The Business Behind Home Services.

Topics  ·  About  ·  Submit a Tip  ·  Advertise

Google Returns Old LSA Numbers. They Show Spend, Not Profit

Google now lets migrated accounts download historical Local Services Ads totals, but the file holds spend and lead counts only, so the profit math has to come from your own job records.

·

· 9 min read

A small business owner at a desk reviews paper invoices beside a laptop showing a spreadsheet, working out profit from job records.

Google told Local Services Ads advertisers their old performance reports would not follow them into Google Ads. Now it is handing back a small piece of that history. The piece it returns shows what you paid for leads. It does not show what those leads were worth.

Search Engine Roundtable’s Barry Schwartz reported on Nov. 17 that Google updated its help page to say advertisers can download historical LSA reports from the LSA redirection page. Per the help text he quoted, the download covers lead totals: total lead spend and the number of charged leads, split by phone, message and booking. It does not include impressions, clicks or other report data. An earlier version of the same page said the reports would not transition and told advertisers to take screenshots before migrating.

For an owner deciding whether LSAs deserve next year’s budget, the file is useful, but only as one half of a calculation. This piece covers the other half and where the published benchmarks help and fall short.

Leads aren’t jobs, and that’s the problem with any report that stops at the lead count. As an owner, I can tell you a lead report can look excellent while the shop is struggling to turn those leads into profitable work. The calls get counted. Whether someone answered, booked, showed up, sold and got paid lives in a different set of records.

Like this? Get it in your inbox every week.

What changed in home services this week, and what it means for your shop.

No spam. Unsubscribe anytime.

Here’s a real one. In a September reporting period, one of my home service companies spent $5,017.96 on Google Local Services Ads and got 48 charged leads. That works out to $104.54 per lead. Google showed 37 of those as phone leads and 11 as message leads, so on the advertising side we could see exactly what we spent and what Google delivered.

Here’s where owners get into trouble. Those 48 leads aren’t 48 booked appointments. They aren’t 48 sold jobs. And they sure as hell aren’t 48 profitable customers. I still need to know how many calls our office answered, how many customers we booked, how many jobs our techs sold and what we actually collected.

That’s why I don’t judge our marketing by Google’s lead count alone. I want to know what it cost us to acquire a paying customer and how much gross profit that customer generated. Everything before that is just part of the story.

A disclosure: I own home service companies and sell lead generation, so I have a stake in how owners judge paid lead sources. Nothing below argues for or against LSAs. The benchmarks come from outside agencies and data publishers, and each is described as what it is.

What the file is, and what Google hasn’t said about it

“Booking” in Google’s description appears to be a lead type, alongside phone and message. It is not a booking rate. Google’s description doesn’t define it further.

Several things a planner would want are not in Google’s description:

  • How far back the history goes.
  • Whether the rows are daily, weekly or monthly.
  • Whether spend is split by trade, job type, location or campaign.
  • Whether the spend figure is before or after credits for bad leads.

Open the file and settle those four questions before building anything on it. A monthly file with no trade split supports a monthly analysis and one blended number. A file that only gives lifetime totals supports one number.

The download is also tied to the redirection page, which, per Google’s help page, is what accounts see after being moved into Google Ads. Google’s Ads Liaison account has said on X that advertisers can’t access historical performance reports after the move and strongly recommended downloading history before the transition. Whether accounts that haven’t moved yet can reach the download is unclear.

Turning spend into cost per booked job

The file gives you only the spend side. The rest has to come from your own records. Here is what to line up:

FieldWhere it comes fromWhat it’s for
Month (if the file has it)Google LSA history downloadLines spend up with jobs
Total lead spendGoogle LSA history downloadThe cost side
Charged leads by phone, message and bookingGoogle LSA history downloadCost per charged lead
Charge status (charged, not charged, in review, credited)Leads page in Google AdsShows whether credits changed the real cost
Lead date and caller or customer detailsLeads page export, if it includes themMatching leads to your jobs
Booked appointment (yes/no) and dateYour CRM or booking softwareCost per booked job
Job sold, revenue and collected dateYour CRM or job softwareCost per paying customer
Gross profit per job and tradeYour job costingWhether the channel made money, by trade

With those in one place, the steps are below.

