Struggling to figure out how to get ROI on your automation? Wondering which of your tests to automate? Or are you trying to determine what level of automation your organization should target? If so, you may already be heading for a costly mistake.

Each of these thought processes is rooted in a commonplace model as to how to think about test automation. Each attempt to solve a particular problem, to mitigate a particular risk. But, whilst some models are useful, every model is wrong. How useful are these models? What are their drawbacks, and how might they prevent us from seeing how we might harness the power of tools?