Picture a company that loses an estimated $200 every time it wins a customer. Wikipedia's entry on customer acquisition cost gives the case: $500 spent to land a customer with an expected lifetime value of $300. The dashboard for that company would show a CAC of $500 and nothing else.
On USS Jefferson City I wrote and signed off more than 200 QA packages. A QA package never asks how cheap the work was. It asks whether the work holds at test depth. I've carried that question into every company I've worked in, and it's the one most CAC reports skip.
Two companies can report the same CAC and sit in opposite shape. One buys customers who stay, pay and refer. The other buys customers who refund in month two and clog the support queue while sales celebrates the close rate. The formula can't tell them apart. You pay for the difference.
I was an operating partner in an e-commerce company that spent heavily on ads every single day, at a scale where the full cost of a customer stopped being an academic question. Every dollar the number leaves out shows up later, in a department that never saw the ad.
How do you calculate customer acquisition cost?
Add up what you spent to win customers in a period, then divide by the number of new customers in that same period. The American Marketing Association describes the core formula as total acquisition spend divided by new customers in the same time period.
The arithmetic is easy. The argument is about what goes in the top line. Hyperline works one example where a SaaS company spends $100,000 on digital advertising, $50,000 on content, $80,000 on sales salaries and $20,000 on CRM tools. That adds up to $250,000, and with 500 new customers the CAC is $500 per customer.
| Cost line | Amount |
|---|---|
| Digital advertising | $100,000 |
| Content marketing | $50,000 |
| Sales salaries | $80,000 |
| CRM and sales software | $20,000 |
| Total acquisition cost | $250,000 |
| New customers | 500 |
| CAC | $500 |
Source for every figure in the table: Hyperline's worked example.
Salaries and software sit right next to ad spend in that example. If your number only counts ad spend, you're reporting a media cost and calling it CAC. It will look better than the truth, and it will keep looking better right up to the quarter when margin doesn't follow.
What counts as a new customer?
A new customer is a person or company that has paid you. Plenty of dashboards count a lead, a trial or a booked call as the denominator, and the number gets friendlier every time they do.
A Harvard Business School background note warns against common errors such as excluding relevant acquisition costs, including retention or overhead costs, counting users too early in the funnel, ignoring timing, or treating low CAC as inherently good. Each of those errors makes the number prettier or muddier, and none of them makes it more accurate.
Pick one definition of customer and one list of cost lines. Write both down and leave them alone for the year. When CAC moves because someone redefined the denominator, you've learned something about your bookkeeping and nothing about your market.
What is a good CAC?
A good CAC is one your customers pay back with room to spare. There is no universal dollar figure, because the right number depends on what a customer is worth to you.
The usual yardstick is the ratio of lifetime value to CAC. Wikipedia calls 3:1 a very good level because the customer relationships are solid and customers are acquired for the right price. It also gives the downside case from the top of this page, where $500 spent on a customer worth $300 loses an estimated $200 per customer acquired.
The ratio is only as honest as the lifetime value inside it. If that value is a guess built on a retention rate nobody has checked, a 3:1 ratio is a guess too. The HBS note says CAC should be treated as a managerial construct and read alongside customer value, retention, contribution and payback to judge whether growth is economically sustainable.
Why does the cheapest lead often become the most expensive customer?
Because CAC stops at the sale, and the cost of a customer keeps going. The cheapest lead is often the most expensive customer. A discount offer, a loose qualification step or a broad ad audience all lower the cost per lead, and they also raise the odds of a bad fit.
I call the result the growth trap. Marketing finds a cheaper source of leads and CAC drops, sales closes more deals because volume covers a weaker close rate, operations gets more customers than it was staffed for and quality slips, finance sees refunds, chargebacks and churn rise, and leadership, still looking at the original dashboard, asks marketing for more leads.
Every department did its job by its own scorecard, and the loop closed anyway. It closed on the one number that looked healthy the whole time. The same loop turns up in financial services, marketing technology and healthcare, anywhere a cheap lead can turn into an expensive client.
