Compare it to the likely alternative for the task/person, probably R or even MS excel in many cases i'd guess. The alternatives should ideally be based on empirical observation of the population. The marginal saving of choosing a higher efficiency than python might look a lot lower.