Never rely only on benchmark data. The case study with Absolute is interesting, though I am not sure it applies to many.