Define and compute what a customer costs to acquire

1. The definition

The cost to acquire a customer is every cost required to sell a new one. Not the advertising bill alone — the whole apparatus that turns a stranger into a buyer.

What is named as belonging inside it:

• Advertising money itself — the media you buy.
• Payroll for whoever buys that media.
• Payroll for whoever makes the creative.
• Software the advertising and sales teams run on.
• Sales commissions.
• Salaries for salespeople.
• The managers over those teams.

The summarizing instruction is blunter than the list: it is everything that getting customers costs you.

2. Both versions of the number, and which costs dominate where

You may keep the figure in both of the forms below, and keeping both is recommended as a matter of opinion rather than as a rule — it is called wise, not required:

Version	What it counts	Where it applies
Media only	The advertising spend alone	Paid media; it does not carry over to other channels
Fully loaded — the name sometimes used for it	Media plus every wage, commission and tool behind it	Any channel

The split matters because the dominant cost moves with the channel.

How you get customers	Where the cost mostly sits
Paid media	Media spend, with payroll beside it
Outbound	Payroll — the majority of the figure
Content	Payroll for the team that edits and distributes

Bigger businesses generally run several of these at once, which is the reason given for the simple approach that follows: total everything, count the customers, divide.

3. The calculation

Add every dollar spent on winning customers in a period. Divide by the number of new customers that period produced. What comes out is your cost per customer.

Three worked computations are given, one per channel. They are worked so that you can drop your own inputs into the same shape.

Line	Outbound	Content	Paid media
Software	$200	none named	$1,000
Outreach or media payroll	$3,000	$10,000	$4,000
Media bought	none	none	$20,000
Commission per sale	$100	$100	$1,000
New customers in the month	8	10	10
Commissions in total	$800	$1,000	$10,000
Everything spent	$4,000	$11,000	$35,000
Cost per customer	$500	$1,100	$3,500

The content column's payroll is two people at $5,000 a month each. The paid-media column's commission is probably attached to a higher-priced sale, which is why it is ten times the other two.

Each of the three sums checks out against its own inputs, and each divides evenly into its customer count.

The instruction on the line items is to edit them. If you carry sales payroll and the worked columns do not, add it. If a line above does not exist in your business, strike it out.

4. What the figure is not

Not a budget. Stated as a judgment, not a finding: if they say they have a marketing budget of so much a month, they have not yet understood how paid advertising works. Why a budget and this figure pull against each other is §5, phase three.

There is one legitimate way to derive a spending figure from this cost, and it is a constraint calculation rather than a budget. If you know how many sales you can handle, you can work back to the spend: delivery absorbing twenty new customers a month, each costing a hundred dollars to win, puts the monthly spend at $2,000, which across thirty days is about seventy dollars a day.

Figure check on the daily number. Two thousand dollars over thirty days is $66.67. Seventy dollars is a rounding, and it is marked as one — the figure is given as seventy dollars a day roughly. Nothing downstream uses the rounded figure, and the constraint is the point: the ceiling there is delivery capacity, not advertising. Once delivery is widened the number is raised again.

5. The three phases of spending, in order

Spending on advertising passes through three phases, and the order is the content.

Phase	What you are doing	Why it comes here
One	Measure before you spend	Without measurement you cannot tell which spending worked
Two	Spend ahead of the return, and expect losses	Spending precedes return; you have to be out of pocket first
Three	Scale with the return	Once return beats spend, the answer to how much is as much as you can

Phase one. Running advertising with no tracking is compared to betting without knowing the odds, which is probably one of the fastest ways to be cleaned out. Many owners of multi-million-dollar businesses know only their total spend and their total return, with no idea which part produced it.

Phase two. Losing is normal and it is lopsided. In general you lose more often than you win, and the wins are large enough to cover the losses. Across three hundred advertisements, maybe 295 fail, five return more than they cost, and one is a standout worth pushing hard. The 295 and the five come to three hundred.

The worked loss: a hundred dollars into each of ten advertisements is a thousand dollars. Nine return nothing. One returns five hundred dollars on its hundred. You are down five hundred dollars, and you have found a five-to-one performer. The instruction at that exact point is to put ten thousand dollars behind the performer, which at the same rate returns fifty thousand — at which point the original five hundred stops mattering. This is where many people stop instead.

Flag: how much to spend testing one ad. A second reading sets the test budget for each new ad by what a new customer brings in over their first thirty days (put both as cash collected and as profit), not by what they are worth over a lifetime. Let a new ad spend up to twice that figure before you switch it off, as long as it is bringing leads; if not a single lead has come from it by the time it has spent once the figure, switch it off then. In an invented example, with a customer worth $150 over the first thirty days, an ad with leads may spend up to $300, and an ad with none is cut at $150. The reason given: ads left running too long waste money, and ads dropped before they had a fair chance waste more. SOP 275, section 5, prepares to spend up to what a customer is worth in order to learn, and in its section 7 an ad showing good early signs but no sale yet may run to about three times the target cost of a customer. This page does not settle which reading is right.

