If you steer your Local Services Ads budget by raw cost per lead, you are steering by the wrong gauge. Raw CPL is easy to see and easy to compare, which is exactly why it is dangerous: it can make a wasteful account look cheap and an efficient one look expensive. The number that actually reflects whether your budget is working is cost per booked lead—what you pay for a lead that becomes a real, scheduled job.
The problem with raw CPL
You pay per lead in LSAs, not per click—but not every lead is a customer. A large share of raw LSA leads are unbookable: wrong service, wrong area, spam, price-shoppers who never book, or calls that go unanswered. Third-party estimates put unbookable leads near 45%. That means if you look only at raw CPL, roughly half of what you are counting may be leads you could never turn into revenue.
The consequence is that two accounts with the same raw CPL can have completely different economics. Raw CPL tells you what a lead cost. It says nothing about whether that lead could become a job—and that is the whole question.
The math that changes the picture
Cost per booked lead is simply your raw CPL divided by the share of leads that actually book. Watch what it does to two accounts that look identical on the surface:
| Account | Raw CPL | Booking rate | Cost per booked lead |
|---|---|---|---|
| A — "cheap" | $40 | 30% | ~$133 |
| B — "expensive" | $70 | 60% | ~$117 |
By raw CPL, Account A wins in a landslide—$40 versus $70. By the metric that pays the bills, Account B is more efficient: it turns leads into jobs at a lower true cost. Steer by raw CPL and you would "optimize" toward the worse account. This inversion is not a rare edge case; it is common wherever lead quality varies, which is everywhere.
Why this shapes every budget decision
Once you accept cost per booked lead as the gauge, several earlier ideas snap into focus:
- The spend sweet spot. The "knee" where extra budget stops paying off is defined by booked jobs, not leads. Past the knee you buy more leads at a rising price—but if those marginal leads don't book, your cost per booked lead spikes even as raw CPL looks stable.
- Scaling vs. protecting. The signal to scale up is efficient booked-lead cost; the signal to protect is a booked-lead cost climbing above target. Raw CPL can't tell you which situation you're in.
- Target CPL. A target set on raw CPL can starve volume while ignoring quality. A target anchored to booked-job economics keeps the business math honest.
- Benchmarks. The widely quoted ~$53 average CPL (with a roughly $12–$180 range by trade) is a raw number. Your booked-lead cost is what determines whether the spend pencils.
Improving cost per booked lead—two levers
There are only two ways to lower your cost per booked lead: reduce what you pay per lead, or raise the share of leads that book. The second is usually the bigger opportunity.
Raise the booking rate
- Speed-to-lead. Answer fast. A lead that reaches a live person—or an instant automated response after hours—books far more often than one that hits voicemail and calls the next provider.
- Geographic and schedule targeting. Concentrate budget on the zip codes and hours that historically convert, and reduce exposure to areas and times that mostly produce unbookable leads.
- Lead quality feedback. Use the "Rate this lead" survey. It feeds Google's machine-learning system, supports auto-credit for invalid leads (assessed in roughly 72 hours, credited within about 30 days), and steers automated bidding toward better leads over time.
Recover what you shouldn't pay for
Google retired manual disputes around July–August 2024 in favor of the ML auto-credit system. Not everything qualifies—job-type and geo mismatches are excluded, and some verticals like healthcare and tax are outside the program—but third-party estimates put recoverable spend around 6–7%. Consistently rating leads and capturing that credit lowers your effective cost per booked lead. Ignore the survey and you leave that recovery unclaimed.
Tracking it in practice
To manage cost per booked lead you have to connect two things most accounts keep separate: LSA spend and job outcomes. That means following each lead through your pipeline—contacted, quoted, booked—so you know the booking rate behind your CPL. Once you can see it, the right question every week stops being "what did leads cost?" and becomes "what did booked jobs cost, and is that trend improving?"
The takeaway: raw cost per lead is a vanity gauge—it can flatter a wasteful account and punish an efficient one. Cost per booked lead is the metric that reflects whether your budget turns into revenue. Steer by it: raise your booking rate through speed and targeting, recover the credit you're owed, and judge every budget move by its effect on the cost of a real job, not a raw lead.
Frequently asked questions
What is cost per booked lead in LSA?
Cost per booked lead is your raw cost per lead divided by the share of leads that become real, scheduled jobs. Because you pay per lead in LSAs but not every lead is a customer, it reflects whether your budget turns into revenue. A $40 raw CPL with a 30% booking rate costs about $133 per booked job, while a $70 raw CPL at a 60% booking rate costs about $117 — so the cheaper-looking account can be the less efficient one.
Why is raw cost per lead misleading?
Raw CPL only tells you what a lead cost, not whether it could become a job. A large share of raw LSA leads are unbookable — wrong service, wrong area, spam, price-shoppers, or unanswered calls — and third-party estimates put that near 45%. Two accounts with the same raw CPL can have completely different economics, so steering by raw CPL can push you to optimize toward the worse account.
How do I lower my cost per booked lead?
There are two levers: reduce what you pay per lead, or raise the share of leads that book — and the second is usually the bigger opportunity. Answer fast with speed-to-lead and after-hours response, concentrate budget on the zip codes and hours that convert, and use the lead-rating survey to feed Google’s auto-credit system, which assesses invalid leads in roughly 72 hours and credits within about 30 days, recovering an estimated 6–7% of spend.