This paper proposes a model to estimate economic losses attributable to front-end complexity and web performance debt in web applications. Technical debt consequences arising from excessive JavaScript execution, deep DOM hierarchies, and high HTTP request volumes were examined. These factors degrade web performance, reduce user engagement and conversion rates, and ultimately diminish revenue. Using an analytical modelling approach, correlations were established between front-end complexity metrics, including JavaScript bundle size, DOM depth, and HTTP request volume, and web performance indicators, specifically Largest Contentful Paint (LCP) and Interaction to Next Paint (INP). Revenue and conversion losses were modeled as nonlinear functions of latency, incorporating margins, control intensity, and threshold effects to represent realistic performance-revenue relationships. The empirical evaluation relied on a three-layered dataset combining real-world performance metrics from Lighthouse with synthetic datasets modeling optimized (best case), typical (baseline), and high-complexity (worst case) front-end scenarios. The results confirmed that increased front-end complexity and performance debt correlate with deteriorated latency and interactivity, leading to substantial conversion and revenue losses. Marginal and threshold analyses revealed nonlinear effects: at lower complexity levels, performance improvements yield higher financial returns, whereas at higher complexity levels, optimisation produces diminishing marginal returns. These findings demonstrate that front-end performance optimisation is an economic imperative rather than a technical consideration. Effective management of front-end complexity reduces performance debt and revenue erosion, providing a framework for engineering decisions and investment strategies in performance-critical environments. This approach transforms performance management into a strategic economic decision, enabling investment optimisation through direct correlation with business outcomes.
A.I. Hryshchenko (Mon,) studied this question.