genpark-cross-platform-allocator

GenPark Cross-Platform Allocator is a zero-dependency quantitative paid media and performance marketing optimization engine implementing Markowitz Modern Portfolio Theory for ad budget distribution, Hill function diminishing marginal ROAS curves, and automated cross-channel spend drift rebalancing.

flowchart LR
    A[Channel ROAS & Volatilities] --> B[Markowitz Mean-Variance Portfolio Engine]
    B --> C[Risk-Adjusted Optimal Budget Share]
    D[Incremental Channel Spend] --> E[Hill Function Saturation Model]
    E --> F[Marginal ROAS & Diminishing Return Warning]
    G[Target Allocation vs Actual Spend] --> H[Cross-Platform Rebalancing Engine]
    H --> I[Pacing Correction & Capital Reallocation]

Features

• Markowitz Mean-Variance Ad Portfolio: Optimizes multi-channel spend based on expected ROAS and channel volatility trade-offs.
• Hill Function Saturation Modeling: Detects marginal efficiency decay ($dRev / dB$) to prevent ad spend waste above saturation thresholds.
• Cross-Platform Rebalancing: Identifies pacing drift across Google, Meta, TikTok, and Amazon to issue corrective capital adjustments.

Quickstart

from client import GenParkCrossPlatformAllocator

allocator = GenParkCrossPlatformAllocator()
res = allocator.optimize_markowitz_ad_budget(
    [{"name": "Google", "expected_roas": 4.5, "roas_volatility_std": 0.3},
     {"name": "Meta", "expected_roas": 3.8, "roas_volatility_std": 0.5}],
    total_budget_usd=10000.0
)
print("Blended ROAS:", res["blended_portfolio_roas"])

License

MIT