Customer Acquisition Cost Formula for SaaS: Calculate and Lower CAC

Most SaaS teams can quote their CAC. Very few can defend it. The number in the board deck is usually a division of ad spend by new signups, with salaries excluded, tools forgotten, and the trial-to-paid conversion happening in a different month than the spend that drove it. It is directionally useful and practically fiction.

This guide is the version we use with SaaS clients, built from years of running paid acquisition for software companies and from the operator side of the fence: years inside the Square software ecosystem, helping business owners integrate and adopt a platform, watching what actually happens to acquisition economics when onboarding and retention are treated as part of the acquisition decision. A 94.1% client retention rate is not a marketing slogan; it is the end state of treating acquisition cost and retention as one number. That lens runs through every section below.

The structure is deliberate: the plain formula first, then the two corrections that change decisions, then the levers. If you only read two sections, read the gross-margin payback correction and the channel-level breakdown. They are the two places where the honest number diverges from the board-deck number.

The customer acquisition cost formula is a ratio between two totals measured over the same period:

Both sides of that division carry more weight than they look. The numerator is everything spent to acquire customers, not just the ad platform invoice. The denominator is new customers, not new leads, not new trials, not new logins. The two most common ways teams get this wrong are undercounting the numerator (forgetting salaries and tools) and overcounting the denominator (counting trials as customers).

Choose one measurement window and stay consistent. Monthly CAC suits ongoing channel management. Quarterly smooths lumpier B2B sales cycles. Cohort-based CAC, where you track the spend that won a specific group of customers against the revenue that group generates, is the most accurate but needs a spreadsheet and discipline. Most teams should run monthly for operations and quarterly for strategy, and graduate to cohorts when retention data accumulates.

The honest numerator includes everything your business spent to win new customers in the period:

The pattern to notice: every exclusion makes CAC look better and every inclusion makes decisions better. Teams that exclude content costs conclude paid is more efficient than organic. Teams that include them often find the opposite. The number is only as honest as its numerator.

The denominator is new paying customers acquired in the period. Not signups. Not trials. Not reactivated accounts. A paying customer who came through a trial that started last month counts in the month they paid, and the spend that drove them counts in the month it was spent. When those months differ, period-based CAC distorts; cohort-based CAC does not. That is the entire argument for cohorts once your volume justifies the spreadsheet.

Two consistency rules matter more than any single number: define "customer" once (a paying account, a paying seat, or a paying contract, chosen once and never changed quietly) and match the cost window to the conversion window. A SaaS with a 45-day trial-to-paid cycle should expect March spend to show up in April and May customers. If that lag is not in your model, every month looks either brilliant or broken depending on which side of the lag you are on.

An illustrative example with transparent assumptions, not a claimed result or a market average. Say a small B2B SaaS spends, in one month:

Fully loaded acquisition cost for the month: A$30,000. The team won 50 new paying customers.

Now the payback question. If average revenue per customer is A$100 per month and gross margin is 80%, the customer contributes A$80 of gross profit per month:

Same spend, same customers, and the payback story is completely different depending on whether you divide by revenue (six months, optimistic, wrong) or gross profit (7.5 months, honest). State the margin. Without it, any payback figure you publish or present is not a number, it is a mood.

Blended CAC is a summary, not a signal. It averages your efficient channels with your expensive ones and reports one comfortable number. The channel that drags it up stays hidden. The channel that looks cheapest stays funded even when its customers churn in two months.

Channel-level CAC assigns every cost line to the channel it served: ads by platform, content to organic and SEO, events to field, outbound tools and salaries to outbound. Each channel then gets its own CAC, and each CAC gets its own retention check. That last step is the one most teams skip, and it is where the real decisions live. The pattern we see repeatedly in acquisition work: the cheapest-signup channel often produces the worst-tenure customers, because it catches people who had the least intent. Paid comparison keywords bring switchers who already knew the category; broad social catches browsers. The CAC on the second group is lower right up until month three.

Cohort CAC closes the loop fully: track the specific spend that won a specific group of customers, then follow that group's revenue over its actual lifetime. It is the only version of CAC that cannot flatter you, because it eventually reveals both what each customer cost and what each customer returned. Start with monthly blended plus channel-level, and add cohorts once retention history exists.

These three numbers measure different things, and mixing them is the most common spreadsheet error in SaaS marketing:

A trial is not a customer. A lead is not a trial. The day a board deck shows CPA in the CAC row, every downstream decision, from channel budget to headcount, inherits the error.

