Initialize the Time Series of a Spread function?

I have a variable, Win%, which shows what percent of a cohort’s Demo wins will close in the months following the month of the Demo. So my wins might occur 0%,10%, 25%, 50%, 15% in the five months following the Demo. Spread(Demos; Win%) gives me the wins for all the cohorts of Demos (I’m not using a system “cohort” just the jargon). This works very well except for the first part of the curve if you are starting in a month that had Demos given in prior months, which don’t get counted by the spread function. How do you initialize the spread function in the first few months so it gives reasonable answers?
My picture to show how the leftmost part of the spread fell to zero did not upload properly here: