The part I stumbled with was in incorporating a variable electricity rate into a model. Cost comparisons are straightforward enough in a flat or tiered rate structure (check your bills for the last 12 months to see what your incremental rate was each month, and guess if your rate would stay the same while charging an EV).
But given the possibility of switching to time-of-use rates, and assessing TOU impact each month on your historical electricity bills, guessing your new incremental rate(s) isn't easy, nor is calculating your updated present electricity cost under a TOU rate structure. If you need to run an air conditioner during the day, your rates become spiky enough that calculating an "average" monthly use will not be enough to accurately model your future costs.
In my case, I installed a 3kW PV system as a hedge to keep my Leaf charging costs reasonable. I wouldn't have done this without the EV purchase. My electricity rate structure, including my household (non-EV-charging) consumption and costs, will change completely as a result.
About this time next year, I expect to have an accurate picture of my charging costs. But I have no easy way to predict it right now to better than about a factor of two accuracy, since I don't have access to my historical month-by-month hourly energy usage and my utility has a steeply tiered TOU rate structure.
The approach I'm going with instead is "don't worry; be happy"