Your customer acquisition cost is the total you spend to win customers over a period, divided by the number of customers you win in that same period. The formula is one line, and that is exactly the problem. Two companies can report the identical CAC while counting completely different things, and the gap comes entirely from what each one agrees, or refuses, to put in the numerator.
The stakes are not academic. According to Benchmarkit (2025), an independent industry survey, B2B SaaS companies spend a median of 37% of revenue on sales and marketing. A line item that size is steered with a written calculation rule, not a back-of-envelope estimate at quarter close. And the number most founders reach for, ad spend divided by customers, is not that rule. This article gives you the formula, then shows what a defensible, fully-loaded CAC actually contains and the three versions of it you need to keep apart.
What is the customer acquisition cost formula?
The formula reads: CAC equals total acquisition spend over a period, divided by the number of new customers won in that period. HubSpot works a small B2B example that already breaks the ad-spend habit: $2,500 in LinkedIn ads, $1,000 to a contractor, $1,500 for the sales-allocated share of a CRM, and $3,000 of founder time on sales, for a quarterly total of $8,000 and 12 new customers, giving a CAC of about $667. Strip out everything but the ads and the same period would read $208, three times lower and completely misleading.
Two conditions make the result usable. Both terms must cover the same period, and the denominator must count only genuinely new customers. A renewal, an upsell or a contract extension is not an acquisition, and slipping them into the count quietly flatters the number.
A third condition is subtler. If your sales cycle runs six months, the spend from January produces customers in July. Dividing one month of spend by the same month of signatures then gives a figure that is wrong in both directions: too high while you accelerate, too flattering while you slow down. On a long cycle, offset the numerator by your average cycle length, or work over periods wide enough to absorb the lag.
What costs go into a fully-loaded CAC?
Include everything that would disappear if you stopped acquiring customers. That single test is the only one that holds up, and it pulls far more into the numerator than media spend.
Six buckets make up an honest numerator.
- Media spend, the budgets paid to ad platforms, social networks and search engines.
- Loaded salaries for marketing and sales teams, prorated to the time spent winning customers.
- External fees: agency, freelancers, consultants.
- Creative production: video, design, copywriting, photography, editing.
- Tools: CRM, email platform, analytics, automation licences.
- Direct sales costs: travel, trade shows, commissions on new business.
Three buckets stay out. Customer support and service, which belong to retention. The cost of producing or delivering the product itself, which belongs to gross margin. And marketing aimed at existing customers, whose economics differ. This is not a stylistic choice: a16z defines CAC as the fully loaded cost of acquiring a customer across all channels, and warns that founders routinely understate it by, in HubSpot’s phrase, “ignoring your time” and overlooking tool costs, an error that compounds as you scale.
This scope is not a bookkeeping detail. It decides what you believe about your own profitability. A media-only CAC is fine for choosing between two campaigns. A fully-loaded CAC is the only one that tells you whether the model holds. Publishing both, and naming clearly which one you are quoting, beats picking one and leaving the reader to guess.
Blended CAC, paid CAC and CAC by channel
You have at least three customer acquisition costs, and they are never equal. Confusing them is the most common source of budget-arbitration error.
| Version of CAC | What you divide | What it decides | Its limit |
|---|---|---|---|
| Paid CAC | Paid media spend, divided by customers from paid channels | Whether you can raise ad budget profitably | Depends on an attribution model, contestable by design |
| Organic CAC | Content, SEO and PR costs, divided by customers from those channels | Judging a long-term investment over several quarters | Long lag between spend and effect, attribution weaker still |
| Blended CAC | All acquisition spend, divided by all new customers | Steering the company, talking to investors, setting a global budget | Says nothing about any single channel |
a16z is blunt about which to lead with. Blended CAC “isn’t wrong”, but it “doesn’t inform how well your paid campaigns are working and whether they’re profitable”, so investors treat paid CAC as more important than blended CAC for judging whether a business can scale its acquisition budget. Their recommendation is to present both, plus a breakdown by channel, for example the cost per customer acquired through a specific platform.
