How to calculate customer acquisition cost

Customer acquisition cost is total acquisition spend divided by the number of new customers it produced. Spend £1,840 in a month and win 92 new customers and your CAC is £20.00. That figure means nothing on its own: the only question that matters is whether £20 is less than what an order leaves behind after every variable cost, which in the worked example below is £21.90 — a margin of £1.90. This page shows how to build both numbers from your own records, how to split CAC by channel, and why the blended figure most sellers quote is the flattering one.

The formula, and what belongs in the top line

The formula is a division and the argument is entirely about what goes into it.

CAC = total acquisition spend in a period ÷ new customers acquired in that period

Two words do the work. Acquisition means money spent to make somebody who had never bought from you buy from you. New means exactly that: a repeat order is not an acquisition, and counting it as one is the most common way the number gets quietly improved.

What counts

• Advertising spend of every kind, including marketplace sponsored listings and boosted posts.
• Anything paid to a person or agency to produce or run that advertising.
• Software whose purpose is finding or converting new buyers.
• Product given away to seed reviews or sent to people who post about it, valued at what it cost you to make, not at retail.
• Discount funded specifically to win a first order, such as a first-purchase code. This one is regularly missed, and it is real money.

What does not count

• Making, packing and posting the goods. Those are costs of fulfilling the order, and they belong in the contribution calculation on the other side of the comparison.
• Spend aimed at people who have already bought — a loyalty email, a re-order reminder. That is retention, and folding it in makes acquisition look dearer than it is.
• General overheads that would be there with no advertising at all.

Pick a period long enough to contain a decent number of orders — a month is the usual choice — and be consistent about it, because the whole value of the measure is the comparison from one period to the next.

A worked month

Kestrel and Fern is an invented candle maker selling on its own store and one marketplace. Here is a full month of acquisition spend.

That month the shop took 140 orders. Of those, 92 were from customers who had never ordered before and 48 were repeat buyers. The two versions of the number:

New-customer CAC = £1,840 ÷ 92 = £20.00
Blended CAC = £1,840 ÷ 140 = £13.14

The blended figure is 34% lower and it is the one that tends to get quoted, because it is the one that comes out when you divide spend by orders without separating the two kinds. It is not useless — it tells you what you paid per order across the whole business — but it cannot answer the question people use it for. If you are deciding whether to spend more on advertising, the money buys new customers, so £20.00 is the price you are being asked to pay. Quoting £13.14 credits your advertising with sales from people who already knew you.

A shop with a strong repeat rate can make its blended CAC look excellent while its actual cost of new customers climbs every month, and it will keep looking excellent right up until growth stalls.

The number CAC has to be compared against

£20.00 is neither good nor bad until you know what an order leaves behind. That figure is the contribution: the price, minus every cost that only exists because the order happened.

Contribution per order = price − cost of goods − fulfilment − platform and payment fees

For a typical Kestrel and Fern order:

Contribution is 45.6% of the order value. Now the comparison that the whole measure exists for:

£21.90 contribution − £20.00 CAC = £1.90 left on a first order

The business is acquiring customers profitably, but only just, and only on paper. £1.90 per new customer across 92 of them is £174.80 for the month — before a single pound of rent, software, insurance or wages. On first orders alone, this shop is not viable. It is viable only if those customers come back, which is the next sum.

Note what happens if the average order value slips by a fiver, to £43.00. Contribution falls to £16.90 and the shop is now losing £3.10 on every new customer it buys, while every dashboard still shows revenue growing. This is the mechanism by which a business can advertise its way to a bigger loss, and the only warning is this arithmetic.

Payback: when the customer has paid you back

Most shops do not recover acquisition cost on the first order, and that is not automatically a problem. It becomes a problem when nobody has worked out how long the recovery takes, because the gap is funded out of cash.

Orders to recover CAC = CAC ÷ contribution per order
Months to payback = orders to recover ÷ orders per customer per month

Kestrel and Fern customers buy 3.2 times a year, which is 0.267 times a month. At £20.00 CAC and £21.90 contribution:

So the money spent winning a customer in January is back in the bank around the middle of April. That number is the one that decides how fast you can grow, because it is the length of the hole you have to fund. Doubling monthly ad spend does not double the shop; it doubles the hole for three and a half months first.

What each of those customers is worth over the whole relationship, and how to discount it for the fact that money arriving in 2029 is not money now, is worked in customer lifetime value.

