The calculator above turns a period's sales and marketing spend into blended CAC, paid CAC, and a payback period. Below is what each number means, what belongs in the spend, and the ways CAC quietly lies to the people reading it.
What CAC is and how to calculate it
The CAC number I've seen in board decks most often is ad spend divided by every new customer. It's the easiest version to pull and the most flattering one. It leaves out the people running the campaigns, the software they run them in, and the freelancers who make the creative. Then it credits paid media with customers who showed up from a podcast mention or a friend's recommendation. In 12+ years of running paid media, I've rarely seen that version come in anywhere near the real cost.
CAC is ad spend divided by new customers everything you spent to win customers divided by the customers who paid.
Customer acquisition cost is the total you spent on sales and marketing in a period divided by the number of new paying customers you won in that same period. Paying is the word that matters. Leads, trials, and free signups aren't customers yet, and counting them turns CAC into a CPA with a different name.
The formula. CAC = total sales and marketing spend / new paying customers. Same period on both sides, and only customers who actually paid.
A worked example
Say a subscription software company spends this much on acquisition in one quarter.
| Line item | Quarter |
|---|---|
| Paid media on Google, Meta, and LinkedIn | $30,000 |
| Marketing salaries, the share spent on acquisition | $15,000 |
| Tools (CRM, email platform, analytics) | $2,000 |
| Freelance designer and copywriter | $3,000 |
| Total | $50,000 |
That quarter brought in 250 new paying customers. Blended CAC is $50,000 / 250, or $200. The attribution report credits 120 of those customers to paid campaigns, so paid CAC is $30,000 / 120, or $250. The ads dashboard, which doesn't know about salaries or organic customers, divides $30,000 by all 250 and reports $120. Same quarter, same company, three different CACs, and the one on the dashboard is the lowest of the three. They're the calculator's default numbers, so you can watch each one move as you change the inputs.
Blended CAC vs paid CAC
Blended CAC divides all acquisition spend by all new customers, whatever brought them in. Organic search, referrals, word of mouth, and direct traffic all land in the denominator. It answers one question well. What does it cost this business, all in, to add a customer?
Paid CAC divides paid media spend by the customers your attribution credits to paid channels. It answers a different question. What does the next dollar of ad spend buy? That's the number to use when you decide whether to raise or cut budget, because organic customers keep arriving whether you spend or not. A stricter version adds the costs that exist only because of paid, like the paid media manager's salary and ad creative. Use it if you can split those out.
| Version | Formula | Use it for |
|---|---|---|
| Blended CAC | All acquisition spend / all new customers | Unit economics and board reporting |
| Paid CAC | Paid media spend / customers from paid | Budget calls and scaling a channel |
| Media-only CAC | Paid media spend / all new customers | Nothing. It's a flattering mix of the two |
The two drift apart for a predictable reason. As organic grows, blended CAC falls even while paid gets more expensive every month. I've watched teams celebrate a falling blended CAC while paid CAC climbed, because a strong content quarter was covering for it. Track both side by side every month. If blended improves and paid gets worse, the ads aren't the reason things look better.
Paid CAC depends on attribution, and attribution is never perfect. Last-click undercounts paid social and overcounts branded search. Pick one model, write down which one, and keep it consistent so the trend means something even when the absolute number is off.
What to include in CAC spend
The honest version counts every cost that exists to win new customers. The test I use is simple. If you stopped acquiring customers tomorrow, would this cost go away? If yes, it's in.
| Count it | Leave it out |
|---|---|
| Paid media on every platform | Customer success and support salaries |
| Sales and marketing salaries and commissions, by share of time on new customers | Account management for existing customers |
| Sales and marketing software | Product and engineering costs |
| Agencies, freelancers, and contractors | Retention and win-back campaigns |
| Content, creative, events, and sponsorships | General overhead like rent and finance |
| Affiliate commissions and referral bonuses | Discounts given to existing customers |
Salaries are where most CAC numbers go soft. A founder who spends half the week on sales calls is a sales cost. A marketer who splits time between acquisition and retention emails counts by the share that goes to acquisition. It doesn't need to be precise to the dollar. An honest estimate beats zero, and that's what the ads dashboard assumes.
Affiliate commissions and referral bonuses belong in the spend too. They're paid per customer, which makes them easy to forget because they never show up as a monthly budget line.
CAC payback period
CAC on its own says what a customer costs. Payback says how long you wait to get that money back. The formula is CAC divided by the monthly gross profit a customer brings in, which is monthly revenue per customer times gross margin.
Payback formula. CAC payback in months = CAC / (monthly revenue per customer x gross margin). A $200 CAC on a $50 a month plan at 70% margin earns back $35 a month and pays back in 5.7 months.
Use gross margin, not revenue. A $50 subscription that costs $15 a month to serve only returns $35 toward the CAC. Payback on revenue in that example says 4 months when the real answer is 5.7, and the gap is cash you don't have.
| CAC | Revenue / mo | Gross margin | Gross profit / mo | Payback |
|---|---|---|---|---|
| $200 | $50 | 70% | $35.00 | 5.7 months |
| $200 | $50 | 30% | $15.00 | 13.3 months |
| $60 | $20 | 40% | $8.00 | 7.5 months |
| $1,200 | $150 | 80% | $120.00 | 10.0 months |
| $3,000 | $200 | 75% | $150.00 | 20.0 months |
Simple payback assumes the customer sticks around long enough to finish paying. It doesn't account for churn. If 8% of customers cancel every month, the average one stays about 12.5 months (1 / 0.08). A 13.3-month payback on that base means the average customer leaves before covering what they cost. Check payback against expected lifetime every time you run it.
