When a home service company should turn on Google and Meta ads

Google will take your money this afternoon. Meta will too. For a plumber, an HVAC shop, or a roofer, that spend can be the highest-return paid channel in the business, or the fastest way to buy jobs for the next name on the list.

The difference is not creative. It is whether a homeowner who taps the ad can find a real company, take it seriously, and get an answer while they still have the phone in their hand. Ads amplify whatever path you already have. If that path is a slow site and voicemail after 5pm, you are paying to lose faster.

This article is a decision framework, not a pitch to "get aggressive on paid." It covers what Google Local Services Ads, Google Search ads, and Meta ads actually buy, what they cost in recent public benchmarks, the math that tells you if a lead is affordable, and the checklist that should be true before you spend. GoodLeed sits at the end, as the system that tells you when that checklist is green and then runs the campaigns in your own ad accounts.

The three paid products are not interchangeable

Homeowners meet you in three different states of mind. The ad product has to match the state, or you will compare cost-per-lead figures that are not the same unit.

Google Local Services Ads (LSAs) sit above the regular results for many home service searches. You pay per valid lead, not per click: a call or a message that Google counts as related to your business. Google's own help pages say invalid or low-quality leads are not charged, and that charged leads can be credited later if they turn out poor. To run them you need a verified Google Business Profile, and you pass screening that can include license, insurance, and background checks. Google says that process averages three to four weeks after you submit documents. The Google Guaranteed or Google Verified badge is the trust signal. You can pause the ad when you are too busy to take work.

Google Search ads are the familiar text ads on searches like "emergency plumber near me." You pay per click. The person already typed the job. Intent is high. You also pay for the clicks that bounce, the wrong town, and "how to unclog a drain yourself" if your keywords are sloppy. Search ads cover queries and towns LSAs do not, and they can defend your own name when a competitor bids on it.

Meta ads (Facebook and Instagram) interrupt someone who was not searching. A lead here is often a form fill from a scroll, a seasonal offer, or a before-and-after that made them curious. Cost per lead is usually lower than search. Close rate is usually worse, because they were not in the middle of a leak. Treat Meta as a different job: fill a slow week, retarget people who already hit the site, put a crew in front of a neighborhood, sell a tune-up before summer. Do not treat a $40 Meta lead as a cheaper version of a $90 Google lead.

What the public benchmarks actually say

National averages hide your city and your trade. Use them as a planning range, then replace them with your own cost per booked job as soon as you have twenty real leads.

LocaliQ's 2025 Search Ad Benchmarks for Home Services put average cost per lead for the category at $90.92. Cost per lead rose for 69% of home services advertisers in that edition, by an average of 10.51% year over year, faster than search ads overall. Inside the category the spread is violent. In the same report, pools and spas sat near $45 per lead and cleaning near $47. Plumbing and heating sat near $129. Roofing and gutters sat at $228. A roofing lead can still be cheap if the job is $18,000. A $90 drain-cleaning lead is not.

WordStream's 2025 Google Ads benchmarks put Home and Home Improvement at the same $90.92 cost per lead, against $70.11 across all industries. Home services is an expensive category on purpose. The jobs are worth more, and every shop in town is bidding.

LSAs are a different meter. Google does not publish an official cost-per-lead table. One of the larger public samples is SearchLight Digital's February 2026 tracking: $6.72 million in LSA spend across 888 contractors and 126,650 leads, a blended $53 per lead, with a 43.9% book rate (lead to appointment, not necessarily to a finished job) and an average ticket around $1,800 in that sample. HVAC in that set was $51 per lead, plumbing $57, electrical $39. That is one firm's accounts, in one month, not a law of nature. Your metro can run $30 or $90. The useful takeaway is the shape: LSAs often land cheaper per lead than search clicks that have to convert, and you only pay when someone actually reaches out.

On Meta, WordStream's 2025 Facebook Ads Benchmarks (leads campaigns, April 2024 through June 2025) put Home and Home Improvement at $41.26 per lead, $2.23 per click, a 1.94% click-through rate, and a 5.22% conversion rate. Facebook lead costs rose 20.94% year over year in that edition, to $27.66 across all industries. Home services sits well above that all-industry average. Cheaper than Google search per form fill. Not cheaper once you count how many of those fills become a day on the calendar.

If a vendor quotes you a single "home services CPL" without saying LSA, search, or Meta, they are mixing units.

