COMPARATIVE ANALYSIS OF AUTOMATED TOOLS FOR CLIENT-SIDE PERFORMANCE TESTING IN MODERN WEB ENVIRONMENTS

Abstract

Relevance. In the era of Single Page Applications (SPA), the traditional approach to performance testing, focusing solely on server response times, is no longer sufficient. The logic of modern web applications has shifted to the client side. However, standard automated tools often produce misleading results under unstable network conditions, failing to capture the true user experience. Object of research: The process of testing and analyzing the performance of the clientside of web applications. Subject of research: Methods and software tools for automated analysis of web performance metrics. Purpose of the article: To conduct a comparative experimental analysis of automated testing tools (Google Lighthouse vs. Sitespeed.io) under constrained network conditions and on real-world e-commerce applications, identifying specific reliability gaps in standard auditing approaches. Research results: Experiments on a controlled slow application revealed that Google Lighthouse timed out, reporting a critical "False Positive" (Score 98/100) despite a 50-second load time. Conversely, Sitespeed.io correctly captured the full load duration with high statistical stability (RSD < 10%). Further tests on real-world platforms (Nike, Zara) demonstrated "False Negative" behavior in Lighthouse, which reported inflated LCP values significantly higher than manual observation. Conclusions: The analysis confirms that while Lighthouse is useful for general audits, a robust CI/CD framework requires the flexibility of Sitespeed.io to handle custom pageCompleteChecks, OS-level controlled network throttling, and complex visual elements without generating misleading pass/fail results.