Payback period should be calculated on gross profit, not revenue. The formula:

The reason is simple: revenue is not what funds the acquisition. Gross profit is. A customer paying A$100 a month on an 80% margin contributes A$80 toward repaying the A$600 it cost to win them. The A$20 that is not yours never repays anything. A team dividing by revenue reports a 6-month payback; the honest gross-profit number is 7.5 months. At 70% margin the gap widens further, and add sales-cycle time and the real number lands well past the board-deck figure. Revenue-based payback numbers are not conservative estimates, they are systematically wrong in one direction: optimistic.

This correction changes real decisions. Channels that looked break-even on revenue-based payback are often materially underwater on margin. Products with lower margins but higher volume can look worse than they are if margin is ignored but also get funded longer than they should if margin is ignored in the other direction. One number, honestly calculated, beats two numbers optimistically calculated.

Lifetime value is what a customer returns over their whole relationship, and the standard SaaS approximation is:

And lifetime, for a subscription business, is the inverse of churn: at 5% monthly churn the average customer stays roughly 20 months; at 2% churn roughly 50 months. This is why retention is not a separate topic from acquisition: churn sits inside the LTV that justifies the CAC. Every point of churn you remove stretches lifetime, lifts LTV, and makes the same acquisition spend cheaper in the only currency that matters.

The LTV:CAC ratio compares what a customer costs against what a customer returns. The conventional healthy band is 3:1 or higher, meaning a customer returns at least three times their acquisition cost. Below 3:1 and growth spends money faster than customers repay it; far above it, maybe 6:1 and beyond, usually means underinvestment in acquisition rather than genius efficiency. Treat the band as directional: a self-serve product at 3:1 and a sales-led enterprise product at 3:1 are completely different businesses.

SaaS benchmark numbers travel fast and lose their context even faster. The ones below are the closest thing to consensus across SaaS finance sources, and every one of them is directional: they vary by go-to-market model, stage, ACV and market. Use them to sanity-check your own trend, never as targets to hit from outside.

If you see a benchmark without a model attached (PLG vs sales-led), a margin assumption, or a date, treat it as a rumour. The only CAC benchmarks that matter are your own last six months, trending.

There are only four honest ways to lower CAC, and they are not equal:

The order matters because most teams reach for lever four first: cut bids, trim budgets, shop for cheaper clicks. It is the weakest lever and the most exhausting. Retention and conversion are where the money is, and they compound. If your CAC is genuinely high, the diagnosis is usually upstream of the ad account.

The most expensive acquisition mistake in software is not overpaying for clicks. It is buying the same customer twice. Every churned customer re-enters the acquisition line, and the CAC you report never shows the repurchase: it shows a fresh new customer with a fresh new acquisition cost, and the graph keeps telling you the funnel works.

This is the lesson years inside the Square software ecosystem drilled in. Helping business owners integrate the platform, train their teams, connect payments and actually use the system is retention work, and it is the same work that determines what a merchant is worth. The merchants who adopted properly stayed; the ones who churned had to be won again by whoever sold to them next. Adoption is acquisition insurance. Our own 94.1% client retention rate exists because we treat the retention side of that equation as part of the acquisition offer: systems people stick with do not need to be re-bought.

For your CAC worksheet, the practical version is this: before you optimise bids, ask what share of last quarter's new customers are still active. If the answer is uncomfortable, marketing is filling a leaking bucket, and the fix-activation-first rule from our founder playbook applies. A percentage point of churn is worth more than a percentage point of CPC.

A repeatable monthly review, small enough to actually happen:

Fifteen minutes a month with consistent rules beats a quarterly deep-dive that reinvents the definitions each time.

Low CAC is not the goal. Profitable growth is the goal. A sales-led enterprise product with a A$40,000 ACV can carry a CAC of A$8,000 and be extremely healthy, because payback lands inside a single contract year and expansion revenue compounds behind it. A self-serve product at A$29 a month cannot carry that same CAC and survive a quarter. The same number is a disaster in one business and a bargain in the other.

So before you try to lower CAC, check you are optimising the right number for your model. If your ACV is high and your payback is inside your justified range, do not celebrate a low blended CAC that came from cheap, churning traffic. And if your ACV is low, do not accept a high CAC on the promise that "brand will fix it later." The formula does not replace judgement. It funds it. For the full channel strategy that sits around these numbers, our complete SaaS marketing guide covers the rest of the unit economics, and the B2B SaaS SEO strategy playbook shows how organic compounds the same maths over time.