The only math that matters: cost per booked job

Cost per lead is a vendor metric. Cost per booked job is an owner metric.

A simple ceiling:

Maximum you can pay per lead = (gross profit on the average job) × (your close rate on that channel) × a safety factor.

Work it with round numbers. Average HVAC service ticket $1,800, gross profit after tech and materials $900, close rate on LSA appointments 40%. Then $900 × 0.40 = $360. You can pay up to $360 per lead before the channel is unprofitable, in theory. A $53 LSA lead in that world is not the question. The question is whether you actually close 40%, and whether "booked" means a tech in the driveway or a no-show.

Flip it for a cheap job. $220 drain clear, $110 gross, 25% close. Ceiling = $27.50. A $57 plumbing LSA lead loses money unless that call regularly upsells, or your close rate is much higher. This is why "our leads are $50" is a meaningless sentence without the job mix.

Track three numbers per channel, every month: cost per lead, lead-to-booked-job rate, average job value. LocaliQ's search figure and WordStream's Meta figure will not save you if you cannot split those three for your own shop.

Speed is the hidden cost of paid media

A paid lead goes cold faster than a referral. The person who just tapped "Call" is still holding the phone. If it rings out, they tap the next ad.

The InsideSales.com / MIT Lead Response Management study (Oldroyd, 2007) is the source of the five-minute figures that get repeated everywhere. Calling a web lead at five minutes versus thirty minutes: the odds of making contact dropped by 100 times, and the odds of qualifying dropped by 21 times. A later Harvard Business Review article by Oldroyd, McElheran, and Elkington (2011), "The Short Life of Online Sales Leads," looked at a much larger set of online inquiries and found firms that tried to contact a lead within an hour were nearly seven times as likely to qualify it as those that waited even an hour longer, and more than sixty times as likely as those that waited 24 hours.

That is not a contractor-only study, and it is not a promise that a 60-second text books every job. It is strong evidence that paid inquiries expire. If your average callback is this afternoon, you are not "bad at ads." You are buying a list for the shop that answers.

LSAs make this worse and better at once. Worse: Google is sending the same homeowner several providers. Better: you can pause when the board is full, so you stop paying for overflow you will lose anyway.

When you should wait

Do not turn ads on, or scale them, if any of these are true.

• You cannot answer in five minutes, including after 5pm and on Saturday. Missed-call text-back counts. A chatbot that says someone will email you does not.
• The Google Business Profile is empty or wrong. LSAs require a verified profile. Search ads land on a listing and a site. If hours, categories, photos, and the phone number are stale, you are paying to lose the comparison in the map pack.
• The site cannot sell on a phone. Google has measured for years that more than half of mobile visits leave if a page takes longer than three seconds. Search clicks onto a brochure that hides the number are a gift to whoever bid less and loads faster.
• You cannot tell which channel booked the job. One number for the website, the profile, LSAs, search, and Meta means you will keep funding the pretty report. Unique numbers or tracked call recording per channel is the minimum.
• The calendar is already full for weeks. Paying for leads you will quote in October trains the town to call someone else. Pause LSAs. Do not "stay visible" at $90 a click.
• You have not done the unit-economics on paper for the jobs you actually sell. If the only job the ads will attract is the $180 one, and your CPL is $90, stop.

Waiting is not anti-growth. It is refusing to rent demand until you can keep it.

When you should start

Start paid when the path already converts some organic or referral work, and you want more of the same job at a known cost.

A practical "yes":

• Missed calls get a text in under a minute, and someone can book a window the same day.
• The profile is finished enough that a stranger would pick you: categories, recent photos, reviews in volume, matching name and phone.
• You can name last month's jobs by source, even roughly, and you know what a booked job is worth.
• You have crew time this week, not a fantasy of hiring later.
• For LSAs, screening is done or submitted. Do not plan on leads tomorrow. Plan on three to four weeks, then a budget you can pause.

There are honest exceptions that start earlier.

You are new in a city. Organic map-pack presence takes months. LSAs and search can rent visibility while the profile and reviews compound. Only if you can still answer the phone. A new shop with voicemail is the most expensive way to enter a market.

Demand spiked and you have capacity. Heat wave, freeze, storm. Search and LSAs put you in front of high-intent queries within hours. This is the textbook case for paid. Turn it down when the weather breaks.

You are not in the three-pack for the jobs that pay, and the site already converts the traffic you have. Paid covers the gap while local search work catches up. Do both. Do not use ads as an excuse to skip the profile.

