What 5 million job advertisements tell us about testing

Abstract

Software testing is a crucial part of business success to ensure final product quality. However, little concrete data exists on technical demands about it in the industry, mostly collected through personal opinion surveys on a restricted sample of professionals. In this paper, we used a different approach: we applied content analysis to a set of about five million job advertisements taken from a popular Web job-search engine. The analysis of job advertisements is more promising than surveys because the data are by far more numerous and distributed geographically. The content analysis results revealed four essential findings on the current practice of software testing: a) Companies search for about six times more Coders than Testers, b) Unit testing is the most required skill for Coders while Acceptance testing is the most popular for Testers, c) Automated testing dominates the job advertisement scene compared to Manual testing and, d) the most valuable testing tools and frameworks are Selenium, JUnit, and Cucumber for both Testers and Coders. We believe that these findings (and other related results from the content analysis study) will be useful for professionals, instructors, and researchers dealing with software testing.