Comparative Evaluation of Selenium Cypress and Tricentis Tosca for Enterprise Test Automation

Abstract

Purpose: Enterprises choosing among Selenium WebDriver, Cypress and Tricentis Tosca routinely rely on feature checklists and short demonstrations that ignore portfolio coverage, maintenance under change, workforce skills, governance and exit cost. This article develops a transparent decision model for that choice. Method: A critical integrative review of software testing theory, empirical software engineering methodology and technical documentation for the three tools informs the Enterprise Test Automation Selection (ETAS) model. ETAS applies non-compensatory knock-out gates, then weights six criteria through pairwise comparison with consistency checks, and aggregates normalized ratings. Three illustrative enterprise archetypes demonstrate the mechanics, followed by one-at-a-time crossover analysis and Dirichlet rank-acceptability simulation. Findings: The tools differ in architecture rather than quality. Selenium offers protocol-level reach across browsers and languages; Cypress offers fast, retry-based feedback for JavaScript teams inside the browser; Tosca offers model-based, codeless coverage of SAP and packaged applications under commercial licensing. In the illustrative cases, rankings were decisive for the JavaScript and SAP archetypes but fragile for a polyglot regulated bank, where the leading tool changed once maintainability or governance weight passed identifiable thresholds. Contribution: The article supplies a gated selection model, an explicit rating rubric, six falsifiable propositions and a staged validation design. Implications: Selection should be documented as a weighted, sensitivity-tested decision that is re-examined as the application portfolio changes.