How to Calculate CAC for SaaS

Almost every published CAC number is uncomparable to yours. Not wrong, exactly. Just built on a different cost base, with nobody stating what that base was.

That’s the real problem with customer acquisition cost. The formula fits on a napkin. The decisions underneath it are where two companies with identical efficiency end up reporting 900 dollars and 2,400 dollars for the same motion.

The three CAC variants and when each one is the right number

There are three calculations in common use, and mixing them up is how board conversations go sideways. Each answers a different question.

Blended CAC takes all sales and marketing spend and divides by all new customers, however they arrived. It answers: what does a customer cost us on average right now.

Paid CAC takes only advertising spend and divides by customers attributable to paid channels. It answers: should we spend more on ads.

Fully loaded new customer CAC takes every acquisition related cost, including people, and divides by genuinely new logos. It answers: what does it really cost this business to add a customer.

My position, and it’s not a mild one: report fully loaded new customer CAC as the headline number, everywhere. Use blended as a directional trend only, and say out loud that it moves whenever channel mix shifts even if nothing about your efficiency changed. Paid CAC belongs in the channel review, not the board deck.

What belongs in the numerator, line by line

This is the table to argue over with finance once, write down, and then never relitigate. The principle is simple: if the spend is aimed at people who are not yet customers, it counts.

Two lines cause real arguments. The first is customer success. Including CS inflates CAC by roughly 15 to 30 percent at most companies, and it hides the thing you actually want to know, which is whether retention spend is producing net revenue retention. Keep them separate.

The second is brand. A sponsorship with no lead capture still does acquisition work you cannot measure. I include it and footnote it, because excluding costs because they’re hard to attribute is how CAC becomes a vanity metric. The blended CAC definition covers where these choices bite hardest.

How to lag spend against a 60 to 120 day sales cycle

Dividing this month’s spend by this month’s closed customers is wrong for anyone with a sales cycle longer than a few weeks, and it’s wrong in a specific, predictable direction: it punishes you for growing.

Here’s why. Say your median sales cycle is 90 days. You double spend in April. Those deals close in July. If you compute April CAC as April spend over April closes, April looks catastrophic and July looks miraculous, and neither number describes reality.

If you sell to two segments with very different cycles, run two calculations. A product doing self serve at 14 days and enterprise at 160 days has two businesses in one P&L, and a single blended CAC describes neither of them.

Worked example one: 3,000 dollar ACV, self serve with light sales

A small team product. Quarterly sales and marketing spend of 210,000 dollars, made up of 120,000 in salaries for four people, 55,000 in paid acquisition, 20,000 in tooling and 15,000 in contractor content. Sales cycle is 21 days, so the lag is effectively zero.

New logos in the quarter: 240. Plan upgrades from free: 310, excluded.

Fully loaded new customer CAC is 210,000 divided by 240, so 875 dollars. Against a 3,000 dollar ACV at 80 percent gross margin, monthly gross profit per customer is 200 dollars, giving a payback of about 4.4 months.

That’s a healthy number for this band. If the same business reported blended CAC including the 310 upgrades, it would show 382 dollars, and someone would use that figure to justify a spending increase the unit economics cannot support.

Worked example two: 25,000 dollar ACV, sales assisted

Mid market. Annual sales and marketing spend of 3.2 million: 1.9 million in salaries and commission across marketing, two SDRs and four AEs, 640,000 in paid and events, 310,000 in tooling, 350,000 in agency and contractor work.

Sales cycle is 75 days, so spend lags closes by roughly one quarter. New logos over the lagged window: 142.

CAC is 3.2 million over 142, so 22,535 dollars. Against 25,000 ACV at 78 percent gross margin, monthly gross profit is 1,625 dollars and payback is about 13.9 months.

That sits inside the range Benchmarkit reports as a median, which is near 16 months across B2B SaaS. It’s fine, not great. The lever here is usually win rate rather than spend, because at this band a two point win rate improvement moves CAC more than any budget cut you’d survive.

Worked example three: 120,000 dollar ACV, enterprise

Enterprise motion. Annual spend of 6.8 million, dominated by 4.4 million in people costs across an enterprise AE team, solutions engineers who support new business, field marketing and an ABM program. Events run 900,000. Paid media is small at 400,000 because the addressable list is a few thousand accounts. Tooling, intent data and agency work make up the remaining 1.1 million.

Sales cycle is 190 days, so the lag is two quarters. New logos: 178.

CAC is 38,202 dollars. Against 120,000 ACV at 74 percent gross margin, monthly gross profit is 7,400 dollars and payback lands near 5.2 months on a gross profit basis, which looks extraordinary until you remember that enterprise logos churn in chunks and one loss removes a year of acquisition work.

That’s the point of running all three examples. A 38,000 dollar CAC is healthy at 120,000 ACV and instantly fatal at 3,000. Any benchmark that does not state the ACV band it came from cannot tell you anything about your business. The same problem runs through most published SaaS metrics and analytics content, and it’s why segmented figures are worth far more than headline ones.

What to do with the number once you have it

CAC is an input, not a verdict. Three things make it useful.

Pair it with payback. The B2B SaaS CAC payback calculator and the CAC payback period calculator both do the arithmetic on a gross profit basis, which is the only correct basis. Payback on revenue overstates your position by whatever your cost of service happens to be.

Pair it with lifetime value. The LTV to CAC ratio calculator gives you the ratio, and how to calculate LTV for SaaS explains why most LTV figures are inflated by optimistic churn assumptions applied over a ten year horizon nobody has observed.

Then break it down by channel, because a single company CAC hides the fact that one channel is producing customers at 400 dollars and another at 9,000. That’s where cost per lead becomes the working metric, and how to lower B2B SaaS cost per lead covers the levers that actually move it.

One honest limitation before you present any of this. CAC assumes you can attribute customers to spend, and for anything involving word of mouth, communities or a founder’s reputation, you can’t. Those customers land in the denominator with no matching cost, which makes CAC look better than the truth for companies with strong brands and worse for companies still building one. Say that in the footnote. It’s more credible than pretending the model is complete, and it’s the difference between a metric the board trusts and one they quietly stop believing.