The problem is that productivity is a function of both time and what useful product was derived. RescueTime is using time to measure both. It has no effective way to tell if anything worthwhile was produced. For example, the statistic about users achieving an average 9% increase in productivity in 3 months actually means that they increased the proportion of time spent in applications labeled productive by 9%, the assumption being that the two are absolutely correlated.
