''We will shift testing left and shrink QA. Developers now own testing for what they ship'' was a statement I once heard from a CTO announcing an organizational change in how QA teams would be structured. It was clearly a top-down decision: nobody in the room (developers, QAEs, SDMs, TPMs, PMs...) had any clear idea what it really meant or what each role should do to make it work. This assumption-that developerowned testing will work out of the box and even raise the quality bar-is the phenomenon this paper examines. Why does dev-owned testing look so compelling in theory, yet fall short in almost every real-world implementation? The problem is not that dev-owned testing is a flawed idea, but that it is usually poorly planned and poorly approached. This paper first outlines signals from industry that show how and where devowned testing is failing in practice, then analyzes the underlying root causes and the points at which existing theory breaks. It concludes with several recommendations to the software engineering research and academic community on how they can better support industry in making dev-owned testing succeed as a socio-technical system
Ahmed El-Deeb (Tue,) studied this question.