 3–4 weeks regardless of how good any single creative is. The math is mechanical: an audience exposed to the same 3–4 creatives at frequency 5+ runs out of attention; CTR collapses; CPA inflates; ROAS follows.

This rule is the inverse of creative_concentration_risk:

creative_concentration_risk catches the after — one creative holding >70% of conversions in a campaign that has plenty of ads.
creative_volume_floor catches the before — the account only has 3–4 distinct creatives running at all, so the math is fragile even when no single ad is over-served yet.

A real account often has both signals together; the two rules are complementary.

How we estimate the recoverable dollars

monthly_at_risk = (total_active_ad_spend × 30 / audit_days) × 0.05

This is a forward-looking estimate, not current-month bleed. We project total active-ad spend to monthly and take 5% as the "next 30 days" exposure — that's the amount at risk when the main creative fatigues. The actual loss may be larger; 5% is a conservative floor.

What would change our mind

DCO accounts. Dynamic Creative Optimization automatically generates many variants from a small input set. Some accounts genuinely run on 3 creatives plus DCO and produce dozens of effective variants. The "distinct creative ID" count under-represents what's actually being shown.
Premium-product accounts. Some accounts (high-ticket B2B, luxury) run very few, deeply-tested winners and rotate slowly. The 6-concept floor is the wrong heuristic for that profile.
Recent account launch. A 2-week-old account hasn't had time to test 6 concepts yet — fragility is real but the rule wants to flag it anyway. Treat as informational, not urgent.

What to do about it

Audit your last 30 days of ad uploads. If you uploaded 3 concepts and tested in-platform variants, that's the data this rule sees.
Schedule 4–6 new concepts in the next sprint. The Pilothouse 3-3-3 framework (3 hooks × 3 angles × 3 formats) is a structured way to fill the floor without scattershot creative.
If you're using DCO heavily, this finding is informational — the underlying concept count is real but the in-platform variant pool may be larger than the rule sees.

References

Pilothouse: "Meta Creative Testing Framework — the 3-3-3 Approach to Finding Winners"
Sam Tomlinson: "The Ultimate Meta Ads Account Audit, Part II"
Common Thread Collective: 27-Point Facebook Ads Audit