Udacity course in predictive analytics says that for seasonality you need to have a year's worth of data plus the cycle plus the test period. So if you have a 12 week cycle and a 12 week test, then you need 76 weeks.
12 + 52 + 12 = 76
The AB Tool documentation isn't very clear, but my course notes are very specific about the needing 1 year plus the trend for history.
Thanks for the reply @treepruner ... but I don't think that explains this behaviour... there is enough data to calculate the seasonality when I add one more store... remember the seasonality is calculating per store so the amount of rows of data per store is not changing. but when I only have the data for the one store it no longer calculates seasonality... This is where it strikes me as odd.
Do Alteryx people pop into the forums to answer questions?