Methodology for an operationally-based test length decision