Mild functional difficulties have been associated with early cognitive decline in older adults and increased risk for conversion to dementia in mild cognitive impairment, but our understanding of this decline has been limited by a dearth of objective methods. This study evaluated the reliability and validity of a new system to code subtle errors on an established performance-based measure of everyday action and described preliminary findings within the context of a theoretical model of action disruption. Here 45 older adults completed the Naturalistic Action Test (NAT) and neuropsychological measures. NAT performance was coded for overt errors, and subtle action difficulties were scored using a novel coding system. An inter-rater reliability coefficient was calculated. Validity of the coding system was assessed using a repeated-measures ANOVA with NAT task (simple versus complex) and error type (overt versus subtle) as within-group factors. Correlation/regression analyses were conducted among overt NAT errors, subtle NAT errors, and neuropsychological variables. The coding of subtle action errors was reliable and valid, and episodic memory breakdown predicted subtle action disruption. Results suggest that the NAT can be useful in objectively assessing subtle functional decline. Treatments targeting episodic memory may be most effective in addressing early functional impairment in older age.