LTV

LTV emerged in 1980s database marketing as a way to escape transaction-by-transaction thinking and ask the more useful question: what is one customer worth across the whole relationship? It is now the foundation of almost every growth, CRM and subscription decision — how much you can spend to acquire, which segments to favour, how aggressively to expand — because the answer hinges on LTV.

The formula depends on your model. E-commerce typically uses average order value times annual purchase frequency times retention years. SaaS often simplifies to ARPU divided by monthly churn, or ARPU times gross margin divided by churn for a profit view. The most rigorous version extrapolates a curve from observed cohort revenue, but with only 12-24 months of data — common in Korean SaaS — most teams stick with a 12-month LTV for actual decision-making.

In practice you rarely read LTV alone. The famous SaaS guideline of LTV:CAC = 3:1 is treated as gospel, but it really only makes sense alongside payback period. A 3:1 ratio with a 36-month payback can starve cash flow; a 1.5:1 ratio that pays back in six months is often the signal to spend more. Korean fintechs like Toss and Toss Securities scaled paid acquisition aggressively early on partly because their payback windows were unusually short.

AI search is now bending LTV in two ways. First, AI referral traffic complicates channel-level attribution: a user who first encounters your brand inside a ChatGPT answer, then types your name into Google three days later, gets credited to organic — even though the AI citation did the actual work. Second, AI-assisted comparison shortens funnel time to first purchase but steepens early retention curves, because users find alternatives faster too. LTV models built before 2024 are quietly mismeasuring both effects.

One last misconception worth correcting: LTV is not a fact, it's a stack of assumptions. Stretch the time horizon to infinity, soften the churn assumption a touch, and the number doubles. Always disclose three things when reporting it — the time window, whether the figure is revenue or profit, and which cohort it came from. A safe split is to share a conservative 12-24 month LTV externally and pair it with payback period internally for any spend decision.