What to turn on first, and what to add later

First, Local Services Ads, if your trade and town offer them. Pay per lead, badge, high intent, pause when busy. Submit screening as soon as the profile is real, even if you will keep the budget at zero until response is wired. Dispute junk. Watch book rate, not vanity impressions.

Second, Google Search ads for the queries LSAs miss: a service you sell that has no LSA category, a neighboring town, your brand name, a high-ticket job worth a $129 click. Use a landing page that matches the search (the job and the town in the first line, number in the header). Track calls as conversions, not only form fills. LocaliQ's category conversion rates sit in the high single digits to around 10%, depending on the cut of the data. If you are far below that, the account is not the first place to look. The page and the follow-up are.

Third, Meta, after Google is producing booked jobs you can see. Use it to retarget site visitors, to push a seasonal tune-up, to geo-fence the neighborhoods you actually drive, and for visual trades where a photo of the work is the argument. Instant forms will look cheap and close worse. Send serious campaigns to a page that can ring, and answer those leads with the same speed as Google. If your entire paid budget is small, put it in Google until that machine is boringly profitable. Meta's algorithm wants volume. Starving it with $15 a day teaches it nothing and still bills you for the lesson.

A sane split once all three are live is not a universal percentage. It is: LSAs and search take the high-intent budget until you know cost per booked job. Meta gets a test budget with a kill date. Whatever cannot show a booked job in 30 days gets cut, including the channel a salesperson likes.

How GoodLeed decides when ads are right, then runs them

GoodLeed does not lead with ads. The Growth Engine is the path a homeowner has to complete: get found, get chosen, get booked. The first 30 days are the site, local search, Google Business, reviews, weekly content, and a 60-second response with missed-call text-back. Rankings and reviews start compounding. The live dashboard ties leads to sources and to booked revenue. That is how you see whether organic is converting before anyone bids.

When that path converts, paid traffic is the accelerator, not the product. Campaigns are geo-targeted to the searches that become jobs. They are managed by people who do this for home service companies, not by a generalist who also "does social." Spend sits in your own Google and Meta accounts. GoodLeed never marks it up. If a campaign cannot survive the dashboard, it does not survive the month.

That sequence is the point. The scorecard grades Found, Chosen, and Booked in about a minute, including the public half of "would an ad click land somewhere usable." Response is marked estimated on purpose: a crawl cannot see who picks up at 6:42. The Deep Audit times that live. Ads wait until those answers are ones you would spend money on.

If you want the short version of whether you are ready to buy traffic: run the Local Growth Scorecard. If Found and Chosen are leaking, fix those first. If Booked is the hole, fix the answer before you rent more rings. When the engine is converting, GoodLeed is built to turn paid on in your accounts, against booked work, with the same team already on the hook for the rest of the path.

Google and Meta ads FAQs

How much should I budget to start?

Enough to buy a statistically useful handful of leads in your trade, not enough to hide a broken path. Using the public ranges: if LSAs in your world look like $50 a lead, $1,500 to $2,000 in a month is on the order of 30 to 40 leads, which is enough to see a book rate. Search at $90 a lead needs more dollars for the same sample. If you cannot attribute a booked job at the end of that first month, do not scale. Cut or fix response, the page, and the profile.

LSAs or Search ads if I can only afford one?

LSAs, if you can get approved and you can answer. You pay for the conversation, not the click, and you can pause when the board is full. Use Search when LSAs do not cover the job, the town, or the query, or when you need to protect your name.

Why are Meta leads cheaper but often a worse buy?

WordStream's 2025 Facebook leads benchmark put Home and Home Improvement at $41.26 per lead, versus LocaliQ's $90.92 for home services search. The Meta person was scrolling. The Google person typed the problem. Cheaper inquiries with lower intent need faster follow-up and a tighter offer, or they become a list you paid to babysit.

Should a brand-new shop skip ads until they rank organically?

Not always. Organic in a new city is slow. Ads can rent the gap if, and only if, you can answer and you know what a job is worth. A new shop with no reviews, no profile, and voicemail should not buy traffic. Finish the storefront, wire the phone, submit LSA screening, then spend.

How do I know GoodLeed will not just turn ads on to look busy?

Ads are not month one. Spend stays in your accounts and is not marked up, so nobody at GoodLeed earns more because you bid more. The dashboard has to show booked work. That is a different incentive than a retainer that needs a campaign "on" to justify itself.