Ask HN: How do you forecast revenue for your SaaS?
I was looking at some data today and noticed the following: after we onboarded a user, time-to-conversion roughly follows an exponential decay curve. Most of our signups convert within the first few weeks of onboarding. A handful convert much later, but the probability of conversion drops off sharply with age. Makes sense. I also assume this is pretty typical (yeah?).
So I did what tech-cofounders do...busted out numpy and scipy and did some curve-fitting ;) Assuming the probability of conversion follows an exponential distribution in time, the expected value of total future conversions should be a sum of exponential decay curves, one for each user, with each curve offset by age. The results of this match intuition so far. I'm curious if anyone uses a similar approach. (Or a dissimilar approach...that would be even more interesting.) Thanks in advance!