For ecommerce, swap monthly revenue for the gross profit a customer generates per month across repeat orders. A store whose customers order $40 every two months at 50% margin earns $10 a month per customer. If they don't come back much, the question that matters is whether the first order covers CAC, and the break-even ROAS calculator answers that one.
What payback period to aim for
These are my rules of thumb. They aren't benchmarks. Under 12 months is comfortable for most subscription businesses. Twelve to 24 months works if churn is low and you have the cash to float it. Over 24 months is a bet on retention that you'd better have data for. The calculator grades your result on those same lines.
What a good CAC is
There's no good CAC in the abstract. A $1,500 CAC is cheap for software that brings in $500 a month and ruinous for a $30 candle. The question is always what the customer is worth compared with what they cost.
That comparison is LTV to CAC. Lifetime value is the gross profit a customer generates over their life, roughly monthly gross profit times the number of months they stay. In the example above, $35 a month over an average 20-month life is $700 of LTV against a $200 CAC, a 3.5 to 1 ratio.
The ratio you'll hear quoted most is 3 to 1. Treat it as a rule of thumb with a reason behind it. Gross profit has to cover CAC plus the costs CAC leaves out, like product, support, and overhead, and still leave a profit. Below 1 to 1 you lose money on every customer. Far above 3 to 1 usually means you're underspending and leaving growth on the table. The LTV to CAC calculator handles the lifetime value side of that ratio.
Two checks I'd run alongside the ratio. Payback, because a 5 to 1 ratio that takes four years to collect can still run you out of cash. And trend, because a CAC that holds flat while competitors bid up the same keywords is better than it looks. A CAC that falls right after you cut budget hasn't improved. You just stopped buying the most expensive customers.
Know your CAC? The other half of the ratio is what a customer is worth.
Open the LTV:CAC calculatorCAC vs CPA
CPA, cost per acquisition, counts whatever conversion you told the ad platform to track. That might be a lead, a trial, a signup, or a sale. It's usually media cost only. CAC counts new paying customers and includes the full cost of winning them.
The two connect through your funnel. A $50 CPA on free trials with a 25% trial-to-paid rate is $200 of media per paying customer. Add salaries, software, and fees and the fully loaded CAC climbs again. A team celebrating a $50 CPA without running that math is reading the wrong number.
Use CPA to run campaigns day to day, since it's what the platforms optimize toward. Use CAC to decide whether the whole acquisition engine makes money. The CPA calculator covers the campaign side, including the most you can afford to pay per conversion.
Five ways CAC misleads you
Blended CAC hides paid getting worse
Organic growth lowers blended CAC even when paid hasn't changed at all. If you only report blended, a paid program can lose efficiency for months before anyone notices. Report paid CAC right next to it.
Media-only CAC leaves out the team
Ad spend divided by customers is the number the dashboards make easy. It skips salaries, software, and contractors, which on a small team can match the media budget. Any deck that uses it overstates how efficient acquisition is.
The time windows don't line up
Spend in March often buys customers in April or June, especially with a sales team or a long trial. Dividing March spend by March customers doesn't measure March. It mixes this month's cost with last quarter's results. If your sales cycle runs 60 days, lag the customer count by about 60 days, or measure over a full quarter so the edges matter less.
The denominator counts the wrong people
Trials, free accounts, reactivated customers, and upgrades from existing accounts all inflate the count when they slip in. CAC is new paying customers only. Check what the customer count actually contains before you trust the number sitting on top of it.
One average covers very different channels
A $200 blended CAC can be $80 from referrals and $600 from LinkedIn. The average doesn't tell you where the next dollar should go. Break CAC out by channel at least once a quarter, even when the channel numbers are rough.
How to lower CAC
Work through these in order. The early steps make the later ones cheaper.
- Fix the count first. Make sure the customer number is new paying customers and the attribution model stays the same month to month. A wrong denominator makes every later decision wrong.
- Cut what's above your ceiling. Sort channels and campaigns by paid CAC and payback. Anything that won't pay back inside your target window, with enough volume to be sure, gets cut or rebuilt.
- Raise conversion down the funnel. Landing page to signup, trial to paid, demo to close. A 20% better trial-to-paid rate lowers CAC by about 17% without touching a bid.
- Grow the channels that don't charge per click. Referrals, content, partnerships, and search traffic you've earned lower blended CAC every month they keep working.
- Shorten the sales cycle. Fewer touches and faster follow-up mean less salary per closed deal.
- Raise price or margin when CAC won't move. It doesn't change CAC, but it shortens payback and lifts LTV, and that comparison is the one that matters.
The bottom line
CAC is only honest when the spend includes everyone and everything it took to win the customer, and the count includes only customers who paid. Run blended and paid side by side. Check payback against how long customers actually stay. Then judge the result against your own LTV instead of someone else's benchmark.