CAC (Customer Acquisition Cost) is the total cost of acquiring a new customer — all the sales and marketing spend it took, divided by the number of new customers won in the same period. It's the foundational unit-economics metric: it tells you what it costs to buy growth, and it's the input behind CAC payback and the LTV:CAC ratio.
How to calculate CAC.
Divide total sales and marketing spend by new customers acquired in the same period:
Total Sales & Marketing Spend ÷ New Customers Acquired
SaaS teams track two variations of this calculation:
- Blended CACAll sales & marketing spend ÷ all new customers (paid + organic)
- Paid CACPaid acquisition spend ÷ new customers acquired from paid channels
A thorough CAC includes ad spend, sales and marketing salaries, agency fees, tool costs, and overhead. A simpler working version is channel ad spend divided by new customers acquired from that channel. Whichever definition you use, always count paying customers, not trial signups.
CAC example.
SaaS channel metrics over a 30-day period:
- Blended CAC
- $12,000 spend ÷ 40 new customers
- $300
- Meta · Prospecting
- channel spend ÷ channel customers
- $233
- Google · Brand search
- channel spend ÷ channel customers
- $141
The blended $300 CAC hides real variance: Google brand search acquires at $141, while Meta prospecting sits at $233. Only per-channel CAC tells you where to invest next.
Why CAC matters for attribution.
CAC is only actionable when it's measured per channel and against real paying customers. Blended CAC averages your efficient and wasteful channels into one number that hides the decision that matters. And if you count trial signups instead of paying customers, CAC looks artificially low — until those trials don't convert.
That's an attribution problem at heart: you need to tie each dollar of spend to the Stripe conversion it produced. Signal Sparrow computes true CAC per channel from reconciled revenue attribution, then pairs it with CAC payback and true ROAS — so CAC becomes a budgeting input you can trust, not a blended average you have to caveat.
