genpark-growth-agent

GenPark Growth Agent is a zero-dependency quantitative growth engineering and experimentation engine implementing viral branching process epidemic modeling ($K$-factor), SaaS cohort retention and CAC payback decay analysis, and Wald's Sequential Probability Ratio Test (SPRT) for continuous A/B test early stopping.

flowchart TD
    A[Invites per User & Conversion] --> B[Branching Process K-Factor Engine]
    B --> C[Viral Cascade Simulation]
    D[ARPU, Churn & Gross Margin] --> E[Cohort Retention Decay Model]
    E --> F[LTV/CAC Ratio & Payback Period]
    G[Continuous Conversion Stream] --> H[Wald SPRT Log-Likelihood Ratio]
    H --> I[Early Stopping Boundary: Accept/Reject/Continue]

Features

• Branching Viral Modeling: Calculates exact geometric cascade curves and super/sub-critical thresholds ($K = i \times c$).
• SaaS Unit Economics: Models cohort retention decay curves, net lifetime value ($LTV = ARPU \cdot Margin / Churn$), and CAC payback months.
• Wald's SPRT Early Stopping: Minimizes experimentation sample size and sample waste by terminating continuous A/B tests immediately upon significance.

Quickstart

from client import GenParkGrowthAgent

growth = GenParkGrowthAgent()
res = growth.simulate_viral_k_factor_branching(initial_users=1000, invites_per_user=3.5, invite_conversion_rate=0.35)
print("K-Factor:", res["k_factor"], "Total Reached:", res["total_users_reached"])

License

MIT