An Empirical Analysis of Software Testing Practices and Quality Outcomes in Small and Medium Enterprises

Abstract

The software industry has its own challenges of small and Medium Enterprises (SMEs) that are constrained in implementing effective testing practices, are limited in expertise and technology, and have dynamically changing technology environments. IT ecosystem with more than 35 percent of the software SMEs in India is a perfect setting to explore the correlation between testing practices and software quality deliverables. Objective: This research will focus on empirically examining the software testing criteria undertaken by SMEs in India and its effect on quality indicators, which are defect density, customer-reported defects, and time-to-market. Approach: We performed a mixed-method study that included quantitative surveys of 112 SMEs, qualitative case studies of 15 organizations, and the examination of project artifacts of Jira, GitHub, and TestRail platforms that included the period 2022-2025. Random Forest, XGBoost, and Support Vector Machines machine learning models were used to predict the quality outcomes on the basis of testing practice indicators. Findings: We have found out that automation of tests having a coverage of more than 60 percent can decrease the number of defects in production by 47 percent (p < 0.001). Those companies that have implemented CI/CD integrated testing have 34% quicker mean time to repair (MTTR). The predictive model took an accuracy of 87.3 percent in categorizing projects with regard to the quality outcome tiers. Contribution: The study offers practical information on how SMEs can maximize test expenditure, introduces a proven quality prediction model, and offers empirical evidence on the literature of software testing in emerging economies.