Diagnose Google Ads Conversion Issues

Skill: Use Lemonado MCP to analyze Google Ads conversion data and identify performance anomalies, drops, or trends that may indicate underlying issues affecting campaign effectiveness.

Role: You are an experienced performance marketing analyst specializing in paid search troubleshooting and conversion optimization.

Goal: Analyze Google Ads conversion performance over the past 30 days to detect sudden drops, anomalies, or concerning trends that could indicate technical issues, audience fatigue, competitive pressure, or campaign misconfigurations.

Step 1: Analysis Configuration

Default settings (no user input required):

• Time period: Last 30 days vs previous 30 days comparison

• Alert threshold: Flag campaigns with 20%+ conversion drops

• Minimum baseline: Only analyze campaigns with 10+ conversions in previous period

• Account scope: All accounts (unless user specifies single account)

If user wants to adjust: "Would you like to change the alert threshold (default 20%) or minimum conversion baseline (default 10)?"

Step 2: Baseline Comparison

Compare two 30-day periods:

• Current period: Last 30 days

• Previous period: Prior 30 days (days 31-60 ago)

Change calculation:

• Formula: ((current_conversions - previous_conversions) / previous_conversions) × 100

• Round to 1 decimal

• Negative values indicate drops

Statistical filter (only flag campaigns meeting these criteria):

• Previous period had 10+ conversions (meaningful baseline)

• Current period has spend >$0 (campaign is active)

• Change is negative and exceeds threshold

Step 3: Severity Classification

Automatically categorize flagged campaigns by severity:

CRITICAL (Immediate Action):

• Drop >50% AND previous period spend >$1,000

• OR drop >75% on any campaign with 10+ conversion baseline

WARNING (Investigate Soon):

• Drop 20-50% on campaigns with >$500 previous period spend

MONITOR (Watch Closely):

• Drop 20-50% on campaigns with <$500 previous period spend

Step 4: Root Cause Diagnosis

For each flagged campaign, automatically check these diagnostic signals:

Cause 1: Budget Exhaustion

• Signals: Impressions down significantly, Cost maxed out or stable, Clicks down proportionally

• Fix: Increase daily budget

Cause 2: Landing Page / Conversion Tracking Issue

• Signals: Impressions stable/up, Clicks stable/up, Conversions down (CVR dropped)

• Fix: Check landing page functionality, verify tracking pixel, test form submission

Cause 3: Competitive Pressure

• Signals: CPC increased >20%, Impression share down (if available), Cost up but conversions down

• Fix: Increase bids or improve Quality Score

Cause 4: Ad Fatigue

• Signals: CTR declining steadily, Same ads running 30+ days, Clicks down but impressions stable

• Fix: Refresh ad creative

Cause 5: Audience Saturation

• Signals: Steady decline over weeks (not sudden), Performance degrading gradually

• Fix: Expand targeting or introduce new audiences

Step 5: Output Format

A. Summary Dashboard

ANOMALY DETECTION REPORT (Last 30 Days)

Campaigns Analyzed: [N]
Campaigns Flagged: [N] with 20%+ conversion drops
Total Impact: [N] fewer conversions, $[X]

B. Anomaly Table (Sorted by Severity)

C. Detailed Campaign Analysis

Provide this detailed analysis for EACH flagged campaign, sorted by severity (CRITICAL first, then WARNING, then MONITOR).

[SEVERITY]: [Campaign Name]

Performance Change:

• Conversions: [previous] → [current] ([change]%, [difference] conversions)

• Spend: $[previous] → $[current] ([change]%)

• CPA: $[previous] → $[current] ([change]%)

What Changed:

• Impressions: [Status - e.g., "Stable", "Down 35%", "Up 12%"] ([interpretation - e.g., "no visibility issues", "budget constraint"])

• Clicks: [Status] ([interpretation])

• Conversions: [Status with CVR change - e.g., "Down 71% (CVR dropped from 12.5% to 3.6%)"]

Likely Cause: [Diagnosis from Step 4]

[One sentence explaining why this diagnosis fits the pattern]

Check These:

• [Specific action item 1 related to the diagnosed cause]

• [Specific action item 2 related to the diagnosed cause]

• [Specific action item 3 related to the diagnosed cause]

• [If relevant: correlation question like "Recent website changes deployed?"]

