From Data to Strategy: A Public Market Framework for Competitive Intelligence

This is the authors' abstract. We don't add key points for this paper.

This paper introduces a dynamic, mathematically grounded, and fully data-driven framework for competitor analysis, index construction, and industry benchmarking—addressing fundamental limitations in traditional strategic approaches that rely on static classifications, internal disclosures, or proprietary data. Built entirely on publicly available financial signals—including firm-level equity prices, macroeconomic indicators, and sector ETFs—the framework replaces rigid taxonomies with a flexible, behaviorally revealed peer selection mechanism governed by coupled ordinary differential equations (ODEs) and speed-weighted Runge-Kutta integration. A stakeholder-aware thresholding scheme integrates strategic priorities without sacrificing analytical rigor. The resulting industry index is constructed using Self-Financing Hierarchical Risk Parity (SF-HRP), producing stable and interpretable portfolio weights. A Gram-Schmidt orthogonality ensures parsimony among macroeconomic factors (|ρ| < 0.05), enhancing robustness. Forecasting evaluations confirm that CNN-LSTM achieves superior performance (RMSE: 0.028, R 2 : 0.95), reducing error by over 80 % relative to baseline LSTM models, while WGANs capture tail risk distributions (R 2 : 0.75). Critically, the framework demonstrates that high correlation does not imply redundancy—validated through multiscale residual analysis, information-theoretic modeling, and geometric submanifold representations. At its core, this research introduces the signal-based strategic interaction model, a new theoretical paradigm that reframes financial performance as real-time strategic communication. It establishes a generalized, industry-agnostic system for dynamic competition management—scalable, transparent, and responsive to market shifts—empowering firms of all sizes to identify peers, track competitive positioning, and inform strategy with unprecedented precision and accessibility.