Local business paid ads that convert

Start with Local Services Ads if your trade is eligible, because the billing model is fundamentally different. Per Google Local Services Help, Google charges when a customer sends a text or email, leaves a voicemail, completes an answered phone call, or submits a booking request. You are also charged when you return a missed call by phone, text or email. You are not charged for clicks. Text, email and booking leads are United States and Canada only.

That difference reorganizes the risk. In a click-billed campaign you buy attention and hope it converts; in a lead-billed campaign you buy the contact itself. Google issues lead credits when a charged lead turns out to be invalid or low quality, which is the mechanism behind the familiar promise that you only pay for good leads. The promise has documented carve-outs: Google states lead credits are not available for health care verticals, tax specialists, or advertisers in EMEA. If you are a dentist, a CPA or outside North America, the refund mechanism you have read about does not apply to you.

Search Ads earn their place second, not never. They are the only format with keyword-level control, they cover trades outside Local Services categories entirely, and they let you bid on the specific problem language customers type at 11pm. Most established local businesses run both. The common mistake is running Search first because it is familiar, then concluding paid ads do not work for local businesses while a cheaper lead-billed channel sat unused above the results.

Local Services Ads are moving into Google Ads proper. Google Ads Help documents that the migration to Performance Max campaigns with pay-per-lead goals began in August 2026, starting with United States home and storefront services: plumbing, HVAC, electrical, appliance repair, house cleaning, lawn care, roofing, pest control and moving. Service-area businesses without a storefront follow in late 2026. Non-U.S. accounts and the remaining categories move in 2027.

Trade coverage of this change has generally reported the timeline and stopped. Four operational details in the same document matter more than the announcement:

1. Your budget number changes shape. Existing weekly budgets are converted to daily averages by dividing by seven. If you have been thinking in weekly terms, the number in your account after migration will look unfamiliar and smaller. It is the same money.
2. Two bidding modes disappear. Manual bidding and vertical-level Target CPA are deprecated. If your account depends on either, you need a replacement plan before the switch, not after.
3. Historical reports do not transfer. Your Local Services reporting history does not come across. Whatever year-over-year comparison you intend to make later has to be exported now.
4. You get 14 days of notice, with a reminder 7 days later. That is the entire window in which the export above is possible, and two weeks is thin warning for an owner who checks the account monthly.

Separately, and quietly, the guarantee changed. Google discontinued the money-back component of the Google Guarantee and renamed the badge Google Verified. Consumers could submit reimbursement requests only for services booked through Local Services Ads before December 7, 2025. Most pages describing this program still promise customers a refund with a dollar coverage limit. That copy is now wrong, and repeating it is a promise you cannot keep. The badge still means something real: it certifies you passed Google’s screening and verification checks, and it still contributes to ad rank.

Rank deserves a note here, because one of its inputs is behavioral rather than financial. Google lists responsiveness among Local Services Ads ranking factors and states explicitly that missed calls may negatively affect it, alongside review rating, number of reviews, average response time, image quality and completed verification checks. On Google’s own account, then, an unanswered phone can cost you twice: the job in front of you, and some of the placement that would have surfaced the next one.

Far more than most published benchmarks suggest, or far less, depending entirely on what you do for a living. Across 3,211 U.S. home-services search campaigns running 1 April 2024 to 21 March 2025, minimum 103 campaigns per category, medians reported to limit outlier effects, LocaliQ’s Home Services Search Advertising Benchmarks reported median cost per lead ranging from $45.15 for pools and spas to $228.15 for roofing and gutters. Plumbing sat at $129.02, air conditioning at $127.74, cleaning at $46.99. We cite that publisher by name and methodology rather than linking it, for reasons a few paragraphs down.

The conversion side varies just as widely. In the same dataset, median conversion rate ran from 2.61% for general construction and contractors to 17.65% for cleaning and maid services, and median cost per click from $5.31 for general construction to $13.74 for painting. Those are different businesses with different arithmetic, and no shared benchmark describes both.

