Forecasting S-Curves
Forecasting s-curves (sigmoid, logistic, etc.) is hard.
This article has an easy explanation of why, and a conclusion that often predicting the specific s-curve given data is hard even up until one is already at the “flattening” portion of the curve.