ChatGPT Ads Benchmarks for Home Services (2026): Lead Volume, Book Rate and Average Ticket from 37 Contractors
Paid ChatGPT leads to home services contractors on SearchLight more than doubled from June to September 2026, from 39 to 93 a month across 37 contractors, and paid placements now account for two thirds of all ChatGPT-attributed leads on those accounts. The leads are real and they book: 95 of 279 paid ChatGPT customers booked a job over the four months, a 34.1% book rate, and the 37 who have paid so far have an average ticket of $4,843. They do not yet book as well as Google Ads on the same contractors, which ran at 48.8% over the same period. ChatGPT Ads today looks like a growing, high-ticket, lower-booking channel: worth testing with tagging in place, not worth moving Google budget to.
Every number on this page comes from SearchLight attribution data tied to the contractor's CRM, not from ad platform reporting, surveys or OpenAI's bid guidance. Spend is not yet available for ChatGPT Ads in SearchLight, so there is no cost per lead here; what you get instead is what happened after the click.
Volume: paid ChatGPT leads are growing, organic ChatGPT leads are not
SearchLight attribution data, 37 contractors with at least one paid ChatGPT lead, June 1 to September 30, 2026.
The growth is all on the paid side. Organic ChatGPT referrals on these 37 contractors sat between 41 and 48 leads a month all summer, while paid placements went from 39 in June to 93 in September, up 66.7% in July, 18.5% in August and 20.8% in September. Paid went from 49% of ChatGPT-attributed leads in June to 67% in September. Part of that is more contractors turning ads on (13 accounts had a paid ChatGPT lead in June, 19 or 20 in each month since), and part is the accounts already running growing their volume.
Lead quality: paid ChatGPT leads book, but not as well as Google Ads
Book rate is the share of attributed customers who booked a job, measured in the contractor's CRM, so it is the same yardstick across every channel in this table. On the same 37 contractors over the same four months, paid ChatGPT leads booked at 34.1%. Google Ads leads on those contractors booked at 48.8%, and Google Organic at 57.5%.
| Channel (same 37 contractors, Jun to Sep 2026) | Customers | Booked | Book rate |
|---|---|---|---|
| Paid ChatGPT | 279 | 95 | 34.1% |
| All ChatGPT (paid + organic) | 469 | 187 | 39.9% |
| Google Ads | 73,893 | 36,063 | 48.8% |
| Google Business Profile | 61,034 | 25,486 | 41.8% |
| Google Organic | 13,454 | 7,741 | 57.5% |
Month by month, paid ChatGPT book rate ran 38.5% in June, 33.3% in July, 30.0% in August and 36.2% in September, so there is no clear improvement trend yet, and at 39 to 94 customers a month a single busy week moves the number several points. The gap to Google Ads is the headline for anyone deciding budget: a ChatGPT Ads lead today is roughly 70% as likely to book as a Google Ads lead on the same contractor. That is consistent with what you would expect from a placement where the consumer is still mid-research rather than typing "plumber near me."
Revenue: a higher ticket than Google Ads, on a small base
Average ticket here is sold plus closed revenue divided by paying customers, the standard SearchLight formula. Revenue per lead is the same revenue divided by every attributed lead, booked or not, which is the number that matters once you know your cost per lead.
| Channel (same 37 contractors, Jun to Sep 2026) | Paying customers | Revenue (sold + closed) | Average ticket | Revenue per lead |
|---|---|---|---|---|
| Paid ChatGPT | 37 | $179,198 | $4,843 | $654 |
| All ChatGPT (paid + organic) | 79 | $411,757 | $5,212 | $913 |
| Google Ads | 16,355 | $61,931,512 | $3,787 | $893 |
| Google Business Profile | 13,183 | $42,315,816 | $3,210 | $740 |
| Google Organic | 3,898 | $14,233,264 | $3,651 | $1,161 |
With those caveats, the pooled ticket is higher than Google Ads, but the monthly numbers do not agree with each other yet: paid ChatGPT average ticket was below Google Ads in June ($2,264 vs $3,878) and July ($2,435 vs $3,879), then above it in August and September. Call it "at least comparable, possibly higher," and revisit when the base is a few hundred paying customers. The lower book rate eats the pooled advantage on a per-lead basis: paid ChatGPT revenue per lead ($654) trails Google Ads ($893) by 27%, which is roughly the book-rate gap. Whether the channel wins on return depends entirely on what the lead costs, and that is the number no one outside the ad platform can see yet.
