Customer acquisition cost = acquisition spend ÷ new customers acquired by that spend. There are three fair versions of it, and one lie:
- Paid CAC = performance-marketing spend ÷ customers acquired through paid channels. It's the marginal price of manufactured demand.
- Blended CAC = all acquisition spend ÷ all new customers, organic included. It's the average price, flattered by every customer who'd have turned up for free.
- Fully-loaded CAC = the blended (or paid) numerator, widened to everything acquisition really costs: sales salaries and commissions (SBC included, as last week's rule says), agency fees, onboarding/installation, first-order discounts, free-trial serving costs. For a SaaS company with a sales force, payroll usually dwarfs the ad budget, so a "CAC" that counts only ads is a fiction.
- The lie is presenting blended while scaling paid. Watch it happen:
GlowKart, year 1. Performance ads costing ₹6.0 cr acquire 75,000 customers, a paid CAC of ₹800. Referral bonuses of ₹0.5 cr acquire 20,000 more, at ₹250 each. Organic/SEO/word-of-mouth brings 55,000 at roughly zero marginal cost. Blended CAC = ₹6.5 cr ÷ 150,000 = ₹433, and the deck says "we acquire customers at ₹433".
Year 2. The company wants 50% more new customers, 225,000 of them. Organic doesn't scale on command, so it stays around 55,000, and referrals stretch to 25,000. Paid has to deliver 145,000. Auctions price the marginal impression, so paid CAC rises to ₹1,100. That's diminishing returns at work, since the cheapest customers were acquired first. Blended CAC = (145,000 × 1,100 + 25,000 × 250) ÷ 225,000 = (₹15.95 cr + ₹0.63 cr) ÷ 225,000 = ₹737.
Advertisement
What does the GlowKart example show about blended CAC?
Look at what happened. Blended CAC rose 70%, and marketing didn't get any worse at its job. Instead, the mix shifted, with organic falling from 37% of adds to 24%, and the marginal paid customer costs more than the average one.
There are three lessons here, and they're permanent. (1) Blended CAC is a mix statement. It's useful for cash maths and useless for judging acquisition efficiency. (2) The marginal customer prices your growth, so always ask what the last tranche of spend bought. (3) Any company that guides to faster growth while projecting flat blended CAC is assuming organic scales linearly with paid. It's history's most expensive unspoken assumption.
When you can get the data, break CAC down by channel and track each channel's CAC and its share. Most filings won't give you that. Then at least separate growth in "advertising & sales promotion" from growth in new customers, and watch the ratio.
How can the CAC denominator be gamed?
One more rule, and you'll meet it again among the games at the end. CAC's denominator is as gameable as its numerator. Count "new users" (downloads, registrations) in place of new transacting customers and CAC can shrink 3–5× overnight. Pin the denominator to the unit you chose at the start, the one revenue actually attaches to.