  1. Lay out the file by month. Put total lead spend and charged leads side by side. Dividing spend by charged leads gives your cost per charged lead for each month.
  2. Export your leads. The updated help text says you can check, download, or interact with leads on the Leads page in Google Ads. Check the export for fields you can match against your own job records, such as dates and contact details, before relying on it.
  3. Match leads to jobs in your own system. Pull the booked appointments, sold jobs and collected revenue that trace to those leads. Group them by the month the lead arrived, not the month the job closed. A lead in late October may not pay until December.
  4. Divide. Spend divided by booked appointments is cost per booked job. Spend divided by paid jobs is cost per paying customer. Then set that second number against gross profit per job. What’s left has to cover your overhead and still leave the contribution you want from the channel.
  5. Split by trade yourself if the file doesn’t. If you run plumbing and HVAC through one account, tag each matched lead by trade from your own records and allocate spend by lead count. It is an approximation, and you should call it one.
  6. Look at the spread. Compare your best and worst months. No published source reports how much a single shop’s cost swings month to month, so your own file is the only place to find out.

A hypothetical example, Digital Insider’s illustration and not real data: a shop spends $9,000 in a month on 150 charged leads. That is $60 per lead ($9,000 ÷ 150). If 42% book, that is 63 booked appointments (150 × 0.42), or about $143 each ($9,000 ÷ 63). If 31 of those become paid jobs, the cost per paying customer is about $290 ($9,000 ÷ 31).

Illustrative example: In this example, the same $9,000 costs more per result at each later stage
One month in a hypothetical example · $
$0$72.5$145$217.5$290Per charged l…Per booked ap…Per paying cu…
Source: Digital Insider worked example (hypothetical numbers, not reported data). Hypothetical numbers from the article's worked example. This is a made-up example, not measured results. The same $9,000 monthly spend is divided by 150 charged leads, 63 booked appointments and 31 paid jobs. The passage calls the last two figures 'about'.
Show the data
Value ($)
Per charged lead$60
Per booked appointment$143
Per paying customer$290

Now set that against gross profit. If the average paid job carries $310 in gross profit, the ad cost takes all but about $20 per job, roughly $600 for the month. That $20 still has to pay for office payroll, insurance, trucks, owner pay and everything else gross profit doesn’t include. It’s thin, and it may not be worth the spend. If the average job carries $700 in gross profit, about $410 per job is left to cover those costs and leave a return. Same lead cost, very different answer, and no cost-per-lead figure would have told you which shop you are.

What the published benchmarks say, and don’t

Two outside sources publish LSA benchmarks. Both are samples, not the industry, and they define terms differently. These are the figures most useful for checking your own file:

Source and periodFigure, as the source defines it
SearchLight Digital, Feb. 2026 (888 contractors, 1,774 campaigns, $6.72M spend, 126,650 leads)Blended cost per lead $53, before credits. HVAC $51, plumbing $57, electrical $39. Book rate (share of leads that booked an appointment) 43.9%.
The Data-Driven Trades, Jun. 2026Cost per unique lead $62.56. The middle 80% of businesses paid $33 to $88.
The Data-Driven Trades, Jun. 2026Book rate of unique leads 42.0%. Cost to acquire a paying customer (a person in a sold or closed state in the month) $302.33.
Plumbing LSA leads cost the most and electrical the least
Feb. 2026 · $
Plumbing$57HVAC$51Electrical$39
Source: SearchLight Digital (888 contractors, 1,774 campaigns, $6.72M spend, 126,650 leads). These are average cost per lead before any Google credits, so what contractors actually paid may be lower. The blended $53 across all trades is left out because it is an average of these segments.
Show the data
Value ($)
Plumbing$57
HVAC$51
Electrical$39

Unique leads are not charged leads. The Data-Driven Trades strips out duplicates and repeat touches. It said about 43% of raw conversions were duplicates or repeat touches in May. Your file counts charged leads, which Google defines its own way, so comparing your cost per charged lead to $62.56 per unique lead compares two different denominators.

