Your blended CAC is hiding your best channel, and your worst one.
One averaged cost-to-acquire number can sit comfortably under target while a winning channel goes underfunded and a losing one keeps its budget. The average is not wrong, it is just answering a different question than the one you are asking. Here is how to split it apart, and a worked example of what changes when you do.
01
The average is doing exactly what an average does
What happens
Add total spend across every channel, divide by total new customers, and you get one number. It looks like a fact about the business. It is actually a weighted average of several very different numbers, and the weighting is just how much you happened to spend on each channel this month, not how well each one performs.
What to do about it
Before anything else, split spend and new customers by channel and compute CAC per channel. Only once those exist does the blended figure mean anything, as a summary of numbers you already have rather than the only number you have.
02
A cheap channel can be subsidising an expensive one
What happens
Say organic and referral bring in customers at a low cost and paid search or paid social bring them in at a high one. Blend the two and the average sits comfortably in the middle, under whatever target you set. Nobody looks twice at it. Meanwhile the expensive channel keeps its budget, because the number that would flag it never gets computed.
What to do about it
Rank channels by their own CAC against your own target CAC, not against the blend. A channel that clears the target on its own merits gets more budget. A channel that only looks acceptable inside the average gets a harder look, and a plan for either fixing it or funding it less.
03
New and returning customers get counted the same way
What happens
Some of the 'new customers' a channel gets credit for were coming back anyway, through a channel that happened to be the last thing they touched. Blended CAC does not distinguish a channel that creates a customer from one that recaptures someone who already knew the brand, so both get the same credit per acquisition.
What to do about it
Separate genuinely new customers, people with no prior order or account, from returning ones before computing CAC. A channel with a high cost per new customer but a low cost per returning one is doing a different job than the number implies, and deserves a different budget conversation.
04
Spend this month, conversions next month
What happens
A channel with a longer consideration window, content, SEO, some paid social, spends now and converts later. Divide this month's spend by this month's new customers and a channel that is actually working looks expensive, simply because the payoff has not landed in the same window as the cost.
What to do about it
Match spend to the cohort it produced, not the calendar month it was paid in. Give slower channels a longer window before judging their CAC, and say what that window is next to the number. A channel judged on too short a window will always look worse than a channel with an immediate payoff, whether or not that is true.
05
Media cost is not the whole cost
What happens
A CAC built only from ad spend ignores everything else that channel actually costs to run: the agency or platform fee, the content or creative production, the tooling. Two channels with identical media spend can have very different fully-loaded costs, and a media-only CAC will rank them as equal.
What to do about it
Decide once whether you are computing media CAC or fully-loaded CAC, and say which on every report. Fully-loaded is the honest number for a go/no-go decision on a channel; media-only is useful for day-to-day bid and budget pacing. Using the wrong one for the wrong decision is how a channel that is not actually profitable keeps its budget.
A worked example, with illustrative numbers
Say a business spends $60,000 across three channels in a month and books 300 new customers. Blended CAC: $200. Under a $250 target, every channel looks fine.
Split it: referral spent $5,000 for 50 customers, CAC $100. Organic content spent $10,000 for 80 customers, CAC $125. Paid social spent $45,000 for 170 customers, CAC $265, over target and carrying most of the budget. The blend never shows this, because $45,000 of cheap-looking spend is doing the averaging.
The decision the blended number could never support: shift budget from paid social toward referral and content until paid social’s own CAC clears target, or fix what is pushing its cost up. Numbers changed for illustration; the arithmetic is the point, not the figures.
Connect your ad platforms and your order data, and Decifer splits acquisition cost by channel and cohort so a losing channel can't hide inside a winning average.