Paid Acquisition Economics: CAC, Payback and Lifetime Value

CAC: blended versus paid

Customer acquisition cost is the full cost of acquiring customers divided by the number acquired, over the same period. Andreessen Horowitz's 16 Startup Metrics (2015) puts the central trap clearly: CAC should be the full cost of acquiring users, and a common problem is reporting a blended CAC — total acquisition cost divided by all new customers across all channels — that includes customers who arrived organically. Blended CAC is not wrong, but it does not tell you how well your paid campaigns are working; strong word of mouth can make a failing paid channel look efficient.

So compute both. Blended CAC tells you whether the business as a whole acquires customers affordably. Paid CAC — paid-channel spend plus the people and tools that run it, divided by customers attributable to paid — tells you whether the money you can scale is working. 'Full cost' means more than media: include the salaries, agencies and tools that exist to acquire customers, or every channel looks cheaper than it is.

Payback and lifetime value

CAC payback is the number of months of gross margin from a new customer needed to recover what it cost to acquire them: CAC divided by monthly revenue per customer times gross margin percentage. It is the more robust of the two headline ratios for a small company, because it relies on near-term numbers you can observe rather than a forecast of how long customers will stay.

David Skok's SaaS Metrics 2.0, based on his experience investing in subscription businesses, offers widely used guidelines: the best SaaS businesses have a lifetime value to CAC ratio above 3, sometimes much higher, and many recover CAC in 5 to 7 months; his modelling shows profitability becomes anaemic when the time to recover CAC extends beyond 12 months. These are rules of thumb from one investor's portfolio, not laws, and he notes himself that many healthy businesses do not meet them — but they are a reasonable first sanity check.

Lifetime value is the gross margin a customer generates over their relationship with you. Sunil Gupta and Donald Lehmann's work on managing customers as investments, summarized by Wharton in 2007, frames customer profitability as driven by three components — acquisition (rate and cost), margin (dollar margin and its growth) and retention (rate and cost) — with trade-offs between them. A cheap acquisition channel that brings customers who churn in three months can be worse than an expensive one whose customers stay for years; compute LTV per channel where you can, not only overall.

Why measured returns are so uncertain

Platforms report return on ad spend by counting conversions from people who saw or clicked an ad. That is correlation: those people were often going to buy anyway, which is why they were targeted. Randall Lewis and Justin Rao's 2015 paper in the Quarterly Journal of Economics analysed 25 large field experiments with major US retailers and brokerages, covering about $2.8 million in digital ad spend. Even with randomized experiments, measuring returns was hard: the median confidence interval on return on investment was more than 100 percentage points wide, because individual sales are highly volatile relative to the per-person cost of advertising. Informative experiments could require more than 10 million person-weeks.

The eBay paid-search experiments by Blake, Nosko and Tadelis (Lesson 3.1) show the direction of the bias: experimental returns were a fraction of non-experimental estimates, and brand-keyword ads showed no measurable short-term benefit.

For a small company the conclusion is not despair; it is humility and conservatism. Assume platform-reported returns are optimistic. Compute paid CAC from your own records of who became a customer. Where you can, test incrementality with holdouts or geographic switch-offs (Lesson 5.3). And make scaling decisions on payback, which is harder to flatter than return on ad spend.

Making it a habit

Unit economics are only useful if they are computed the same way every month. Write the definitions down once — what counts as acquisition cost, what counts as a paid-attributed customer, which margin you use — and compute the same numbers on the same day each month. Changes in definition are the main source of spurious improvement in these metrics, so log any change as a decision with its date.