Building a growth model that survives board questions

Growth models fail when they are ornamental. A useful model connects spend to marginal outcomes, states assumptions plainly, and predicts what should be observed weekly.

Start with unit economics, validate tracking, then layer channel curves—not the other way around.