No problem, I can create that index 
And what kind of pivot would you suggest?
A non-perfect but working solution is to use a Lens with 2 layers (one unique count of emails from the original index, and one with count of my emails that are churn).
Doing a bar vertical percentage will do the computation from me (so the visualization isn't great but I have the result in %). If this is doable via a visualization, I assume I should make it another way 
It's not perfect but enough for me (except if someone got a better idea).
How about a similar concept to detect… new customers? Since the "latest" transform will find me the latest record (which is fine for leaving customers in a month), what would be the best approach to detect the fresh new customers per month?