Where HVAC marketing money actually goes.
Cost per lead is where most budget conversations start and end. It is also the number least connected to whether marketing paid for itself. This is the ledger we use to trace one dollar from ad platform to bank account, and the four leaks that make every channel report look better than the business it describes.
No invented benchmarks here. Cost per lead varies wildly by market, season and trade, and any article quoting you a universal number is guessing. What does not vary is the structure of where money leaks between spend and revenue. That structure is what this framework maps, and you can verify every step with your own books.
Cost per lead is a division of two slippery numbers.
Divide spend by leads and you get a clean-looking figure built from two unclean inputs. "Leads" counts whatever the report wants it to count: form fills, ten-second calls, wrong numbers, tire kickers. "Spend" usually omits the fees, the tools and your own time. A precise ratio of two vague quantities is still vague. It is just confident about it.
Easy to compute, easy to flatter. Redefine what counts as a lead and the number improves without a single extra job sold. It measures the platform's output, not your outcome.
Everything means everything: ad spend, agency fees, software, the labor of answering and chasing. Revenue means money that cleared, not pipeline. This division is harder to compute and impossible to flatter, which is why it is rarely on the report.
Follow one dollar through the whole machine.
Every marketing dollar has to survive six stages before it becomes revenue. Channel reports describe stage one and two. Businesses live or die at stages three through six, where the money leaks in ways no platform dashboard will ever show you.
When we examine a company, we are looking for the stage where the most value dies. That stage, not the channel, is the real cost problem.
Money buys attention
Ads, rankings, profiles, referral goodwill. This is the only stage most reports describe, and the only one platforms are paid to optimize.
Attention becomes inquiries
Calls, forms, messages. Quality varies enormously and the count is easily inflated by loose definitions.
Inquiries get answered, or not
The first leak that never appears on a report. An unanswered call during peak season is spend converted directly into a competitor's job.
Leak: the missed callAnswered inquiries become booked estimates
Speed and intake skill decide this stage. The same inquiry, answered in five minutes or five hours, is two different products with two different close rates.
Leak: slow follow-upEstimates become sold jobs
Quotes that are never followed up die quietly. Most companies have a folder of open estimates worth more than next quarter's ad budget.
Leak: the unworked quoteSold jobs become banked revenue
The only stage that funds payroll. Marketing performance means one thing: what reached this line, divided by everything it took to get here.
Four leaks that flatter every channel report.
These four failures share a property: each one makes the marketing look better while making the business poorer. That is why they persist. Nobody's dashboard is punished by them.
The missed call disappears from the record
A call that rings out during a heat wave never becomes a bad metric. The click was paid for, the call was counted, the revenue went elsewhere. The channel gets credit for a lead. The company gets nothing. Peak season, when your team is busiest, is exactly when this leak runs widest.
Speed decay is invisible in monthly numbers
An inquiry answered within minutes closes at a different rate than the same inquiry answered tomorrow, because the homeowner kept calling down the list. Monthly reports average this away completely. Two companies with identical spend and identical lead counts can book wildly different revenue on response time alone.
Attribution theater misassigns the credit
The buyer was won at the review stage, or by a referral, or by your name being everywhere for years. Then they clicked an ad to reach your site, and the ad claimed the job. Last-click reporting systematically over-rewards whatever sits at the end of the journey and starves whatever did the persuading.
Peak season launders bad economics
In July, demand is so intense that almost any spend produces jobs, and every channel looks brilliant. The honest read of a channel is its shoulder-season performance, when demand is scarce and the machine has to work. Judge marketing by its worst months. The best months would have fed you anyway.
Rented visibility, owned visibility, and what stopping costs.
Beneath every channel debate sits one structural distinction. Some visibility you rent: the moment payment stops, it stops. Some visibility you own: rankings, reviews, brand searches, a maintained profile. It compounds slowly and keeps working after the invoice. Neither is superior. They answer different questions, on different clocks.
| Rented (ads, lead services) | Owned (rankings, reviews, brand) | |
|---|---|---|
| Time to first job | Days. Demand can be bought this week. | Months. It is built, not bought. |
| Cost trajectory | Flat at best. Auctions get more crowded, never less. | Falls over time as the asset compounds. |
| What happens when you stop | Visibility ends the day the budget does. | Keeps producing, decaying slowly if neglected. |
| Who captures the value | The platform, in rising auction prices. | You, in an asset a buyer of your company would pay for. |
| Honest first use | Immediate volume while owned assets mature. | The long game that reduces rented dependency. |
The mature position is both, sequenced: rent demand while you build the asset, then let the asset lower the rent. Companies that treat this as either/or pay for the mistake in whichever direction they chose.
Five questions before you judge any channel.
Ask these in order. If the answer to any of the first four is no, more spend will amplify a leak, not fix a channel.
-
Do you know your answer rate during peak weeks?
Not a feeling. The number. If a meaningful share of calls ring out in July, that is the cheapest revenue you will ever recover, and no channel change touches it.
-
Do you know your minutes to first response?
From inquiry to a human reply. If it is measured in hours, your marketing is subsidizing whichever competitor answers faster.
-
Can you connect spend to booked jobs, not inquiries?
If the report stops at "leads", the most important half of the ledger is dark. Insist on seeing spend against booked and sold work before believing any verdict on a channel.
-
Is anyone working the open estimates?
Count the quotes issued in the last ninety days with no follow-up. Price that pile. It is usually the strongest argument in the building against buying more traffic.
-
Are you judging the channel on its worst months?
Only shoulder-season numbers tell you whether a channel works or whether the weather does. Pull March and October before renewing anything.
A report that cannot tell you what a sold job cost is not a report. It is a receipt.
The rest of the machine.
Find the stage where your money dies.
The Marketing Diagnostic traces this ledger through your actual market: your visibility, your competitors, your paid landscape. $1,000, credited in full to month one of any program. Ad spend always goes from you to the platforms directly, never marked up.