Blended CAC keeps one virtue the others lack: it depends on no attribution model. It adds up what you actually spent and counts the customers you actually signed, which makes it the hardest figure to dispute and the right guardrail. A simple reading rule prevents most illusions: when paid CAC improves while blended CAC stays flat, you have probably not gained efficiency, you have moved conversion credit from one channel to another. That reporting gap is the same one we unpack in our comparison of the acquisition metrics, ROAS, MER, CAC and LTV.
What is a good CAC?
There is no universal good CAC, because the figure swings hard by channel and by deal size. First Page Sage (2026), drawing on roughly 120 firms, puts B2B paid search near $802 per customer and LinkedIn ads near $982, while thought-leadership SEO lands around $647 and account-based marketing near $4,664. Averaged out, the same dataset places organic channels around $942 and paid channels around $1,907 for B2B. A CAC that looks alarming on one channel can be perfectly healthy on another.
Which is why a CAC never means anything on its own. It only takes on meaning against what a customer is worth, and the reference most people reach for is the LTV:CAC ratio, with a widely cited floor of 3:1. That convention was popularised by investor David Skok in his SaaS Metrics framework, drawn from mature public SaaS companies, alongside a companion rule to recover acquisition cost inside about 12 months. Treat 3:1 as a discussion marker rather than a grade, and track it over time on your own business. We take that ratio apart, floor and ceiling, in which acquisition metric to run on.
The mistakes that distort a CAC
A handful of errors recur often enough to warrant a warning. None is a rounding detail: each can move the result by a large factor.
- Counting media spend only. The most common mistake, and it produces a flattering CAC that ignores the heaviest bucket, salaries.
- Mixing blended and paid. Comparing last quarter’s paid CAC to this quarter’s blended figure is comparing two different measurements and calling the difference progress.
- Counting renewals as acquisitions. The denominator swells, the CAC drops, and you mistake retention for conquest.
- Reading a platform’s CPA as your CAC. An ad platform reports the conversions it credits to itself. Across 663 randomised experiments at Facebook, Gordon, Moakler and Zettelmeyer (Marketing Science, 2022) found observational attribution overstated the real lower-funnel effect by a factor of 4.8 to 12.8. For scale, Triple Whale (2025) measures a median Meta CPA of $38.19 across roughly 35,000 brands: a useful campaign signal, but not a company acquisition cost. The same gap drives much of why GA4 and Meta conversions never match.
- Changing the scope midway. A CAC that improves because you quietly dropped salaries from the maths is not an improvement, it is a rule change. Document the scope and keep it stable.
How do you reduce CAC?
The durable levers work on the ratio itself, not on the ad budget in isolation. Cutting spend usually raises CAC rather than lowering it, because you lose the learning volume and efficient reach before you lose the cost.
Four levers actually move the number. Lift conversion on the traffic you already pay for, so the same spend buys more customers. Shift mix toward channels with a lower loaded cost, which is a measured decision once you hold budget split between Google and Meta to a constant scope. Invest in brand, so demand arrives warmer and costs less to convert, an effect we weigh in does branding lower CAC. And tighten lead qualification, so sales time, one of the largest buckets in a fully-loaded CAC, is spent on prospects who can actually close.
In short
- Set your scope before you calculate. A cost enters the numerator if it would vanish the day you stopped acquiring customers. Salaries, agency, creative and tools included, not media spend alone.
- Keep three numbers apart. Paid CAC to judge whether you can scale spend, organic CAC to judge the long game, blended CAC to steer the company. a16z treats paid CAC as the more telling one for growth.
- Read CAC against value, not against a universal target. First Page Sage shows B2B channel CAC ranging from about $647 to over $4,600, so the real test is your LTV:CAC ratio, with 3:1 as a marker rather than a grade.
A defensible acquisition cost is not declared, it is built with a written scope rule held steady over time. This is the groundwork we lay in our B2B paid acquisition work before arbitrating a single dollar of spend. If you want to pin these calculations to your own numbers and see what your customers really cost to win, book a diagnostic.