The spread across shops is wide. The top of the June range is about 2.7 times the bottom ($88 ÷ $33), so an average says little about where your shop sits. The same publisher’s cost per unique lead was $55.08 in March, but the samples differ from month to month, so treat that as direction, not a measured rise.

A booked appointment is not a paying customer. At a 42.0% book rate, $62.56 per unique lead works out to about $149 per booked appointment ($62.56 ÷ 0.42, Digital Insider’s math), against $302.33 per paying customer from the same publisher. The paying-customer figure covers people in a sold or closed state during the month, not that month’s leads, and the sources don’t confirm that the lead-cost and paying-customer figures cover the same businesses. The roughly 2-to-1 ratio between the two numbers is Digital Insider’s arithmetic only, and it is not a close rate. The gap is the reason to find your own booked-to-paid numbers.

Credits: the part that doesn’t fit the old playbook

An analysis of the share of leads disputed no longer fits well. Schwartz’s report quotes Google saying manual lead disputes are no longer supported. Google’s help documentation on pay-per-lead campaigns says leads carry a charge status on the Leads page (charged, not charged, in review or credited), and that Google keeps assessing lead quality for 30 days and automatically credits poor-quality, duplicate or invalid leads to the account balance.

Two consequences for the file. Google’s description of it doesn’t mention credits, so you can’t assume the spend figure is net. And because credits can land up to 30 days after a lead, the most recent month’s spend may look higher than it will end up. Don’t judge a month until its 30 days have passed.

No credit rate comes from Google. SearchLight Digital cites “industry data” suggesting about 6% to 7% of LSA spend comes back as credits, and says its own cost-per-lead figures are before credits. BG Collective, an agency that sells LSA management, says roughly 15% to 25% of the bad leads flagged on accounts it manages end up credited. These measure different things (share of spend versus share of flagged leads), come from interested parties, and are not Google statistics. Treat them as a reason to check your own credited leads, not as a benchmark.

October breaks the comparison

Art of Strategy Consulting, reporting Google’s notice to advertisers, says that since Oct. 1 a missed call during business hours can be charged as a valid lead if the caller stays on the line more than 20 seconds, with exceptions. It also reports that Google’s Ads Liaison later said repeat calls from the same user within 15 days are charged only once.

That means charged-lead counts from October on include calls your team didn’t answer, and earlier months did not. It is not yet clear which way this moves your cost per lead or your booked rate, and it will depend on how many calls you miss. Run the analysis in two blocks, before and after Oct. 1, and don’t treat a change across the line as a trend. Migration to Google Ads adds another break, since Google’s help page says the old reports won’t carry over and the new campaigns run on different mechanics.

What to do with it

If your account has already moved, get the file from the redirection page and settle the four unknowns above. Then compare cost per paid job with gross profit per job, by trade where you can. What’s left after acquisition cost has to cover overhead and leave the profit you want from the channel. If it does, the question becomes how much more spend it can absorb; Digital Insider’s earlier piece on break-even lead cost covers how to set that limit. If it doesn’t, the file at least shows which months and trades were the problem, which memory rarely does.

If your account hasn’t moved, the redirection-page download isn’t something Google describes as available to you yet. Google’s Ads Liaison account has urged advertisers to save their history before the transition, so save your existing reports now. Whether the download will still be there for you after you migrate is unclear.

Sources

Written by

Meet the Digital Insider team →

The business behind home services, in your inbox

Your weekly read on the marketing, AI, private equity and tech changing how homeowners hire and how shops get bought.

No spam. Unsubscribe anytime.

Latest

More from Digital Insider

Run a home service business? Get the Insider.

One email a week on the marketing, AI, private equity and tech reshaping the trades. Written by an owner, for owners.

No spam. Unsubscribe anytime.