CallRadius
Measurement & ROI

From Cost Per Lead to Cost Per Booked Job: The Metric That Matters

May 20, 2026 · CallRadius LSA Institute · 4 min read

Ask most home-service owners how their Google Local Services Ads (LSA) are performing and they will quote a cost per lead. It is the number Google surfaces, it is easy to say, and it feels like performance. It is also the wrong metric — or at least a dangerously incomplete one. Cost per lead can improve while your business gets worse, and it can look expensive while you are printing money. The metric that actually reflects reality is cost per booked job, computed on net spend after credits. This article explains why, and how to make the shift.

Why cost per lead misleads

Cost per lead counts every charged contact equally — the emergency customer ready to book, the robocall, the wrong number, the tire-kicker, the caller wanting a job you do not do. But those leads are worth wildly different amounts. A channel can deliver a low cost per lead precisely because it is delivering cheap junk, and a high cost per lead can come from expensive, high-intent leads that book at a great rate. The number moves for reasons that have nothing to do with whether you are making money.

Recall the structural context: third-party estimates suggest around 45 percent of raw LSA leads are unbookable, and average cost per lead is often cited near 53 dollars with a wide 12-to-180-dollar range by trade and metro. When nearly half your leads cannot become jobs, cost per lead is measuring the wrong denominator.

The two corrections

Turning cost per lead into a metric you can trust takes two adjustments.

1. Change the denominator: booked jobs, not leads

Divide spend by booked jobs, not by leads. This automatically prices in your unbookable share, your conversion skill, and your response speed. A business that answers fast and closes well will show a far better cost per booked job than its cost per lead suggests; a business drowning in mismatches and missed calls will show a far worse one. That gap is the truth cost per lead hides.

2. Change the numerator: net spend, not gross

Use spend after credits. Because the ML auto-credit system returns a modest slice of spend — roughly 6 to 7 percent by third-party estimates, on a lag of up to about 30 days — gross spend overstates your true cost. Net spend, computed on a trailing basis once credits have posted, is the honest numerator.

MetricFormulaWhat it reflects
Cost per lead (gross)Gross spend / all leadsAlmost nothing useful alone
Cost per lead (net)Net spend / all leadsEfficiency, still counts junk
Cost per booked jobNet spend / booked jobsTrue acquisition cost
Return on ad spendJob revenue / net spendWhether the channel pays

What cost per booked job reveals

Once you compute it, the number tells you things cost per lead never could:

Tying leads to outcomes

The catch is that computing cost per booked job requires connecting leads to what happened next. That means tracking each lead through its pipeline — contacted, quoted, booked, completed — and, ideally, back to revenue. For businesses using field-service software, syncing booked and paid outcomes closes the loop. Without that connection, you are stuck with cost per lead by default, measuring effort instead of results.

Using it to make decisions

The payoff of the better metric is better decisions. Budget choices should be judged on whether they produced booked jobs at an acceptable net cost, not on whether they produced cheaper leads. A campaign change that lowers cost per lead but raises cost per booked job is a step backward, however good the headline looks. Grading decisions by their effect on booked revenue — and being willing to reverse the ones that did not deliver — is what keeps an LSA program honest over time.

The bottom line

Cost per lead is a vanity metric dressed as an efficiency metric. It counts junk as if it were opportunity and ignores the credits and conversions that determine your actual economics. Cost per booked job, on net spend, is harder to compute and far less flattering — which is exactly why it is worth the effort. It is the number that tells you the truth about whether LSA is building your business or quietly draining it.

Frequently asked questions

Why is cost per lead a misleading metric for LSA?

Because it counts every charged contact equally — a ready-to-book emergency customer, a robocall, and a wrong number all weigh the same. Since third-party estimates suggest around 45 percent of raw LSA leads are unbookable, a low cost per lead can simply mean cheap junk, while a high one can come from expensive, high-intent leads that book well. The number moves for reasons unrelated to whether you are making money.

How do you calculate cost per booked job?

Divide net spend by the number of jobs you actually booked, not by leads. Booked jobs as the denominator prices in your unbookable share, your conversion skill, and your response speed. Net spend as the numerator accounts for the roughly 6 to 7 percent that Google's ML auto-credit returns on a lag of up to about 30 days.

Why use net spend instead of gross spend?

Google's ML auto-credit system returns a modest slice of spend, roughly 6 to 7 percent by third-party estimates, but it posts on a delay of up to about 30 days. Gross spend therefore overstates your true cost. Net spend, computed on a trailing basis once credits have posted, is the honest numerator for any efficiency metric.

How CallRadius helps. CallRadius ties leads to booked-revenue outcomes and computes net cost per booked job automatically, grading its own budget decisions on whether they produced real jobs — not just cheaper leads. See it live at callradius.io.
CallRadius — autonomous AI for Google Local Services Ads · Total AI Marketing LLC, Scottsdale, AZ · Patent-pending closed-loop optimization (U.S. Provisional 64/063,539).