average treatment effect is the average of individual treatment effect across all units
It is often interpreted as the typical effect on a typical unit, even though this estimand may never be actually be experienced by any unit, even when estimated correctly.
Naive average treatment effect
This average treatment effect is inherently unknowable, because it requires both potential outcomes for each unit, while we only know one of these (observable outcome) by the switching equation.
But it can be estimated, using the simple average treatment effect, which is the difference between average response among subjects assigned to treatment AND average response among subjects assigned to control.
Depending on the research question, other parameters might be interesting.