Cross-Platform Budget Optimizer

Analyzes ROAS performance across all connected ad platforms and generates budget reallocation recommendations to maximize overall return. Supports Google Ads, Meta Ads, LinkedIn Ads, X Ads, Reddit Ads, TikTok Ads, Amazon Ads, The Trade Desk, and Amazon DSP.

Safety Guardrails

Read-only by default. This skill:

• Analyzes performance data and outputs recommendations only
• Never auto-applies budget changes without --apply flag
• Requires explicit --confirm for applying changes over threshold
• Generates detailed reports before any modifications

Capabilities

1. Cross-Platform ROAS Analysis - Pull and compare ROAS across all platforms
2. Budget Reallocation Recommendations - Suggest moving spend from low to high ROAS platforms
3. Campaign-Level Optimization - Identify top/bottom performers within each platform
4. Constraint-Aware Planning - Respect minimum budgets, daily caps, and platform limits
5. Historical Trend Analysis - Consider 7/14/30-day ROAS trends
6. Apply Budget Changes - Execute approved reallocations across platforms

Workflow

Step 1: Pull Performance Data

Fetch ROAS data from all connected platforms:

python3 scripts/pull_platform_metrics.py \
  --platforms all \
  --date-range 30d \
  --output metrics.json

Step 2: Analyze & Generate Recommendations

python3 scripts/analyze_budgets.py \
  --input metrics.json \
  --target-roas 3.0 \
  --min-budget 10 \
  --reallocation-threshold 0.15 \
  --output recommendations.json

Step 3: Preview Recommendations

python3 scripts/preview_reallocation.py \
  --recommendations recommendations.json

Step 4: Apply Budget Changes (Optional)

python3 scripts/apply_budget_changes.py \
  --recommendations recommendations.json \
  --confirm \
  --dry-run

Remove --dry-run to execute changes:

python3 scripts/apply_budget_changes.py \
  --recommendations recommendations.json \
  --confirm

Common Operations

Quick ROAS check across all platforms

python3 scripts/pull_platform_metrics.py \
  --platforms all \
  --date-range 7d \
  --summary-only

Optimize for a specific ROAS target

python3 scripts/analyze_budgets.py \
  --input metrics.json \
  --target-roas 4.0 \
  --aggressive \
  --output recommendations.json

Reallocate between specific platforms

python3 scripts/analyze_budgets.py \
  --input metrics.json \
  --platforms google,meta,linkedin \
  --target-roas 3.0 \
  --output recommendations.json

Weekly optimization report

python3 scripts/generate_optimization_report.py \
  --date-range 7d \
  --format markdown \
  --output weekly_report.md

Reallocation Algorithm

The optimizer uses a ROAS-weighted reallocation algorithm:

1. Calculate Platform Scores: score = ROAS × conversion_volume_weight
2. Identify Donors: Platforms with ROAS < target × 0.8
3. Identify Recipients: Platforms with ROAS > target × 1.2
4. Calculate Shifts: Proportional to score difference, capped at thresholds
5. Apply Constraints: Minimum budgets, max reallocation %, platform caps

Score Formula

platform_score = (
    roas_normalized × 0.6 +
    conversion_volume_normalized × 0.25 +
    trend_score × 0.15
)

Where:

• roas_normalized = platform_roas / max(all_roas)
• conversion_volume_normalized = platform_conversions / max(all_conversions)
• trend_score = 7d_roas_change / 30d_avg_roas

