CAC: the acquisition cost that almost everyone miscalculates

Most companies I see reporting CAC are reporting something else. They take what they spent on ads in the month, divide it by the number of customers who came in and call it the acquisition cost. The number looks nice, fits on a slide and hides half of the expense that actually occurred to bring in that client. Undervalued CAC is the most elegant way to lie to yourself about the health of the business.

The customer acquisition cost answers a direct question: how much did your company need to spend, in total, to acquire a new paying customer. The definition is trivial. The difficulty lies in deciding what counts as "disbursement to achieve" and in what window of time. This is where the math separates between those who are measuring reality and those who are deceiving themselves.

The account that seems obvious and isn't

The CAC formula, in prose, is the total invested in acquisition in a period divided by the number of new customers acquired in that same period. Two terms, one division. The problem is in the numerator.

Almost everyone puts paid media there and stops. But what brought that customer was not just the ad. It was the salary of the marketing team that ran the campaign, the commission of the salesperson who closed it, the cost of the automation and CRM tools, the content budget that warmed up the lead months before, the events, the giveaways, the pre-sales team that qualified. All of this existed for that name to become a client. Removing these lines from the account does not make the CAC smaller, it makes it false.

The rule I use is simple to state and uncomfortable to apply: every cost you only incurred because you wanted to acquire customers goes into CAC. If the expense disappears in a world where you stop prospecting, it is an acquisition cost. Salesman's salary comes in. Salary for those who support existing customers, no. Media enters. Server that runs the product, no. The boundary is the intention of the expense, not the department it comes from.

The denominator is also deceiving. A new customer is a new paying customer, not a lead, not a registration, not a trial. Mixing people who signed up with people who paid inflates the number of acquisitions and artificially lowers the CAC. You think you buy things cheaply because you're counting people who haven't even opened their wallets as wins.

The time mismatch that distorts everything

There is a more subtle trap, and it is temporal. The money you spend on acquisition in one month rarely reaps customers in the same month. A content campaign planted in January could close accounts in April. A hired salesperson now takes a quarter to fill the pipeline.

When you divide a month's expenditure by customers from the same month, you are crossing today's expense with a result that came from previous efforts, and ignoring that today's expense will appear as a customer later on. In stable operations, this is diluted and the error becomes small. In operations that accelerate or slow down investment, it is very distorting: those who are scaling acquisitions seem to have a CAC that is too high, because they reap tomorrow what they sowed today; whoever is cutting appears efficient, because he still harvests what he planted before cutting.

The honest fix is to look at cohorts and give the achievement a maturation window compatible with your sales cycle. For a low-ticket self-service product, the month ends quickly. For a six-month consultative sale, comparing spend and customer in the same month means nothing.

Blended hides where the money works

The blended CAC is the average of everything: the entire acquisition cost divided by all new customers, without separating where they came from. It is the number that appears on the board and serves as a general health reading. It has a real value, which is to say whether the entire business is purchasing within a sustainable range.

But blended mixes what shouldn't be mixed. Inside it coexist customers who arrived for free through word of mouth and customers who cost dearly at an advertising auction. The average dilutes the two. You look at a healthy blended CAC and conclude that your acquisition is going well, without realizing that it is going well because of organic and that every dollar invested in paid media is resulting in loss.

CAC per channel dispels this illusion. You separate the cost and customers by origin: how much does a customer coming from paid search cost, how much does one coming from a referral cost, how much does one closed by the outbound sales team cost. Then the truth appears. There is almost always a channel that carries the average and one that bleeds. Blended tells you if the boat floats. The per channel tells you which motor pushes and which hole sinks.

The practical consequence is where the money is. Budget decisions, hiring decisions, where to double the bet, none of that happens with blended. If I don't know which channel brings me good and cheap customers, I can't reallocate funds intelligently, I can only increase or decrease total spending in the dark.

Why isolated CAC lies

A low CAC is not necessarily good news, and a high CAC is not necessarily a problem. The number alone is meaningless because it only answers half the question. Knowing how much it costs to acquire a customer without knowing how much that customer is worth is like knowing the price of something without knowing what it does.

A company could have the lowest CAC in the market and be dying if the cheap customers it attracts cancel within two months. Another can pay a lot per customer and prosper, if each customer stays for years and expands the account. The acquisition cost only becomes useful information when it meets the customer's value. This is why it should almost never be read alone.

The reading that matters crosses the CAC with two other things. First, with the value that the customer generates throughout their life, in the LTV to CAC ratio, which tells whether the account closes. Second, over time, in the CAC payback period, which says in how many months the customer returns what it cost. Without these two crosses, the CAC is an orphan number. It is part of the unit economics of SaaS, and it is there, in the confrontation between cost and value, that it finds meaning.

There is also the scale effect that isolated CAC hides. A channel's first customers are often cheap because you capture the hottest, most obvious demand. As you scale, you need to look for less ready people, and the cost per client goes up. A CAC that was great can deteriorate precisely because you grew, and anyone who looks at the number for an isolated quarter doesn't see this curve rising.

What does this change in the mind of those who decide

The discipline here is not to calculate the CAC with more decimal places. It means deciding, frankly, what goes into the bill and defending that definition against the temptation to embellish. The most useful CAC is the most complete, even though it is the ugliest. A high and true number protects you; a low, made-up number pushes you to climb a hole.

In the routine of those who lead, this becomes three habits. Always report the CAC per channel alongside the blended, because the average is for the board and the detail is for the decision. Recalculate the frontier of what goes into the cost every time the acquisition structure changes, because a new sales team changes the entire account. And never present CAC without customer value on the side, because one without the other is a half-truth that leads to error.