At the trade's own filed average a bath remodeler spends $29,700 a month, gets 169 leads and sells 17 jobs. The other 152 leads cost about $26,700 and left by one of four doors. Here is how to work out which one, and which to fix first.
Re-Bath's 2026 filing covers 122 units. Spread the filed year over twelve months and the average unit's month looks like this.
| The average unit's month | Filed or derived |
|---|---|
| Spent on marketing | $29,700 |
| Leads it produced | 169 |
| Cost per lead | $176 |
| Jobs sold | 17 |
| Revenue at $18,091 a job | $307,500 |
| Gross profit at 53.5 percent | $164,500 |
| Left after the ad spend | $134,800 |
That is a good month. Now look at it the other way round.
169 leads came in and 17 became jobs. 152 did not. At $176 each, those 152 cost about $26,700.
Ninety percent of the month's ad budget went on leads that did not sell.
That is not a broken account. It is what this trade looks like at its own filed average. So the question is never whether the leads are any good. The question is which step lost the 152.
A month has four steps. A click becomes a lead. A lead becomes a booked estimate. A booked estimate becomes a visit that happened. A visit becomes a sold job.
Your ad account can see the first. Your books can see the last. The two in the middle are the ones nobody records, which is the subject of its own post.
So most owners argue about cost per lead, because it is the only step with a number on it. It is the cheapest step in the month and the one that matters least.
Each step's cost is the month's spend divided by the count that reached it. Spend over leads is your cost per lead. Spend over booked is your cost per booking. Spend over visits that happened is your cost per visit. Spend over sold jobs is the only one that pays your crew.
Then price the leaks. Every lead that dropped out cost you the same $176, wherever it dropped. Three groups. The ones that never booked. The ones that booked and never happened. The ones whose visit did not sell.
Those three plus the leads that became jobs add up to the month's spend, exactly. That is the check that the arithmetic is honest.
Price a lost booking at your cost per booking instead and you count the leads that never booked twice. That makes the middle of your month look worse than it is.
Put your seven numbers in and see where your own month goes
Owners fix the step with the ugliest percentage. That is the wrong instinct, because the steps do not move by the same amount for the same effort.
Move a rate by a realistic step, not by an equal share. Five points on a rate. A tenth on a price.
Five points on a 70 percent show rate gets you 7 percent more jobs. Five points on a 32 percent close rate gets you 16 percent more. The close rate is the uglier number and the better move. Five points is a bigger slice of 32 than of 70.
Anything upstream of the sale moves everything below it. A tenth more leads is a tenth more sales, all else equal. So the ranking is not which number is worst. It is which realistic move is worth the most money this month.
Turn it around and it gets easier to act on.
To cut your cost per sold job by a fifth at today's spend, you need a fifth more sold jobs. Nothing else.
Now ask what each rate would have to reach on its own to produce that fifth, with the others held still. Four answers, four sizes of move.
The one with the smallest move is the one to work on this month. It is usually not the one you were going to pick.
It prices the ad money and nothing else. A visit that never happened also costs you an estimator's hour, a drive and a vehicle. Another page counts those.
It also assumes the month is a month. An ad from six weeks ago pays for a job that signs in week one, so one month on its own reads noisy. Read three.
Get five of the seven numbers off your existing reports: spend, leads, jobs sold, average job value, gross margin.
The other two are booked estimates and estimates that happened. Booked is in your calendar. Happened has to be counted by hand, once, this month.
With all seven you will know which of the four steps loses the most money. You will also be able to say the number out loud.
Because that is what each one cost you. A lead that dies at the booking step cost the same $176 as one that dies at the kitchen table. Pricing the later ones higher feels right and double counts the earlier ones. Then the three groups stop adding up to your spend.
On its own it tells you nothing. Cheaper leads that book less often leave you a higher cost per sold job. That is the number that pays your crew. Run the whole month through and read the last row, not the first.
Partly. Spend, leads and sold jobs give you cost per lead and cost per sold job. They also show how much of the month bought nothing. What they cannot show is which of the two middle steps lost it, and that is where most of the fixable money sits.
Neither, on its own. Fix the one where a realistic move is worth the most money. Five points is a bigger proportional gain on a 32 percent rate than on a 70 percent one, so the ugly number is often the right answer for that reason and not because it is ugly.
A bath job signs weeks after the click. So a change at the top of the month takes a month or more to reach sold jobs. The middle two steps move faster: a booking that happens this week is a visit this week.
Lovro Lucic, on US hours from Zagreb. Your page, your number and your calendar stay in your time zone. Call +1 (650) 643-4497, or pick a time on the calendar.