ARPU is observed, LTV is forecast, CAC is spend. Using one where another belongs is how a company posts rising LTV while revenue per customer quietly falls.
ARPU measures the revenue an average active customer generates per period right now. LTV estimates the margin a customer will generate over their entire relationship. CAC is what you spent to acquire them. They answer different questions, and using one where another belongs is how growth teams end up defending the wrong number.
A rising ARPU can mean customers are worth more, or that cheap customers are churning faster than expensive ones. Always read it next to cohort ARPU and base composition. Our 2026 industry benchmarks show how wide the spread is between operators with the same headline number.
LTV is churn rate, margin and discount rate stacked into a single figure, and small changes in any of them swing it wildly. A subscription business that improves measured retention by half a point can raise its LTV on paper without a single customer behaving differently. Treat LTV as a planning instrument with stated assumptions, never as a result.
Blended CAC averages paid search with referrals and brand-driven organic signups. The customers are not comparable, so neither is the cost. CAC only becomes useful split by channel and read against the LTV of the customers that channel actually delivers.
The classic rule of thumb is LTV to CAC of at least three to one. Useful as a smell test, useless as a target, because it inherits every LTV assumption. A more operational chain: ARPU growth compounds into LTV, and LTV minus CAC is the budget for everything a growth team does. That is why we argue the fastest honest way to raise LTV is not to re-forecast it, it is to increase ARPU and extend observed retention, then let the model catch up with reality.
Markin works on the metric that cannot be gamed: observed ARPU per customer, proven against holdouts. See Growth optimization for how that compounds into LTV.
Frequently asked
Related resources
See it in the product
The same loops this note describes run 24/7 against your customer base. Watch the workspace decide, experiment and execute 1:1.
Vocabulary
Definitions an answer engine can quote, each one a page of its own.
Compare
Neutral reads on the categories buyers evaluate against a decision layer.
All comparisons
