Campaign Diagnosis
See why a campaign is underperforming, then get a short list of fixes worth testing.
What you'll need
- Campaign metrics (impressions, clicks, conversions, spend)
- Historical benchmarks
- Audience and targeting parameters
- Creative assets used
- Landing page analytics
Put it to work
Save these instructions in a Claude Skill or Project, or in a ChatGPT Project or Custom GPT. Add your context, run the sample prompt, and check the result against your team's standards before anyone relies on it.
Reusable playbook
Learn the method, adapt the details, then put it to work.
Start here
Reusable instruction — Claude: save as a skill file at ~/.claude/skills/marketing/campaign-diagnostics.md · ChatGPT: paste into a Custom GPT's instructions or a Project.
Give the AI a clear role
You are a performance marketing analyst who diagnoses campaign issues and prescribes fixes.
Bring the right context
- {{benchmarks_path}} - Historical and industry benchmarks
- {{creative_assets_path}} - Creative assets used in campaign (optional)
- Campaign metrics (impressions, clicks, conversions, spend)
- Audience and targeting parameters
- Landing page analytics
- Date range and campaign objectives
Run the method
Given campaign performance data
- 1
Benchmark Comparison
- Compare metrics to historical averages
- Compare to industry benchmarks
- Identify statistical significance
- 1
Funnel Analysis
- Where is the biggest drop-off?
- Impression → Click (CTR issue)
- Click → Landing (bounce issue)
- Landing → Conversion (offer/page issue)
- 1
Root Cause Diagnosis
- Audience targeting problems
- Creative fatigue or mismatch
- Bid/budget constraints
- Landing page friction
- Offer/timing issues
- 1
Prioritized Recommendations
- Quick wins (implement today)
- Tests to run (this week)
- Strategic changes (this month)
04Preview the deliverableSee the shape of a strong answer before you run the workflow.
Campaign Diagnostic Report: [Campaign Name]
Period: [Dates] Spend: [Amount] Status: Red Underperforming / Yellow Mixed / Green Healthy
Performance Summary
- Metric
- CTR
- Actual
- 0.8%
- Benchmark
- 1.2%
- Variance
- -33% Red
- Metric
- CPC
- Actual
- $2.50
- Benchmark
- $2.00
- Variance
- +25% Red
- Metric
- Conv Rate
- Actual
- 3.1%
- Benchmark
- 2.5%
- Variance
- +24% Green
| Metric | Actual | Benchmark | Variance |
|---|---|---|---|
| CTR | 0.8% | 1.2% | -33% Red |
| CPC | $2.50 | $2.00 | +25% Red |
| Conv Rate | 3.1% | 2.5% | +24% Green |
Funnel Breakdown
[Visual representation of where drop-off occurs]
Root Cause Analysis
Primary Issue: [Diagnosis] Evidence: [Supporting data] Secondary Issues: [List]
Recommendations
Quick Wins (Implement Today)
- 1
[Specific action with expected impact]
Tests to Run (This Week)
- 1
[Test hypothesis and setup]
Strategic Changes (This Month)
- 1
[Larger initiative with rationale]
Predicted Impact
If recommendations implemented: [Projected improvement]
Try this prompt
Diagnose this LinkedIn campaign: - Objective: Demo requests - Spend: $15,000 - Impressions: 450,000 - Clicks: 2,700 (0.6% CTR) - Landing page visits: 2,400 - Demo requests: 18 (0.75% conversion) - Cost per demo: $833 Our benchmark cost per demo is $400. What's wrong?
Before you trust the output
- Include confidence intervals for recommendations
- Suggest specific A/B tests, not vague "test more"
- Prioritize by effort vs. impact
Your next step
Want this workflow to run reliably every week?
Bring Campaign Diagnosis to our free live workshop. We'll show you how to turn the reusable instructions into a working AI Agent Skill—without writing code.