Every retention curve falls. The only question that matters is whether it flattens, and how high.
New members at a gym. Most stop coming in weeks; the question is whether the line ever goes flat, and how many people are on it when it does.
The height of the plateau decides whether acquisition compounds or whether you're renting users forever.
Plot the share of a cohort still active against days since signup and you get a curve that drops steeply, then bends. The steep part is people who were never going to stay. The bend is the product finding its actual users, and the height of the plateau is the fraction for whom the product works. That plateau is the single most predictive number in a consumer business, because growth compounds only when the curve flattens above zero — a curve that decays to nothing means every new user must be replaced forever, and acquisition becomes a treadmill rather than an investment.
A retention curve is the share of a cohort still active by age. It always falls; what matters is whether it flattens and at what level. A flat tail means a real core of users, so acquisition compounds. A curve decaying to zero means you are renting users. Improving the plateau beats improving day-1, because the plateau multiplies every future cohort.
#Tableau - Calculate Customer Retention & Cohort Analysis — Andy Kriebel, 5:20