Marketing & Advertising / Facebook Ads
Diagnose what’s wrong with your Facebook Ads using CTR, CPC, CPM, conversion rate, and frequency — then get a specific fix, a test alternative, and a kill recommendation.
Why This Prompt Exists
Most ad audits fail because they guess instead of diagnose.
You get:
- “Make the ad better” (useless advice)
- no link between metrics and specific problems
- fixes that don’t address the root cause
- no kill recommendation — so money keeps burning
- advice that works for one metric but breaks another
But ad performance is not mysterious.
Metrics point to specific problems.
- Low CTR = hook or creative problem
- High CPM + high frequency = audience fatigue
- Low conversion rate despite good CTR = offer or landing page problem
- Every metric has a lever — pull the right one
Without diagnosis, you optimize in the dark.
This framework forces AI to be a performance doctor who prescribes specific fixes.
The Prompt
Assume the role of a Facebook Ads auditor who reads metrics like a doctor reads vitals. Your task is to diagnose what's wrong with an ad based on its performance data. Generate: 1. DIAGNOSIS (one sentence) Hook problem / Creative problem / Audience problem / Offer problem / Creative fatigue 2. SPECIFIC FIX What to change (e.g., "Rewrite the hook using a curiosity gap") 3. TEST THIS INSTEAD ALTERNATIVE A specific variation to test against the current ad 4. KILL RECOMMENDATION Should this ad be turned off entirely? YES / NO — with rationale INPUTS: CTR (Click-Through Rate): [INSERT %] CPC (Cost Per Click): [INSERT $] CPM (Cost Per 1,000 Impressions): [INSERT $] Conversion Rate (if known): [INSERT % OR "UNKNOWN"] Frequency (avg. times same person saw ad): [INSERT NUMBER] Ad Spend to Date: [INSERT $] RULES: - Diagnosis must be one of the five categories - The fix must be specific, not "improve the creative" - If frequency > 3 and CPM is rising, diagnose creative fatigue - If CTR < 0.5%, diagnose hook problem first - Kill recommendation must have a numeric threshold (e.g., "If CTR doesn't improve to 1% by $500 spend, kill")
How To Use It
- Run this diagnosis every $500-$1,000 in spend per ad.
- If multiple metrics are bad, diagnose in this order: CTR → CPM → Conversion Rate.
- The kill recommendation is the most valuable output — use it.
- Save diagnoses to build a pattern library for your account.
- If you disagree with the diagnosis, the data might be insufficient (wait for more spend).
Example Input
CTR: 0.4%
CPC: $1.85
CPM: $7.40
Conversion Rate: 2.1%
Frequency: 1.2
Ad Spend to Date: $450
Why It Works
Most ad optimization fails because it's guessing, not diagnosing.
This framework improves outcomes by forcing:
- metric-specific diagnosis (hook, creative, audience, offer, fatigue)
- specific fixes (not general advice)
- test alternatives for validation
- kill recommendations with numeric thresholds
- prioritization of which metric to fix first
Great ad performance doesn't come from working harder — it comes from knowing which lever to pull.
Build Better AI Systems
Subscribe for advanced prompt engineering, AI marketing tools, Facebook Ads frameworks, and practical strategies for advertisers and business owners.