Customer Acquisition Cost (CAC)

Definition, formula, a worked example, and the mistakes that make it wrong. For DTC brands.

What is customer acquisition cost (cac)?

Customer acquisition cost (CAC) is acquisition spend divided by the new customers it is counted against, and blended, attributed, and incremental CAC each count those customers differently.

Formula

Acquisition spend ÷ new customers, over the same period

What each part means

Blended CAC
All acquisition spend divided by all new customers in the month. A health check for the whole business that needs no model, but it cannot tell you which channel to change.
Attributed CAC
Channel spend divided by the new customers a model credits to that channel. Platforms credit the same customer, so never add platform credits together.
Incremental CAC
Channel spend divided by the new customers the spend caused, usually estimated from a lift test that compares people who saw ads with a group who did not.
New customer
A customer placing a first order, using the same definition everywhere. Amazon new-to-brand uses a different rule, so keep those counts apart.
Acquisition spend
Media spend, plus any agency fees or creative you choose to include. Decide once and use the same rule everywhere.

Worked example

A store spends $60,000 in a month and gets 1,500 new customers, so blended CAC is $40. Social spent $30,000 and a dedup model credits it with 800 new customers, for attributed CAC of $37.50. If a lift test found social caused only 400 of them, incremental CAC is $30,000 ÷ 400 = $75.

Where the numbers come from

  • Ad platform spend by channel and day (search, social, video, retail media)
  • Agency and creative invoices, if your definition includes them
  • Shopify customer records for first orders
  • An attribution model or platform credit reports for attributed CAC
  • Lift test results (such as Google Conversion Lift or Meta lift reporting) for incremental CAC

Mistakes that make it wrong

Reading cost per purchase as CAC
Ad platforms usually report cost per purchase, which includes repeat buyers. $30,000 over 1,500 purchases is $20, but over 900 first-time buyers it is $33.33.
Adding up platform credit
Platforms each claim the customers they touched, so their credits can add up to more new customers than the store got.
Mismatched periods or customer definitions
Spend and new customers must cover the same period, and new customer must mean the same thing in both. Do not mix new-to-brand with first order.
Treating unknown as zero
If spend or new customers is unknown, CAC is unknown, not $0. If new customers is zero, CAC is undefined.

What good looks like

We do not publish a target CAC, because it only means something next to what a customer is worth. Compare each channel's CAC with its contribution LTV at the same age, track the trend, and label every CAC as blended, attributed, or incremental.

Common questions

What is the difference between blended and incremental CAC?

Blended CAC divides all spend by all new customers, including people who would have come anyway. Incremental CAC divides a channel's spend by only the customers it caused, usually measured with a lift test.

Which CAC should guide the next dollar of spend?

Blended, attributed, and incremental CAC are all averages. Marginal CAC, the extra spend divided by the extra customers it brings, is the one built for the next dollar.

Why is attributed CAC lower than incremental CAC?

Attribution credits customers who would have bought anyway, while incremental CAC counts only customers the spend caused. A channel can look cheap on credit and expensive on cause.

Can I compare DTC CAC with marketplace CAC?

Only with care. Shopify new customers are first orders with your store, while Amazon new-to-brand looks back 12 months on Amazon. Watch each over time rather than ranking them against each other.

Should CAC include agency fees and creative?

That is your choice, but make it once, write it down, and use it everywhere so CAC is comparable month to month.

Want customer acquisition cost (cac) from your own data?

We build the reports that calculate it the same way every month, from the systems you already use.

Book a Call