Customer Acquisition Cost: Count Hours, Not Just Cash

The standard formula for customer acquisition cost is total acquisition spend divided by new customers. Run it on an ordinary month in which you added 12 customers and spent nothing on advertising. Zero divided by 12 is zero, and the formula reports that each customer was free.

That answer is not a rounding problem or a sign you did the arithmetic wrong. It is the formula working exactly as designed on a business it was never designed for. Before you can get a useful acquisition number, you have to settle a question almost no guide asks: which unit is your acquisition actually denominated in? For a growing share of small businesses the answer stopped being dollars some time ago, and nobody moved the formula.

The two conditions the formula needs

Customer acquisition cost is sound arithmetic under two conditions, and both are worth stating because most explanations skip to the division.

The first is that you buy customers. There is a spend line somewhere: ads, a sponsorship, a paid listing, a commission on a booking. The second is that the spend is separable, meaning you can look at a dollar and say it went to winning work rather than to doing it, or to rent, or to payroll.

Where both hold, the number is genuinely decision-grade. You can raise the spend and watch what the cost per customer does. Where neither holds, the formula does not break loudly. It returns something small and tidy, and a small tidy number does more damage than an error message, because a number invites a decision and an error does not.

What changed, and why the number is deflating

The formula was built for a world in which reaching people cost money. Two findings from the Federal Reserve’s 2026 Report on Employer Firms, published March 3, 2026 from a 2025 Small Business Credit Survey that drew 6,525 responses, describe a different one.

Reaching customers and growing sales was the most commonly reported operational challenge, ahead of hiring or retaining qualified staff (Federal Reserve, 2026 Report on Employer Firms). And among the 46 percent of firms whose business or employees now use AI, the most common task by a wide margin is writing or marketing at 83 percent, ahead of individual productivity at 61 percent and planning or analysis at 51 percent (Federal Reserve, 2026 Report on Employer Firms).

Put those side by side. The single most common thing small firms currently point AI at is producing marketing material, which happens to be the part of acquisition that used to arrive as an invoice. Copy, images and posts that once meant paying someone now come out of a subscription bought for other reasons. The cash in the numerator falls.

What has not fallen is the part that actually converts: the follow-up call, the quote that gets chased, the conversation where somebody decides to trust you. That work was never in the numerator. So the dollar figure drifts down while the job itself gets harder, and a measurement that improves while the underlying task worsens has stopped tracking the task.

Two questions that tell you which unit you are in

Answer both about your last 20 customers, using your inbox and booking record rather than memory.

One: could you have got those customers without spending the money? If you switched the ads off, would most of them still have arrived through referral, repeat business or walk-in?

Two: could you have got them without spending the hours? If you had stopped quoting, chasing and asking for a month, would the ad spend have delivered them anyway?

If the money is the thing you cannot remove, you are in the dollar case and the standard formula is the right instrument. Use it as written, and put your effort into the payback question: how many months of that customer’s business it takes to repay what you paid to get them.

If the hours are the thing you cannot remove, you are in the hours case, and no amount of care with the standard formula will produce a usable number, because the dominant input is missing from it. The rest of this is for you.

Converting hours into a number you can use

The correction costs nothing and needs no software. Three steps.

1. Tally two weeks of winning-work time. That means quoting and estimating, chasing quotes and inquiries that went quiet, asking for referrals and reviews, writing or posting anything promotional, and networking. Delivering work you were already hired for does not count, even when it is the thing that earns the next referral. A note on your phone is enough, and rough beats abandoned.
2. Price an hour with your own figures rather than a national average, using one of three methods: your draw, meaning what you paid yourself last month divided by the hours you actually worked; your billable rate, if you bill hourly, which is what the hour earns when pointed at paid work instead; or replacement cost, the hourly rate you already pay, or would have to pay, somebody capable of that specific task. Any of the three works. Switching between them from month to month does not, because the comparison across months is the entire point of building the number.
3. Add any real cash spend and divide by new customers. The cash line stays in; you are widening the numerator, not replacing it.

With placeholder figures to replace with yours: six hours a week of winning-work time is about 26 hours a month, which at 40 dollars an hour is 1,040 dollars. Add the cash spend, nothing in this case. Divided across 12 new customers, acquisition costs about 87 dollars each rather than zero.

Eighty-seven dollars is actionable in a way zero never is. You can hold it against what a customer is worth to you across a year, and against what a paid channel would quote to deliver the same 12. That is the same instinct behind choosing a growth metric you can already count: the number worth having is usually already in the building rather than waiting behind a subscription.