The cross-industry number you have probably seen is $66.69, alongside a $5.42 cost per click, a 6.64% click-through rate and an 8.18% conversion rate. Those come from the same publisher’s 2026 Google Ads benchmarks: 13,474 U.S. search campaigns running 1 April 2025 to 31 March 2026, medians, minimum 52 campaigns per subcategory. It is a defensible dataset for what it is: a national median across every industry that advertises. It is not a target for a roofer, and using it as one will make a viable campaign look like a failure.

Read benchmarks with the publisher in mind

Two honesty notes, because this corner of the internet has a circular sourcing problem. Both datasets above are published by marketing services companies that sell the advertising they are benchmarking. That does not invalidate the numbers, but it is a disclosed interest, and the figures are self-reported client medians rather than an audited census. It is also why we name them instead of linking them: we sell the same service, so treat a benchmark from anyone in this trade, us included, as evidence to check rather than a fact to adopt. And the identical $66.69 figure also appears on a companion page carrying no sample size, no date range and no definition of a lead, which is the version most widely quoted. Look for the denominator and the date window before you plan against any paid-media benchmark.

There is no official minimum, and anyone quoting a universal dollar figure invented it. But two numbers Google publishes let you calculate your own. Google Ads Help recommends measuring Target CPA performance over the last 30 days including at least 30 conversions in that window, because the average Target CPA metric is unreliable over periods with insufficient conversion volume. And Google Local Services Help states that a minimum budget of 10 leads per week is recommended for optimal results with Maximize Leads bidding.

Multiply the first by your trade’s median cost per lead and the floor appears.

A cleaning company at a $46.99 median cost per lead needs roughly $1,410 a month to reach 30 leads. A roofer at $228.15 needs roughly $6,845 for the same 30. Both call themselves a small local service business, and their viable budgets differ by nearly five times. Below your own floor you are still spending real money, but on a sample too small for Google’s own stated evaluation window to mean anything — which is the situation an owner is usually describing when they say they tried Google Ads and could not tell whether it worked.

One more number keeps the monthly math honest. Google states that a campaign may spend up to twice its average daily budget on a given day, but that at the end of the month you will have spent no more than 30.4 times your average daily budget. Plan against the monthly ceiling. The daily figure exists so Google can chase a Tuesday spike in emergency searches on your behalf, and an owner watching the daily spend line will otherwise pause a campaign on the exact day it worked hardest.

Most of the answer arrives by phone, which is precisely the channel default analytics cannot see. The fix is the Google forwarding number: a temporary number Google substitutes for yours when a visitor arrives from an ad, so the call is attributed back to the campaign that produced it. Without it, phone leads are invisible in reporting and your bidding algorithm optimizes toward the form fills it can see while ignoring the calls that pay your bills.

Set a minimum call length threshold, and set it from your own call log rather than a default. A 26-second wrong number should not count as a lead when your threshold is 30 seconds and your real jobs start with a longer conversation. Recheck it quarterly: a threshold that fit one service is wrong once you sell three.

Google’s newer approach goes further than duration. AI-qualified call leads use the call recording as the primary signal to filter out robocalls, spam calls and misdials, falling back to call duration when a call cannot be recorded. The constraint worth knowing before you plan around it: call recording, and therefore the AI evaluation, is currently only available when both the dialing and receiving phone numbers are located in the United States or Canada.

Whatever the tracking layer, answering matters twice over. Google lists responsiveness among Local Services Ads ranking factors and says directly that missed calls may negatively affect it. Google does not publish how much weight that carries, so treat it as a documented input rather than a measured penalty. Either way the job itself is gone. That matters most for construction and trades businesses.

Fast. But the number everyone quotes is the most confidently misattributed statistic in local marketing, and its provenance changes how much weight it deserves. The ubiquitous claim is that calling within five minutes makes you 100 times more likely to reach a lead. That figure originates in a 2007 InsideSales.com executive summary analysing more than 15,000 leads and more than 100,000 call attempts across six companies over three years, which reported that the odds of contacting a lead drop 100-fold between a 5-minute and a 30-minute callback, and the odds of qualifying drop 21-fold. Three corrections to how it is usually cited: it is from 2007, not 2011; it is a vendor-funded study of B2B web leads, not local home services; and it is observational, not experimental. Nobody randomised callback speed. Fast responders and slow responders differ in many ways besides speed, and staffing is one of them.

