Interesting, congrats! And also thanks for sharing the methodology! For the seasonality analysis, how much historical data do you reckon is a good enough?