Two practical notes. Keep the tally running for a second two-week stretch in a different part of the year if your trade is seasonal, because a January number and a June number can differ enough to reverse a decision. And write the figure down somewhere with its date and the method you used, since a number whose method you cannot reconstruct in six months is not a series, just a memory.

If you want the tally to outlive the first two weeks

A business adding 12 customers a month does not need a system to hold 12 rows, and paper is a legitimate final answer. If you would rather it persisted, two time trackers cover it, both prices taken from their own pricing pages today:

• Clockify is free for up to 5 users, including unlimited tracking, timesheets and reports. Paid tiers begin at Basic, 3.99 dollars per seat per month billed annually or 4.99 billed monthly, then Standard at 5.49 annually or 6.99 monthly.
• Toggl Track has a free tier covering a limited number of users, with Starter at 9 dollars per user per month.

For most owner-operators the free tier of either is the whole answer. Before adding anything, check what you already pay for: a “how did you hear about us” field on an existing booking form or point of sale does the attribution half of this job at no extra cost. And since an hour spent winning work has to be recovered in what you charge, the figure you land on feeds directly into pricing, which is the argument a standard costing system makes at length.

The argument this leads to

Measurement ends at the previous heading. This part is argument, and it should be read as one.

A cash-only acquisition number does not merely understate the cost. It actively recommends the wrong plan. If the unpaid channels appear to cost nothing, then doing more of them is the rational move, and doing more of them means more owner hours, which is the single input that cannot scale. Owners who over-commit to referral chasing and weekend posting are not being undisciplined. They are following the instrument, and the instrument has a blind spot precisely where their week goes. Getting the number right is what makes it possible to argue for spending money instead, or for leaving a channel alone.

That is also the honest case for pointing AI at this work, and the Federal Reserve’s figures are more specific than the usual claims. Among firms using AI, the vast majority reported no change in their labor costs, while 71 percent said it led to increased productivity, 39 percent noted improved quality of goods and services, and 31 percent reported higher sales (Federal Reserve, 2026 Report on Employer Firms). Same payroll, more output. The hours an AI drafting tool gives back on the writing are worth having because they can be moved to the conversations that actually close, which is the half of acquisition no tool has yet shown it can do. Anyone selling the other reading, that the number exists so you can work out whose hours to stop paying for, has misread both the arithmetic and the survey.

Where does the hours figure send you? Usually toward whichever call on your week has the weakest claim on it, which is the decision the urgent versus important priority rule exists to settle.

Where the prompt fits, and who wrote it for whom

If the two questions above leave you with a channel mix you do not like and no obvious alternative, BusinessPrompter’s Customer Acquisition Loop prompt exists to generate candidates. Its stated purpose is designing customer acquisition loops that scale efficiently, and what it promises to hand back is a blueprint naming channels, metrics and optimization strategies.

Two things to know first. It is filed under Innovation and Growth, and its own “Who It’s Best For” line names startups and growth-stage companies building cost-effective acquisition engines, which is not the same reader as the owner of an established local business. And it sits behind an upgrade wall as a Pro prompt, along with its five-step chain, rather than among the free ones. Use it to widen the list, then run the hours arithmetic on whatever it proposes, because a loop that runs on your attention has a price too. The wider catalog is at BusinessPrompter.com.

Questions owners ask about this

What is the standard customer acquisition cost formula?
Total acquisition spend divided by the number of new customers acquired in the same period. The arithmetic is sound, but it only counts cash, so a business acquiring through referral and owner effort gets a figure near zero that hides where the cost actually went.

What counts as winning-work time when I tally hours?
Quoting and estimating, chasing quotes and inquiries that went quiet, asking for referrals and reviews, writing or posting anything promotional, and networking. Delivering work you have already been paid for does not count, even though it is often what earns the next referral.

How do I put a dollar value on my own hours?
Pick one of three and stay with it: your monthly draw divided by hours actually worked, your billable rate if you bill hourly, or the hourly rate you would pay somebody capable of that task. Each is defensible on its own. Switching between them breaks the month-to-month comparison the number exists for.

Do I need software to track this?
No. A note on your phone covers a two-week tally, and 12 new customers is not a data problem. If you want it to persist, Clockify is free for up to 5 users and Toggl Track has a free tier, and an existing booking system or point of sale can usually record how a customer found you without a new subscription.

Open last week’s calendar and mark every block that was about winning work rather than doing it. Before you price anything or buy anything, that total answers the question the formula could not: whether you are running a business that buys its customers or one that pays for them in hours. Most owners already suspect which, and two weeks of notes simply settles it.