Parameters

Scripts Reference

Output Format

Recommendations JSON

{
  "generated_at": "2026-01-18T10:00:00Z",
  "analysis_period": "30d",
  "target_roas": 3.0,
  "total_budget": 1500.00,
  "platforms": {
    "google": {
      "current_budget": 500.00,
      "recommended_budget": 650.00,
      "change": 150.00,
      "change_pct": 30.0,
      "roas": 4.2,
      "conversions": 125,
      "revenue": 2100.00,
      "spend": 500.00,
      "trend_7d": 0.15
    },
    "meta": {
      "current_budget": 400.00,
      "recommended_budget": 350.00,
      "change": -50.00,
      "change_pct": -12.5,
      "roas": 2.1,
      "conversions": 84,
      "revenue": 840.00,
      "spend": 400.00,
      "trend_7d": -0.08
    }
  },
  "reallocations": [
    {
      "from_platform": "meta",
      "to_platform": "google",
      "amount": 50.00,
      "reason": "Google ROAS 4.2x vs Meta 2.1x"
    },
    {
      "from_platform": "reddit",
      "to_platform": "google",
      "amount": 100.00,
      "reason": "Reddit ROAS 1.5x vs Google 4.2x"
    }
  ],
  "estimated_impact": {
    "current_blended_roas": 2.8,
    "projected_blended_roas": 3.4,
    "projected_revenue_increase": 450.00,
    "confidence": "medium"
  }
}

Platform-Specific Notes

Google Ads

• Uses GAQL to query campaign performance
• Includes Search, Display, pMax campaign types
• Respects shared budgets

Meta Ads

• Queries campaign insights via Graph API
• Combines Facebook and Instagram data
• Handles CBO (Campaign Budget Optimization) campaigns

LinkedIn Ads

• Rest.li 2.0 protocol for metrics
• Includes Sponsored Content and Message Ads
• B2B conversion tracking

X Ads

• Campaign analytics via v12 API
• Includes promoted tweets and video views
• Engagement-to-conversion attribution

Reddit Ads

• API v3 for campaign metrics
• Includes subreddit targeting performance
• User-Agent header required

TikTok Ads

• Business API for reporting
• Video-first metrics (VTR, engagement)
• Interest targeting performance

Amazon Ads

• Sponsored Products/Brands/Display metrics
• ACOS (Advertising Cost of Sale) converted to ROAS
• Product-level attribution

The Trade Desk

• Programmatic display, video, CTV, audio
• Uses delta sync for efficient metric pulls
• Campaign-level budget optimization
• Supports GraphQL for complex queries

Amazon DSP

• Programmatic display and video on Amazon
• Async reporting (request → poll → download)
• Order-level budgets (lifetime, not daily)
• Region-aware (NA, EU, FE)

Minimum Budget Requirements

Environment Variables

Required environment variables:

# Platform Connection IDs

# Platform Account IDs
GOOGLE_ADS_CUSTOMER_ID
META_AD_ACCOUNT_ID
LINKEDIN_AD_ACCOUNT_ID
TWITTER_AD_ACCOUNT_ID
REDDIT_AD_ACCOUNT_ID
TIKTOK_ADVERTISER_ID
AMAZON_ADS_PROFILE_ID
TTD_ADVERTISER_ID
AMAZON_DSP_PROFILE_ID
AMAZON_DSP_ADVERTISER_ID

# Google Ads specific
GOOGLE_ADS_DEVELOPER_TOKEN
GOOGLE_ADS_LOGIN_CUSTOMER_ID

Error Handling

Best Practices

1. Run weekly - Weekly optimization cycles balance data freshness with stability
2. Start conservative - Use --conservative flag initially
3. Review before applying - Always preview recommendations first
4. Respect learning periods - Don't optimize campaigns in learning phase
5. Consider seasonality - Adjust targets for seasonal patterns
6. Monitor after changes - Watch performance 48-72 hours after reallocation
7. Set realistic targets - Base target ROAS on historical blended performance
8. Keep minimum budgets - Never starve platforms completely

Integration with Other Skills

• campaign-preflight - Run before applying budget changes
• multi-channel-reporting - Generate reports after optimization
• bid-optimization - Optimize bids within reallocated budgets
• cross-platform-launcher - Deploy new campaigns with optimized allocation

About this skill

Part of the Synter free skills collection — open-source agent skills for advertising, PPC, and marketing automation.

Want this skill (and 40+ more) running on autopilot against your live ad accounts? Try Synter — AI Agent Media Buyers that connect to Google, Meta, LinkedIn, TikTok, Reddit, Amazon, and 7+ more platforms.

• 🌐 syntermedia.ai
• 📚 All free skills
• 💬 Built by @syntermedia — questions? Open an issue.