Phase three. When the return exceeds the spend, the question of how much changes shape: you feed the machine as much as it will take.

6. The levers that move it, in the order given

Lever	What it changes
A better offer	Higher response rate
A better offer	Higher share taking the next step
A better offer	In general, a lower cost per impression, because the offer reads as better
Cheaper labor	Lower payroll inside the figure
Cheaper software	Lower tooling inside the figure

The grouping is stated, and so is the reason for it. The first three are named as the biggest levers. You could pay the sales team or the outreach team less, and you could buy cheaper software, and either could move the figure. The bottom two are visited rarely, because the changes available at the top are large enough that the bottom ones sometimes stop being worth anything — and cutting pay costs morale for no real gain. Cheaper labor is marked as not necessarily related to the offer-side levers at all; it is listed because it sits inside the arithmetic, not because it belongs to the same family.

On the advertising side specifically, the things that lower the figure are better advertisements, better targeting, fewer steps in the process, better selling, split tests on the landing page, cheaper platforms, and media types that get distributed more widely. The count of such things is left open rather than fixed.

The measurement itself is a lever. Beginning to measure a thing at all is named as one of the best ways to improve it.

7. How much room the lever really has

Separate readings are given of how far apart competitors sit on this number, and both carry scope.

Reading	The spread	Its scope
Competitor pairs	Might be $2,500 against $3,500; rarely $6,000	Unless somebody has only just entered the market
Measured across an industry	Almost always within one to three times each other	Businesses running paid advertising without a brand

The second reading was taken across agencies buying small-business customers through one piece of software, and its condition is stated with emphasis: it holds in the absence of a brand. A large brand can drive the cost toward almost nothing, and where that is true the comparison does not hold.

The conclusion drawn from both readings is the same. The cost of winning customers within one trade is often fairly similar, and usually pretty tight for a given kind of buyer, so the difference between a business that is thriving and one that is not sits on the other side of the ratio — in what each customer is worth. That side is SOP 55.

A warning travels with this. Owners commonly diagnose an advertising problem where a business-model problem is what they have — named as one of the biggest mistakes made. And a second misdiagnosis is named beside it: where the right buyers are reaching a sales conversation and not buying, the advertising is working and the selling is not.

8. What this page does not decide for you

• An acceptable figure. No threshold is given here for what this cost may reach in absolute terms. The test it has to pass is a ratio, and that ratio is SOP 55.
• The period to measure over. The worked computations are monthly. No rule is given for shorter or longer windows.
• How to split a shared cost. Where one salesperson serves more than one channel, no allocation method is given.
• What to do about customers won by referral, who arrive without advertising spend against them.
• How to treat a customer who returns after leaving — whether they are a new customer in the denominator. Not established on this page.

9. The checklist

Question	The answer
What the figure counts	Every cost required to sell a new customer
Named cost lines	Media, media payroll, creative payroll, software, commissions, sales salaries, managers
The operation	Total spend on winning customers ÷ new customers
Versions worth keeping	Media only, and fully loaded
Where outbound's cost sits	Payroll, as the majority
Where content's cost sits	Payroll for editing and distribution
Outbound worked answer	$500
Content worked answer	$1,100
Paid-media worked answer	$3,500
The three phases	Measure; spend ahead of the return; scale with it
Advertisements that fail, per three hundred	Maybe 295
Test budget for one ad, second reading	Up to twice a new customer's first-thirty-day value while it brings leads; flagged in §5
The biggest levers	Offer quality — response rate, conversion, impression cost
The levers rarely used	Cheaper labor, cheaper software
Spread between competitors	One to three times, without a brand
Deriving a daily spend	Customers you can serve × cost per customer ÷ days
An acceptable absolute figure	This page gives no figure for it.

10. What this page does not cover

The other half of the ratio — what a customer returns — is SOP 55, and the ratio itself is decided there. How quickly the money comes back is SOP 57, and the thirty-day tests are SOP 58. Costing an engaged lead off payroll, and the learning budget you spend while the figure is still bad, are SOP 42. Choosing the platform is SOP 35, the landing page is SOP 37, testing cadence is SOP 41, and choosing between doing more, doing better and doing something new is SOP 40. Which channel to run at all is SOP 12. Free and discounted entry offers are SOP 59 — Assemble a money model from four prongs. Commission priced off this figure is SOP 47.

Pricing and delivery staffing are not covered on this page.