Customer acquisition cost formula for ecommerce

Customer acquisition cost (CAC) is what you pay in advertising to win one customer. For a shop running Google Shopping or Performance Max, the formula is the easy part. The harder question is what you can afford to pay for a first order, because a first order is often the only purchase you can count on.

nCAC (new customer acquisition cost) is how this post writes CAC, to make clear that only first-time buyers are counted, and section 3 turns it into a break-even you can set for your own products. The worked example uses round numbers you can check by hand, so swap in your own. For the short definition, see CAC in the glossary.

1. How do you calculate customer acquisition cost?

The formula is customer acquisition cost = ad spend ÷ new customers. Google Ads defines it the same way: according to Google Ads Help, Customer Acquisition Cost is “the ad spend allocated to new customers divided by the total of unique new customers acquired through a campaign.”

Take an example with round numbers (illustrative data). A shop spends €10,000 on Google Ads in a month and receives 200 orders at an average of €100. Of those, 120 came from customers who had never bought before, and each of them ordered once.

CPA is €10,000 ÷ 200 orders, nCAC is €10,000 ÷ 120 new customers, and ROAS is €20,000 revenue ÷ €10,000 of ad spend. To run the same sums on your own account totals, use the MER and nCAC calculator.

Cost per order is the number that flatters you. It spreads the spend over 80 orders from returning customers, who were not acquired that month, so it understates what each new customer cost. The same split is what separates cost per acquisition from CAC.

2. nCAC, CPA and ROAS answer different questions

nCAC tells you what a new customer costs, CPA what an order costs and ROAS what revenue each euro of ads returned. They use the same ad spend but divide it by different things, so they can disagree about the same campaign.

The example looks acceptable on CPA and ROAS, but at a 40% margin its break-even ROAS is 2.5, so a 2 ROAS loses money. POAS is ROAS multiplied by the contribution margin, here 2 × 0.40 = 0.8; see POAS vs ROAS for the maths. For where each metric sits at account level, see ROAS, MER, POAS and nCAC.

nCAC adds what ROAS cannot: how much of the spend went on customers who had never bought. A campaign can show a strong ROAS because it reaches people who would have bought anyway, and a high nCAC shows you are not reaching new ones.

3. Break-even nCAC

The most a new customer can cost on their first order is the profit from that order. That is the break-even nCAC: basket value × contribution margin, with prices excl. VAT and the margin after cost of goods and variable order costs.

In the example, the basket is €100 and the margin is 40%, so first-order profit is €40.00. The nCAC is €83.33, so the first order loses €43.33 per new customer.

A first-order loss can be a deliberate bet. If customers come back, later orders can repay it, and shops that can wait for repeat orders may accept a gap. Place the bet on the number of orders you actually see, not on hope. If you value new customers on a forecast instead, predicted LTV explains how to check it against the orders that follow.

At €40.00 profit per order, this customer needs a third order: two orders reach €80.00, which is still €3.33 short of €83.33. If your new customers rarely place a third order, the nCAC is too high, whatever the average nCAC says. LTV:CAC for ecommerce covers how to count those later orders. To find the break-even for a product, start from its margin with the break-even ROAS calculator.

4. How Google counts new customers

Google Ads can report customer acquisition, but only for Purchase conversions. According to Google Ads Help, “Customer acquisition reporting only works with ‘Purchase’ conversions in Google Ads.”

Three parts of Google’s setup decide how nCAC is counted:

• Detection: for each conversion, Google can auto-detect whether it is new, or you can report it yourself via the conversion tag, which Google says allows better accuracy. The tag parameter is new_customer, and for Google Analytics or uploads it is customer_type with the values new and returning.
• Lapse window: 540 days is recommended and set by default, but not required. It decides how long a past customer still counts as existing.
• Goals: per Google Ads Help on customer lifecycle goals, New Customer Value tells Google to “bid higher for new customers than existing ones”. New Customer Only means “only bid for new customers”. New Customer Value works with Target ROAS and Maximise conversion value in Search, Performance Max, Shopping and Demand Gen campaigns.

A first order is the signal Google gets at bid time, so the quality of the new or returning flag matters. With Full Signal Tracking every order is marked new or returning, so you can increase priority for first orders. It gets through every adblocker on a unique endpoint on your own domain, so the new customer purchases a blocker would otherwise remove are still counted. Details are on the Full Signal Tracking page.

New customers get their own conversion action in Google Ads. Product Metrics delivers these as separate conversion actions: Product Metrics - New Customer, New Customer Profit, Returning Customer, Returning Customer Profit, Revenue, Profit, Revenue LTV and Profit LTV. That lets Smart Bidding value a first order differently from a repeat one.

5. A weekly nCAC routine

Keep nCAC under what a first order earns, or know exactly why you accept the gap.

Keep the weekly check short. Take last week’s ad spend and new customers, calculate nCAC and put it next to first-order profit. Allow for conversion lag before you judge a recent week, because orders can arrive days after the click.

Judge a change over several weeks, not one. A single week with few new customers swings nCAC sharply. Write down the window you compare, so a month of promotions is not set against a quiet one.

6. nCAC per product

nCAC per product shows which products bring new customers cheaply and which do not. Product Metrics shows nCAC per product and as an average.

In this example, Product C costs nearly four times as much to win a new customer as Product A. The average of €83.33 describes none of them. A product with a low nCAC and a good margin is a place to increase priority. A product that only wins customers at a high nCAC needs a closer look at its price, margin or ads before it takes more budget.

Seeing nCAC per product next to the average is what makes Product C’s €150.00 visible before it takes more budget. Product Segmentation then places each product in one of six segments by its ad clicks and return, so you see where every product stands without working through them one by one; it suggests, and you decide what each product gets.

Read product-level nCAC with the same care as the average. It rests on the orders and new customers your ad platform and shop recorded, so a product whose purchases are missing from either looks cheaper or dearer than it is. So get the purchase data complete first, with server-side tracking where a blocker eats orders, and only then act on these numbers.

Start with one best seller

The break-even ROAS calculator gives you the margin-based floor for each product, and break-even ROAS per product walks through the maths. To send new and returning customers to Google Ads, see Full Signal Tracking, which comes with the Pro subscription at no extra charge; see the pricing page.

I’d start with one best seller. Work out its break-even nCAC (price excl. VAT times contribution margin), divide last month’s ad spend by its new customers, and compare the two. If nCAC comes out higher, count how many later orders those new customers actually place.