Why saturation curve is not cumulative ?

Hi Team,

I wanted to ask some questions on the saturation curve of robyn :
If you change the date period for the saturation curve in budget optimizer then the initial reference point changes and because of that the mROAS value changes.
So the mROAS becomes dependent on the base period that has been chosen budget optimizer, Is this a right approach ?(Wanted to know the reasoning behind it)
2. Why we are plotting this : the mean channel spend per period (per day, if you're modeling daily, per week if modeling weekly, etc.).
Shouldn't it be cummulative, Like Meridian :