How do you read CAC across the whole business?
Read it next to what happens after the sale. One number from marketing can't show you the constraint, and four numbers from four departments can.
| Department | What it reports | What it can't see |
|---|---|---|
| Marketing | Cost per lead, CAC | Whether the leads fit the product |
| Sales | Close rate, cycle length | Which closed deals churn |
| Operations | Delivery capacity, support load | Which channel sent the load |
| Finance | Refunds, margin, collections | Which campaign caused them |
The table is my own framing of how most companies are organized. Each row holds a true number and a blind spot, and the blind spot always sits in the next row's data.
That's the idea behind the Four-Pillar Constraint Method. A company is a four-sided scale of marketing, sales, operations and finance, and its output is set by the one limiting step, which usually sits between two departments. I start with the money, then follow a customer from first touch to sale and from sale to delivery. Connect marketing spend, close rate, delivery capacity and refunds for the same cohort, and the picture changes. A channel with a high CAC can be your best one. A channel with a low CAC can be the reason operations is on fire.
Every fix moves the bottleneck. Cut your CAC without checking the next stage and you may only hand the problem to operations.
What does CAC look like to a buyer?
A buyer doesn't care that your CAC fell. A buyer cares whether the revenue behind it will still be there after you leave. If you're two to five years from a sale, a diligence team will rebuild your CAC from the raw data, count the costs you left out, and check whether your best customers came from a channel you can repeat or from your own network.
Buyers pay for systems, not heroes. If your acquisition runs on one salesperson's relationships or your own phone calls, the number may be real and still be worth little. This is the value cap in practice: a company's value is capped at the level its CEO is comfortable running, and a CAC that only works while you're in the room is part of that cap. A CAC you can reproduce through a documented process is worth more than a lower one you can't.
What are the common mistakes with CAC?
The most common one is counting only ad spend. Salaries, software and agency fees are acquisition cost, and leaving them out makes the number look better than the business feels.
Timing trips up almost as many people. If your sales cycle runs ninety days, this month's spend didn't produce this month's customers, so match spend to the customers it helped create. The HBS note lists ignoring timing among its common errors.
A blended number for everything hides the channel that loses money and the one that carries the company, so split it by channel and by cohort before you cut a budget. And a falling CAC means little until you've checked refunds, churn and support load for those same customers.
Last, watch for departments keeping their own definitions. Marketing's CAC and finance's CAC will disagree until somebody writes one down, and the disagreement usually surfaces in a board meeting.
FAQ
What does customer acquisition cost mean?
Customer acquisition cost is the total you spend on sales and marketing to win new customers, divided by how many you won in the same period. It tells you what one customer costs to get. It says nothing about what that customer costs to keep or serve.
What is a good CAC?
CAC is a dollar amount, so a percentage doesn't apply. The closest benchmark is the ratio of lifetime value to CAC, where Wikipedia cites 3:1 as a very good level. Your own margins and retention decide what is good for you.
How is CAC different from cost per acquisition?
Cost per acquisition usually means the cost of one conversion, such as a lead, sign-up or sale, often measured per campaign. CAC is stricter. It counts only paying customers and usually includes salaries and tools, not just ad spend.
What is acquisition cost in simple words?
It's the price you pay to land one new customer. Take everything you spent on getting customers this quarter, divide by the number of customers you got, and you have it. If you spent $250,000 and won 500 customers, each one cost $500.
What should you do this week?
Pull four numbers for your last full quarter, for the same group of new customers: what you spent to acquire them, how many closed, how many refunded or churned, and what they cost to serve. Put them side by side on one page. If you can't get all four, that's your first finding, because it means no one in the company is looking at the whole cost of a customer.
If you want to see how I read a company as one system, the Field Notes are where the writing lives, and the home page lays out the value-cap thesis. The value-leak assessment asks related questions about the rest of the business. Which of your four numbers would be hardest to pull?
Jeff Barnes · Growth Operating Partner · Greater Seattle-Tacoma