You know what a customer costs. How long until they pay you back?

Cohort payback is the number that quietly runs a subscription or repeat-purchase business: it folds acquisition cost, retention, and margin into one figure, and it decides how fast each dollar can be recycled into the next cohort. This models the curve, and shows what one point of retention is actually worth.

Acquisition cost per customer ($) Fully loaded CAC, media plus the incentives it took to convert.

Cycle length Your billing interval, or the average gap between orders.

Order or subscription value per cycle ($)

Contribution margin (%) After product cost, shipping, fulfillment, and payment fees.

Retention shape

Real curves drop steeply and then flatten, so a single churn rate models the wrong thing. These two numbers give you the shape.

Customers who survive the first cycle (%) The activation cliff. This is where onboarding either worked or did not. 78% Ongoing retention per cycle after that (%) The flat part of the curve, cycle over cycle among customers who stuck. 93% Test a retention improvement (percentage points) Applied to the ongoing rate. Small numbers here do more than they look like they should. +2.0 pts

Cohort payback period

Contribution per cycle-

Lifetime contribution-

LTV to CAC-

Cumulative contribution With the retention lift Acquisition cost

Scenario	Payback	Lifetime contribution	LTV:CAC
Today	-	-	-
With retention lift	-	-	-

The model treats the cohort as an activation cliff followed by a steady ongoing rate, which is the shape real retention curves take and the one a single blended churn rate cannot represent. Contribution is counted per surviving customer per cycle and is not discounted, so treat long paybacks as optimistic. It is a directional model: the real version gets built from your own cohort table, by acquisition source, because retention that degrades cohort over cohort is usually a media problem wearing a retention costume.

More on the math behind this: subscription unit economics · churn reduction · Churn Decomposition Model