Cannibalization: no sign Google Ads is losing leads to ChatGPT Ads
If ChatGPT Ads were pulling demand away from search, Google Ads lead volume on these 37 contractors should have fallen faster than the organic Google channels as paid ChatGPT volume rose. It did the opposite. Google Ads leads fell 7.3% from June to September, less than Google Business Profile (down 6.1%) and well under Google Organic (down 12.6%), while paid ChatGPT leads rose 138%. That is the normal summer-to-fall HVAC slowdown showing up in every channel, not a shift between them.
| Channel (same 37 contractors) | June leads | September leads | Change |
|---|---|---|---|
| Paid ChatGPT | 39 | 93 | +138% |
| Google Ads | 17,664 | 16,371 | -7.3% |
| Google Business Profile | 14,352 | 13,471 | -6.1% |
| Google Organic | 3,252 | 2,843 | -12.6% |
Scale matters here too. At 93 leads in September against 16,371 from Google Ads, paid ChatGPT is 0.6% of the volume of Google Ads on the contractors running both. Even a contractor seeing real growth from ChatGPT Ads is not yet seeing enough of it to move the total.
What to do with this if you run a contractor's marketing
Tag the campaigns before you spend a dollar. The only way a paid ChatGPT lead can be told apart from an organic ChatGPT referral, or from direct traffic when the parameters get stripped, is the advertiser's own tagging: utm_source=chatgpt, utm_medium=cpc, and a campaign name that says what it is. Several of the 37 contractors in this data are identifiable only because their tracking-number pool or CRM campaign literally says "ChatGPT Ads." Contractors whose tags were missing or clipped are not in this data at all, which means the paid numbers above are a floor.
Set expectations at a 30 to 40% book rate, not a Google Ads book rate. Leads from the channel are real jobs at a real ticket, but roughly one in three books today versus one in two from Google Ads on the same business. Judge the channel on cost per booked job once spend is visible, not on cost per lead.
Do not fund it from Google Ads. The volume is a rounding error against search, nothing in four months of data says it is replacing search demand, and the leads that do come through book less often. Treat it as an incremental test with its own budget line, and give it 60 to 90 days, because a single month of 1 to 5 leads tells you nothing.
Watch the ticket. If the higher average ticket holds up as the base grows past a few hundred paying customers, ChatGPT Ads starts to look like an install and replacement channel rather than a service-call channel, and that changes which campaigns belong there.
Methodology and caveats
Source. SearchLight attribution data, pulled live on October 8, 2026, for June 1 through September 30, 2026. Every lead is a call, form or chat matched to a customer record in the contractor's CRM, and booked, paying and revenue figures come from the CRM, not from any ad platform.
What counts as a paid ChatGPT lead. A lead whose attribution channel is ChatGPT (utm_source or referrer naming chatgpt) and whose attribution detail also carries a paid marker: a medium of cpc, ppc, ads or paid, a campaign or tracking name containing "ChatGPT Ads," or a CRM medium of PPC. Organic ChatGPT is every other ChatGPT-channel lead. Paid leads whose UTM parameters were stripped in transit land in Direct or organic ChatGPT and are not counted, so paid volume is understated.
The account set. The 37 contractors (HVAC, plumbing, electrical, mixed-trade) that had at least one paid ChatGPT lead in any of the four months. The Google Ads, Google Business Profile and Google Organic comparisons are for those same 37 accounts, so channel differences are not explained by which contractors happen to be in each group. Between 13 and 20 of the 37 had paid ChatGPT leads in any given month.
Definitions. Book rate is booked customers divided by attributed customers. Average ticket is sold plus closed revenue divided by paying customers. Revenue per lead is the same revenue divided by leads. Pooled figures sum the four monthly buckets; a customer whose activity spans a month boundary can be counted in two buckets, which inflates customer counts slightly (paid ChatGPT: 279 customers against 274 leads) and does not change the rates materially.
What is missing. September revenue is incomplete: jobs booked in late September close in October, so the latest month understates revenue and paying customers in every channel. Spend for ChatGPT Ads is not yet available in SearchLight, so there is no cost per lead, cost per booked job or ROAS on this page; those will be added when the ad platform exposes spend. This page is updated monthly.