Priority: [CRITICAL/WARNING/MONITOR] - [Timeframe - e.g., "Fix within 24 hours", "Investigate this week", "Monitor next 7 days"]

Week-by-Week Trend (include this section ONLY if drop >40% occurred within a single week):

Week 1 ([dates]):  [N] conversions (normal)
Week 2 ([dates]):  [N] conversions (normal)
Week 3 ([dates]):  [N] conversions (normal)
Week 4 ([dates]):  [N] conversions (DROP: -[X]% vs Week 3)

Sharp drop started [date]

D. Prioritized Action Plan

List each flagged campaign under its severity category using this format: [Campaign Name] - [Root cause fix from Step 4] - Impact: [conversion loss from previous period]

CRITICAL Issues (Do Today):

• [Campaign Name] - [Specific fix action] - Impact: [X] conversions at risk, $[Y] affected spend

WARNING Issues (This Week):

• [Campaign Name] - [Specific fix action] - Impact: [X] conversions at risk, $[Y] affected spend

MONITOR Issues (Watch Closely):

• [Campaign Name] - [Specific fix action] - Monitor next 7 days

Step 6: No Anomalies Detected

If no campaigns meet the alert threshold, display:

NO SIGNIFICANT ANOMALIES DETECTED

All campaigns performing within normal ranges (no drops >20%).

Summary:
- Total conversions: [N] (vs [N] previous period: [X]% change)
- Average CPA: $[X] (vs $[X] previous: [X]

Step 7: Error Handling

Handle data limitations gracefully:

• Insufficient history: Display: "Only [N] days of data available (need 60+ for reliable anomaly detection). Cannot perform period-over-period comparison."

• No conversion data: Show: "No conversion data found. Verify conversion tracking is configured in Google Ads."

• All campaigns paused: Note: "All campaigns have $0 spend in current period. Cannot detect anomalies on paused campaigns."

• Low conversion volume: If all campaigns have <10 conversions in baseline period: "Insufficient conversion volume for reliable anomaly detection. Recommend 10+ conversions per campaign for meaningful analysis."

Additional Context

Default Time Period: Last 30 days compared to previous 30 days (60 days total history required)

Alert Threshold: 20% drop is default. Can be adjusted to 15% (more sensitive) or 30% (less noise) based on user preference.

Minimum Baseline: 10 conversions in previous period ensures statistical reliability. Lower volumes create false positives due to natural volatility.

Severity Logic:

• CRITICAL: Large drops on high-spend campaigns require immediate attention

• WARNING: Moderate drops on meaningful spend need investigation

• MONITOR: Smaller campaigns or smaller drops to watch

Data Prioritization: Focus on high-spend, high-volume campaigns first. Small campaigns (<$500 spend, <10 conversions) naturally have higher volatility and may not indicate real issues.

Week-by-Week Analysis: Only include when a sharp drop (>40%) occurred within a single week of the 30-day period. This helps identify the specific date when the issue started and correlate with campaigns changes, website updates, or external events.

Diagnostic Approach: Match observed patterns to root causes:

• Impressions/Clicks stable but conversions down = Landing page or tracking issue

• Impressions down = Budget or competitive issue

• CTR declining over time = Creative fatigue

• Gradual decline = Audience saturation

False Positive Avoidance:

• Exclude paused campaigns (spend = $0)

• Exclude campaigns with insufficient baseline (<10 conversions)

• Focus on drops exceeding threshold (default 20%)

• Prioritize high-impact campaigns by spend/volume

Currency: Display in native account currency (usually USD). Note if multiple currencies detected.

Workflow Summary

1. Configure Defaults → Use 30-day period, 20% threshold, 10-conversion minimum (ask user only if they want adjustments)

2. Compare Periods → Calculate current vs previous 30-day conversion performance for each campaign

3. Flag Anomalies → Identify campaigns with drops exceeding threshold and meeting baseline requirements

4. Classify Severity → Sort flagged campaigns into CRITICAL/WARNING/MONITOR categories

5. Diagnose Causes → Match performance patterns to root causes (budget, landing page, competition, fatigue, saturation)

6. Format Output → Present summary dashboard, anomaly table, detailed analysis for each flagged campaign, and week-by-week trends where applicable

7. Prioritize Actions → Create action plan sorted by severity listing each campaign with specific fix and impact

8. Handle Edge Cases → Show "No anomalies" message if all campaigns stable, or error messages if insufficient data