Cost Per Booked Job: The Only Number Your Ad Account Answers To

Ask a painter what his marketing costs, and he’ll tell you the monthly bill. Ask him what a booked job costs him from Google, and the room usually goes quiet. Not because he’s careless — because nobody who sends him reports has ever put that number on one.

That’s not an accident. The metrics agencies lead with are the ones that are easiest to make look good. The metric that matters is the one that can’t be dressed up.

The report you probably get now

Somewhere in your inbox is a PDF with impressions, clicks, click-through rate, and a line called “conversions” or “leads.” Every one of those numbers can improve while your business gets worse.

Impressions rise when your ads show for looser searches. Clicks get cheaper when they come from people who were never going to hire anyone. And “leads” — the proudest line on the page — counts every form fill and phone ring, including the guy pricing one accent wall, the homeowner four towns past your service area, and the telemarketer who found your tracking number.

None of those lines know whether a crew ended up on a ladder. That’s the tell: a report that can’t tell you what booked is a report about Google’s performance, not yours.

Why cost per lead rewards the wrong behavior

Cost per lead sounds like accountability. It’s actually an incentive problem wearing a lab coat.

Here’s the move: loosen the targeting, widen the match types, soften the ad copy so more people click, and CPL drops. The chart goes down and to the right. The monthly call goes great. Meanwhile your estimator burns three afternoons a week quoting work that was never real, and your close rate “mysteriously” sags.

The agency didn’t lie. The metric did. CPL pays whoever optimizes it to deliver volume, and junk is the cheapest volume there is. When we take over painter accounts, the pattern is common enough that we check for it first: healthy-looking CPL, hollowed-out booking rate.

Why ROAS is usually fiction for a painting company

Return on ad spend at least aims at the right thing — money out versus money in. The problem is what feeds it.

To compute real ROAS you need revenue tied back to campaigns: this keyword produced this call, which became this estimate, which closed at $6,800. Most painter accounts have none of that plumbing. So the “revenue” side gets estimated — average job value times assumed close rate times counted leads. Three guesses multiplied together, presented to two decimal places.

A confident number built on guesses is worse than no number, because you’ll steer by it.

The number that can’t hide

Cost per booked job is brutally simple: ad spend for the period, divided by jobs that booked. Not clicked. Not “converted.” Booked — a homeowner said yes and a date went on the calendar.

What makes it honest is what it forces you to build:

1. Call tracking that ties every call to the keyword and ad that caused it, with spam flagged out.
2. Estimate logging — which calls turned into scheduled quotes. A call that never becomes an estimate is a cost, and the report should say so out loud.
3. Outcome capture — which estimates signed. This is the only step that needs you, and it’s a two-minute monthly conversation.

Run that loop and something better than reporting happens: the outcomes feed back into Google’s bidding as offline conversions, so the machine starts optimizing toward jobs instead of toward whatever is cheapest to click. The metric doesn’t just measure the account. It improves it.

What a healthy number looks like

It varies by market, service mix, and close rate, so treat any universal benchmark with suspicion. The way to think about it is against your gross profit per job. If your average repaint carries $1,800 in gross profit and a booked job from Google costs you $250–$400 in ad spend — a range we consider typical, not promised — you’re trading one dollar for several, on demand, in a channel you control. If a booked job costs you $1,200 in a market where jobs carry $1,500 of profit, the same channel is quietly eating you.

Same platform. Same ads, even. The difference is whether anyone’s watching the number that settles the argument.

A note on small numbers

One honest caveat, because this metric gets abused too. A painting company booking eight or ten jobs a month from ads is working with small samples, and small samples swing. One $14,000 exterior landing in March and not April doesn’t mean the account collapsed — it means the denominator moved by one.

So read cost per booked job in quarters as well as months, watch the trend rather than worshiping any single figure, and keep the intermediate steps (calls, estimates) on the same page so you can see where a bad month happened. A spend-to-booked pipeline with all its stages visible is nearly impossible to spin. A single month’s headline number, stripped of its stages, can still tell a flattering lie in either direction. The point of the metric is honesty; give it enough data to be honest with you.

Three questions that expose any report

You don’t need to become an analyst. Ask whoever runs your ads:

• “Which keyword booked which job last month?” If the answer involves the word “attribution” and no job names, the plumbing doesn’t exist.
• “How many of last month’s leads became estimates?” Counting stops at “leads” exactly when the numbers after it are embarrassing.
• “What did a booked job cost me, by service line?” Cabinets, exteriors, interiors — they book at different costs. Budgets should follow the winners.

An account wired for cost per booked job answers all three in one page. We’ve written up exactly how that wiring works on our how we measure page, and the tracking stack behind it lives on the call tracking service page.

Clicks don’t cash checks. Make your account answer to the number that does.