Unbounded Retention

Unbounded retention is a retention definition that counts a user as retained on day N if they were active on day N or on any later day.

The definition looks forward from day N to the end of the observation window, so the resulting curve is always flat or falling and much smoother than an N-day curve. It suits products with an irregular rhythm, where a user genuinely comes back but not on a specific date. It carries two traps. First, it is the most generous of the retention definitions, so its value must never be placed next to an N-day figure. Second, recent cohorts have less future time in which to be observed, so the number sags artificially near the edge of the window, and cohorts with unequal observation windows cannot be compared at all.