CAC in complex B2B sales: how to calculate it when the lead goes through several touchpoints

CAC in complex sales goes wrong when you divide one month's investment by the same month's sales. In a long cycle, the customer who closed in September came from budget spent in June, and that lag makes the number lie in both directions: it understates the cost when you have just cut budget, and overstates it when you have just scaled up.

We have already written about why complex sales do not scale without measurement. Here the subject is the calculation itself, and it involves three decisions that need to be made before any math.

It is worth saying what makes this scenario different. In a short-cycle sale, the person clicks, buys, and the report closes the same week. In a complex sale, the contract goes through several people on the customer's side, the cycle lasts months, and the lead touches several channels before deciding. Each of these three characteristics breaks a part of the standard calculation, which is why the math needs to be redone.

Decision 1: what goes in the numerator

The CAC formula is a simple division, and almost every error is in what you put on top.

The most common mistake is reporting only media. Ad spend divided by new customers gives a nice but incomplete number. In a B2B operation, a good part of the acquisition cost is people: the salaries of the pre-sales and sales team, commissions, prospecting and enrichment tools, event participation.

The rule that avoids confusion is to declare the scope along with the number. "Media CAC" and "fully loaded CAC" are two different metrics, and the second is usually several times larger than the first. Swapping one for the other in a presentation to leadership is the kind of error that only shows up when someone compares it with the financial results.

Decision 2: when the cycle ends

It seems like a detail, and it changes the denominator. Does a lead who signed the contract but has not yet paid count? What about one who signed and canceled within thirty days?

It is worth defining a clear milestone and sticking to it. Exact Sales, for example, considers the cycle closed only when the customer pays the first installment, which excludes mid-way dropouts. Any criterion works, as long as it is the same in all of the company's reports.

Decision 3: cohort or period

This is the decision that changes the result the most, and the one that almost nobody makes consciously.

Looking by period means taking September's investment and dividing it by the customers who closed in September. It is the automatic calculation of almost every tool, and it works well in a short cycle.

Looking by cohort means taking the leads who came in during June, tracking how many of them became customers over the following months, and dividing June's investment by that number. It is more laborious and it is the only approach that makes sense when the cycle lasts months.

The difference shows up quickly in an example, with values used only as an example. Suppose an average cycle of three months. In June you invested R$ 60 thousand and brought in 200 leads. In September, 15 customers closed, almost all from that June cohort. If you look by period, you will divide September's investment, which was R$ 30 thousand because the budget dropped, by the 15 customers, and you will celebrate a CAC of R$ 2 thousand. Looking by cohort, the real cost of those 15 customers is R$ 60 thousand divided by 15, or R$ 4 thousand.

CAC did not drop by half. You just measured the wrong thing.

Resource · In PortuguesePrompt Guide for Marketing and Sales70 prompts with questions about marketing, sales and behavior to ask your data.Download the guide

The middle path: lag the numerator

Not every operation has the structure to track cohorts from the start. There is an honest and simple approximation: lag the numerator by the average cycle.

If your average cycle is three months, divide June's investment by September's sales. It is not as precise as the cohort, because not every September customer came from June, but it corrects the biggest distortion. It is the path recommended by those who work with unit economics in long cycles: lag the spend by N months, where N is the average sales cycle.

One caveat that changes the decision: this approximation gets worse when your budget swings a lot from month to month. If the investment is stable, it works well. If it doubles and halves without a pattern, the cohort is the only reliable path.

To find your N, use the median time between lead creation and closing, not the average. In a long cycle, the average is pulled up by two or three contracts that took a year, and they do not represent the operation. It is worth recalculating this number every semester, because a change in ICP or in the sales process moves the cycle.

How the attribution model changes CAC by channel

So far the text has dealt with aggregate CAC. When you break it down by channel, a second variable comes in: which touch gets the credit for the sale.

The same customer may have discovered the company through a LinkedIn ad, come back through organic, and converted on a brand ad on Google. In first-click, that customer's cost is attributed to LinkedIn. In last-click, to Google Ads. CAC by channel changes entirely, without a single sale having changed.

That is why CAC in complex sales must always come with two labels: the attribution model and the time window. We cover the choice between models in the article on last-click versus first-click.

The most common practical mistake here is comparing the CAC of two channels calculated under different models. It happens when someone pulls the cost from the Google dashboard, which works on last click, and compares it with a lead source report from the CRM, which stores the first touch.

What to do with the number once it is calculated

Calculating it right serves three concrete decisions.

Where to increase investment. The channel with the lowest CAC deserves more budget, as long as the average ticket of its customers is not proportionally lower. Low cost with a low ticket is not an advantage, and this reading only appears when the deal value goes from the CRM back to the marketing report.

Where to investigate the process. CAC rising with stable volume almost always points to a drop in conversion rate at some stage, not to more expensive media. It is worth looking at the funnel before blaming the auction.

How much the company can spend. This is the reading that only exists when CAC talks to the customer's value over the course of the relationship. A high CAC is not a problem in itself: it is a problem when the customer returns little, or when the return takes so long that the company does not have the cash to wait. In a long cycle, payback time usually weighs more in the decision than the ratio between the two numbers.

The feature that solves this in practice

The difficulty of calculating CAC in complex sales is rarely the math. It is being able to look at the same dataset through two different time rulers.

In Laiki, the date filter by business rule exists precisely for this: you choose whether the cut considers the lead creation date, the sale date, or the date of the last event. Each one answers a different question, and the base is the same. The cohort stops being spreadsheet work and becomes a filter choice.

If you want the formulas and a calculation example for each funnel indicator, including CAC and LTV, the Marketing and Sales Metrics Guide is free.

Frequently asked questions

What is the CAC formula in complex sales?

The same as always, investment divided by new customers, with two corrections: include the sales team's cost in the numerator and match the investment period to the period in which that investment generated customers.

Does the sales team's salary go into CAC?

In fully loaded CAC, yes, along with commissions and tools. In media CAC, no. Both numbers are useful, and the problem appears when you use one under the name of the other.

How often should you calculate it?

Monthly to track the trend, and always over a rolling window of three months or more in a long cycle. A single month, in a long-cycle operation, has little volume and fluctuates by chance: two extra contracts in a weak month bring CAC down without anything having improved in the operation.

How do you compare your CAC with the market?

With great care. Most public numbers come from American SaaS, in dollars, with a cost structure different from yours. The comparison that decides anything is with your own history and with the value the customer generates over the course of the relationship.