Channel CAC, and what the blended number hides

A single CAC for the whole business averages good channels with bad ones and hides both. Split the spend by where it went and attribute the new customers to the same places. Creative costs belong with the channel they were made for, so the freelancer sits with paid social.

The overall £20.00 was concealing two different businesses. Paid social is buying customers for £27.04 and getting £21.90 back, losing £5.14 every time it works — 54 times that month, which is £277.56 of loss. Marketplace ads and email are both comfortably profitable and are the only reason the blended figure looks positive at all.

The arithmetic of moving money is worth doing before moving it. If £310 on marketplace ads produced 26 customers, that is one customer per £11.92. Moving £500 from paid social to the marketplace would, if the rate held, buy about 42 more customers there and lose about 18 from paid social: a net gain of 24 customers and roughly £525 more contribution.

The words if the rate held are carrying real weight. Acquisition costs rise as you spend more into the same channel, because the cheapest, most obviously interested buyers get reached first. A channel at £11.92 for £310 a month is not a channel at £11.92 for £810 a month, and the only way to find the new figure is to move the money in steps and measure. Treat the sum as a hypothesis with a number attached, not a forecast.

What customer acquisition cost cannot tell you

CAC is a divided average. It is genuinely useful and it is routinely asked to answer questions it structurally cannot.

• It cannot tell you which spend caused which customer. Attribution is an estimate, always. Somebody sees an advert on Monday, searches your name on Thursday and buys on Sunday; every platform involved will claim that order, and the totals will add up to more sales than you made. Channel CAC is directionally useful and it is not evidence.
• It cannot separate your advertising from your reputation. Some of the 92 would have found you anyway, through a recommendation or a search. They are counted as bought customers, which makes advertising look better than it is. The only honest way to size that is to vary spend deliberately and watch what happens to total orders, not to attributed ones.
• It is an average, so it hides the spread. A £20 average can be forty customers at £8 and fifty-two at £29. The average tells you nothing about which kind you get more of when you spend more, and that is the thing you actually need to know.
• It says nothing about whether the customer is any good. A channel with a low CAC that delivers people who buy once and never return can be worth less than a dear channel that delivers people who stay. CAC has to be read next to what those customers go on to spend, or it will reliably point at the wrong channel.
• It does not account for time. Spend leaves your account this week and contribution arrives over months. Two businesses with identical CAC and identical margins can have completely different cash positions, and the one with the longer payback runs out of money first.
• It is not a target in itself. Driving CAC down by cutting spend will improve the number and shrink the business. The number to grow is total contribution after acquisition cost; CAC is only one of the two things that produce it.

Arithmetic and general information only — not financial, tax, legal or investment advice. Your own figures, and your own accountant, decide what any of this means for you.

Frequently asked questions

What is the formula for customer acquisition cost?

CAC = total acquisition spend in a period ÷ the number of new customers acquired in that period. Spending £1,840 in a month and winning 92 first-time customers gives a CAC of £20.00. Use new customers rather than total orders, because repeat orders were not bought by this month’s advertising.

What costs should be included in CAC?

Advertising of every kind, anything paid to people or agencies to produce and run it, software whose job is winning new buyers, product given away to seed reviews valued at cost, and discounts funded specifically to win a first order. Leave out making, packing and posting the goods, and leave out spend aimed at existing customers.

What is a good customer acquisition cost?

There is no general figure, because the only meaningful comparison is against your own contribution per order. If an order leaves £21.90 after goods, fulfilment and fees, then a CAC of £20.00 is just about viable on the first order and a CAC of £27.04 loses £5.14 every time it works. Work out your contribution first; the CAC number is meaningless without it.

What is the difference between blended CAC and new-customer CAC?

Blended CAC divides spend by all orders, including repeat ones; new-customer CAC divides it only by first-time buyers. In the worked example the two are £13.14 and £20.00 for the same month. The blended figure flatters shops with strong repeat business and is the wrong number for deciding whether to spend more, because extra spend buys new customers.

How long should it take to recover acquisition cost?

That is a cash question rather than a right-answer question. Divide CAC by contribution per order to get the orders needed, then divide by how often a customer buys. At £20.00 CAC, £21.90 contribution and 3.2 orders a year, recovery takes about 3.4 months — and that gap is the hole you have to fund every time you increase spend.

Other guides

• How much does Etsy take from a sale?
• How to build a cash flow forecast
• Agency utilisation rate: the formula, and the denominator that decides it