A Study on Data-Driven Budget Optimization for U.S. Enterprises’ Cross-Border Marketing

Abstract

This study constructs an integrated theoretical framework of “Dynamic Institution-Capability-Resource Allocation,” employing a double machine learning dynamic panel model with 1.5 million project-level data points to systematically examine the optimization mechanisms of cross-border marketing budgets in U.S. enterprises. Based on 2015-2024 multinational operations data from 2,800 S&P 500 firms, complemented by in-depth case studies of six corporations including Nike and Tesla, empirical findings reveal that for every one-standard-deviation increase in the institutional distance friction index, marketing ROI deteriorates by 12.7%. However, when dynamic data processing capability surpasses the 0.73 threshold, 58% of efficiency losses can be reversed. This research pioneers the theoretical subfield of dynamic institutional marketing and develops an interpretable AI budget optimization system. Validated through A/B testing, the system yields an average ROI improvement of 22.3% with a 79% manager adoption rate. Theoretically, it introduces a new dimension of digital institutional distance and operationalizes dynamic capabilities into a three-stage mechanism of “real-time sensing-algorithmic seizing-agile reconfiguration.” Methodologically, it integrates causal machine learning with dynamic panel estimation to resolve the dual challenges of endogeneity and dynamic effects. Practically, it constructs an intelligent decision-making tool that balances predictive accuracy and interpretability, providing a paradigm transformation pathway for marketing strategy in the Globalization 4.0 era.