Brightline's ad bill says $120 a customer. Adding back what else it took says $268 — and on one honest input, a range instead of a number

The moment

Brightline Heating & Air runs six vans out of a yard in Dayton, Ohio — the same eleven people as the price-increase story: six technicians, a dispatcher, an office manager, two apprentices and the owner. Alongside the service calls and the 1,380-household maintenance plan, Brightline runs one more line: full system replacements, sold from paid leads rather than from an existing relationship. The office manager quotes and follows up on these between dispatch and invoicing; a marketing agency runs the ads.

It is the second week of October. The agency's monthly report says $3,000 spent on Google Local Services and Facebook, 25 replacement systems closed, $120 a customer. The pitch riding in on the same report is to double the budget to $6,000 a month — the same $120, the agency says, at twice the volume. The owner is not looking at a bill. The owner is looking at a decision to spend $3,000 more a month, resting on one figure nobody at Brightline had checked.

What they did instead

The $120 came pre-computed, in a report, from someone whose job is that number, so it went unquestioned the way most vendor arithmetic does when it arrives already divided and already rounded. Nobody at Brightline had ever sat down and added to it.

Not the office manager's time. Roughly a fifth of her week goes to quoting and following up on replacement leads; the rest is dispatch, invoicing and ordering parts she would be doing whether or not the phone ever rang about a new system — and that fifth never got separated from her salary in anyone's head, let alone in a spreadsheet next to the ad bill.

Not the $60-per-install commission agreed with her a year ago, to reward closing rather than just quoting. It sits in payroll. Payroll does not talk to the marketing report, so the two numbers have never once been in the same place.

Not the agency's own $500-a-month management fee, billed on a separate line and filed mentally as "the marketing bill" rather than as part of what a customer costs. Not the $200-a-month quoting software. And not the discounts given to close hesitant homeowners — $750 that month, mostly a same-day financing incentive the office manager offers when someone hesitates on the number.

None of that is dishonest. It is simply what "$120 a customer" leaves out, and none of it shows up until someone adds it back on purpose.

The questions a CFO asks first

1. What did the ad bill actually buy, and what else got spent winning the same 25 customers that never arrived as a marketing invoice?
2. Of the office manager's time on these leads, how much is genuinely selling, and how much is dispatch and invoicing she would be doing anyway?
3. Is the per-install commission a real, fixed cost per customer, or a number nobody has actually pinned down?
4. What did discounts given to close these deals add up to, and would the deals have closed without them?
5. If the honest figure is higher than $120, does doubling the ad budget still double the result, or does something else move first?

Question 2 is the one with no clean answer on Brightline's books, because a salary is not a bill with a "selling" line on it the way an ad invoice is.

The arithmetic, shown

Brightline's October figures for the replacement-install line: $3,000 in ad spend, 25 systems closed. Ad spend alone, per customer: $120. That is the number on the agency's report, and it is also the number the whole rest of this arithmetic exists to test.

Scenario one — the office manager's selling time, priced. A fifth of her $3,750-a-month salary, spent on these quotes and follow-ups, is $750. On top of salary: the $60-per-install commission on 25 installs, $1,500. The quoting software, $200. The agency's management fee, $500. Discounts given that month to close hesitant homeowners, $750.

That is $148 more than the $120 on the agency's report — 2.2x it — and none of the $148 sits on any single bill anyone would point to as "the cost of a customer."

Scenario two — the selling time counted as zero, the way the ad report effectively treats it now. Even with none of the $750 selling-time share included, the fully loaded figure is $238 — $118 above ad spend alone, 2.0x it — because the commission, the software, the agency fee and the discounts are real money regardless of what anyone believes about the office manager's hours.

Scenario three — the one honest guess. Back in August, the office manager was out for two weeks and the owner took over quotes personally. 18 systems closed that stretch. Ad spend held at $3,000, commission $1,080 (18 × $60), software and agency fee unchanged, discounts $600. Nobody can say precisely how much of the owner's own time that fortnight was selling rather than everything else an owner does in a day — so the honest input is a range, $500 to $2,000, not a single figure invented to look exact.

At $500, the fully loaded figure is $327 a customer. At $2,000, it is $410. Ad spend alone, that stretch: $167. The gap between $327 and $410 — $160 to $243 above ad spend alone — is not all guesswork: $80 of the difference between the two ends is the guess about the owner's time, and the rest is the commission, software, agency fee and discounts that were never in question either way.

What changed

The ad budget did not double on the $120 figure. The owner recomputed first and found the real range was $238 to $268 for an ordinary month, not $120 — enough of a difference that doubling spend on the strength of the smaller number stopped looking like a settled decision and went back to being an open one, still unresolved at Brightline.

What did change is smaller. The owner now asks the agency to show its $500 management fee and the $200 software cost on the same page as the ad number, rather than filed as a separate line nobody connects to "cost per customer." And looking at the $750 in October discounts line by line turned up something worth watching rather than acting on yet: three of the twenty-five deals accounted for most of it, not an even spread across all twenty-five, which raises a question about those three jobs that a single monthly total had been hiding.

What the tool will not do

It will not say whether $268, or $238, or $410, is worth paying to win a customer. That depends on what a customer is worth over time, which is a different question with its own tool — see Is My Marketing Working? or Unit Economics & Payback for the payback side of it.

It will not average the office manager's low and high guess into one invented middle figure. It shows both ends — $327 and $410 — and how far apart they are, because the swing is information and a fake midpoint would throw it away.

It will not decide whether the office manager's selling time is worth separating from the rest of her role. That is a staffing question, not an arithmetic one, and the tool stops at the number.

It will not treat a genuine $0 on commissions, agency fees, tools or discounts as a warning sign. A business that pays no commission and gives no discount gets a true $0 on those lines, not a flag.

It will not produce a per-customer figure with no customers won that month. Dividing spend by zero customers is not a cost of zero — it is undefined — and the tool says so rather than showing a number that looks like an answer.