None of which means answer slowly. The direction of the association is consistent, the mechanism is obvious to anyone who has submitted a form to three contractors on a Saturday, and Google’s ranking documentation treats responsiveness as a factor in its own right. Speed to lead is a well-documented association with an unmeasured causal size. Treat it as a strong operational default rather than a law of physics.

The practical version costs nothing. Route ad calls to a phone a human answers during the hours your ads run, and schedule ads to those hours if you cannot cover all of them. Return missed calls the same hour, remembering that in Local Services Ads a returned missed call is itself a chargeable lead event, so the callback has already been paid for either way.

Less reliably than the industry claims, and the study everyone cites says so plainly if you read past the headline. The Google-commissioned Deloitte Digital and 55 report Milliseconds Make Millions is quoted everywhere for an 8.4% conversion lift from a 0.1-second mobile speed improvement. That figure is retail only: the improvement was associated with an 8.4% increase in retail transactions across 20.5 million sessions and a 10.1% increase in travel conversions across 7.4 million sessions, within a study of 30 million sessions and 37 brands across Europe and the US, published in 2020.

The lead-generation vertical is the nearer analogue to a local service enquiry, and it went the other way. Across 505,000 sessions and 6 brands, mobile conversion rate fell 1.9% with the same speed improvement, while progression from the first form step to form submission rose 21.6%. The report states it without hedging: conversion rates on both mobile and desktop decreased when site speed was seen to improve.

The useful reading is narrower than the slogan. Speed appears to help people who already started your form finish it. It does not appear to persuade anyone to start. So fix the form step before the milliseconds: fewer fields, a phone number that dials on tap, an offer above the fold. Then meet Google’s performance threshold and stop. Per web.dev, the good thresholds are a Largest Contentful Paint of 2.5 seconds or less, an Interaction to Next Paint of 200 milliseconds or less, and a Cumulative Layout Shift of 0.1 or less, each assessed at the 75th percentile of page loads and segmented by mobile and desktop.

On homepage versus dedicated page: a homepage has to serve every visitor, and a paid click has exactly one job. A page built for one service in one service area, phone number and form above the fold, gives the visitor nothing to decide except whether to contact you. That is the entire case for dedicated landing pages, and it needs no speed statistic.

Fewer than the national benchmark implies, and the number is public. The Bureau of Labor Statistics Quarterly Census of Employment and Wages publishes establishment counts by county and industry code, the cleanest way to estimate the density of a local auction. Monterey County averaged 702 private specialty-trade contractor establishments in 2024 under NAICS 238, employing 4,339 people at an average annual pay of $71,911.

Summed across the five Central Coast counties, that is 3,027 specialty trade establishments: San Luis Obispo 721, Santa Barbara 910, Monterey 702, Santa Cruz 532 and San Benito 162. Not all of them advertise, and the ones that do are not all bidding on your service in your radius. But it bounds the problem in a way a national benchmark cannot.

The pattern repeats in adjacent trades. Monterey County averaged 340 private establishments in services to buildings and dwellings, the NAICS 5617 code covering janitorial, landscaping and pest control, and 236 in automotive repair and maintenance, against 43 and 26 respectively in neighbouring San Benito County. Roughly eight and nine times as many firms in two adjacent counties. A Hollister contractor competes in a thinner auction than a Salinas one and should expect a longer wait to accumulate the 30 conversions Google’s evaluation window assumes.

So build local campaigns from local data. Set the floor from your trade’s cost per lead, adjust the timeline for your county’s density, and measure against your own account rather than a national median assembled from thousands of businesses that are not you. The campaign side of that is what our Google Ads and Local Services Ads work covers.