Ask people what goes wrong in a charging pro forma and they'll talk about hardware prices, installation overruns, utility upgrades. Those are real. They're also the lines you can refine: get quotes, get a second quote, get the utility to put something in writing.
The line almost everyone gets wrong is utilization.
The assumption
Everyone assumes their site will get three, five, ten cars a day. A forecast from Stable, or whoever did the study, says ten, fifteen or twenty sessions a day, and that number goes straight into the model starting in month one.
A forecast like that describes what a site might do once it's established. It is not a day-one number. On day one you might get one charge a week.
A charging site is real estate
A charging site behaves like any other piece of real estate with a business on it. A coffee shop, a barber shop, a grocery store, a gas station. Nobody opens a coffee shop and expects the line out the door on the first morning.
People have to know it exists. They have to be marketed to. They have to try it, and then trust it enough to plan around it.
The ramp looks like this: at first almost nobody uses it. Then people start noticing it. A few get curious. The site proves it works, consistently. Drivers start routing to it. That consistency is what brings the volume, and it takes time.
One of mine
One of my sites took almost a year to go from one session a week to about one a day. Nothing about the hardware changed. EV adoption in the area grew, and people figured out that the site was there, that it worked, and that they could rely on it.
That got it to one or two sessions a day, and three to four on a weekend. Today it does two to three a day. The whole ramp took two to two and a half years.
Adoption was the main thing. A forecast that shows two to three a day is describing year three, not month one.
Why this line is the one that's wrong
Capex and opex are anchored to quotes and bills. Utilization is anchored to nothing but the forecast and your own optimism.
Frankly, it's the lever you can turn to make the model say what you want. Need the payback to hit five years? Add two sessions a day. Nobody can prove you wrong until the site is open, and by then the money is spent.
For context, Paren's Q2 2026 report put the national average for fast chargers at 15.76% of the time in use. That average includes sites that have had years to ramp.
What to do about it
Model the first year as a ramp, not a flat line. Ask what the site does if it takes twice as long to get where the forecast says it ends up. If the project only works when the forecast shows up on opening day, it doesn't work.
The utilization forecast shows every assumption it makes, so you can argue with it. The site planner lets you see what your economics look like at the low end, before a customer does it for you.
