Google Ads for home services: 886,000 searches a month. 747,000 have already decided not to hire anyone!

PPC for a service business fails quietly, because the wasted clicks look like customers.

There is a crowd in home services search that has already decided not to hire anybody, and what it costs you has nothing to do with how big it is. It is cheaper per click than your customers, there is no end to it, and it reaches the budget first. 186,000 people a month type “pressure washer”, and they want a machine off a shelf at the big-box store. 109,000 type “mulch”, by the yard, thrown in the back of a pickup. 114,000 want the ants gone without a technician standing in the kitchen. 108,000 are painting the spare room on Saturday.

It would be manageable if those searches looked different from your customers’. They mostly do not. They carry the same nouns you bid on - pressure washing, mulch, garage door opener, moving boxes - and broad match cheerfully connects the two, because those clicks are cheap and there are an enormous number of them. Behind that sits a slower crowd: 107,000 a month collecting landscaping ideas, months away from letting anyone quote. And a small, stubborn tail of 32,000 pricing up a trade career or looking up what the letters HVAC stand for.

AdPilot: automated Google Ads for a service business, whichever trade it is

AdPilot hooks into the account and works through it the way a consultant would if you were paying one by the hour. Searches that were never a job. Hours when the ads run and the phone rings out. Ads that get clicks and book nothing. Each finding comes back as a card, in the words you would use on site. You approve it or you skip it. Nothing reaches your account until you do.

And you stay in control the whole way:

• You approve every change. The AI proposes; you decide. Every budget, bid, keyword, or ad edit arrives as a plain-language card you approve or skip with one tap. Nothing is ever applied on its own.
• A snapshot before every change. AdPilot saves exactly how things looked before anything is touched, so you always have a clean “before.”
• One-click revert. Don’t like a change once it’s live? Undo it in a single click, straight back to that snapshot.
• A full audit log. Who, what, when, the old value and the new value, every change recorded. Nothing hidden.
• You keep the keys. You connect with one secure Google sign-in, and it only ever makes changes you approve. The account stays yours start to finish - you never transfer ownership, and you can disconnect anytime.

Where home services ad budgets actually leak

Real US search volumes, pulled on August 7, 2026. Every line contains a word home service advertisers bid on, and eleven of the twelve end somewhere other than your invoice. The twelfth is mulch, and whether that is a drain or a job depends entirely on whether you own a dump truck. Twelve rows, 886,000 searches a month.

Sort them and the shape of the problem is obvious. 486,000 are buying the product or the material: the pressure washer, the mulch, the opener, the boxes, the spare remote. 261,000 are doing the job themselves - the ants, the paint, the grout. 107,000 are still planning, and planning is not hiring. 32,000 want a salary or a definition. Those four groups add back to 886,000, which you can check on your phone in about a minute.

None of that says this crowd outnumbers your customers. It probably does not: “plumber near me” on its own runs 390,000 searches a month, and that is one trade out of nine. What the table says is that these searches cost less per click, arrive in far greater numbers per dollar, and look identical to a real job on the screen where you pay for them.

Now read what is missing from every row. No town. No “near me”. No “today”, no “cost”, no “quote”. That absence is the whole method: your buyers name a place or a price or a deadline, and the 886,000 above name a product or a verb. Substitute your own trade’s machine and your own trade’s Saturday job into those query shapes, and you will find your version of these rows in your search terms report this week.

US monthly search volume, Ahrefs, August 7, 2026. 886,000 searches a month, and the only row on it worth bidding on is the mulch, and only if you deliver it.

Negative keywords for home service businesses. Free, no email, no form.

Take it and use it, customer or not. Ten minutes, costs nothing, and in the trades it is the best-paid ten minutes you will spend on the business this week. Nearly every line below is a phrase rather than a bare word, and that is deliberate: in home services your junk searches and your money searches are built out of the same nouns, so blocking a word instead of a phrase is how owners quietly delete their own best traffic. Read the paragraph under the list before you paste. It names the words this list leaves alone on purpose, and why. The reasoning behind phrase over bare word, and how to test a negative before it costs you a call-out, is in the negative keyword guide.

Buying the machine or the material

pressure washer for sale, best pressure washer, electric pressure washer, gas pressure washer, pressure washer parts, garage door opener price, best garage door opener, garage door opener home depot, chamberlain remote, liftmaster remote, opener remote, remote programming, moving boxes, packing boxes, bagged mulch, rubber mulch, mulch calculator, home depot, lowes, menards, harbor freight, tractor supply, for sale, best price on

Doing it themselves this weekend

how to, how do i, diy, do it yourself, tutorial, step by step, youtube video, step by step instructions, owners manual, fix it myself, home remedy, baking soda, vinegar, how to paint, how to clean grout, how to get rid of ants, how to get rid of roaches, how to get rid of mice, how to get rid of wasps

Still planning, months from hiring anyone

landscaping ideas, front yard ideas, backyard ideas, design ideas, patio ideas, paint color ideas, pinterest, houzz photos, mood board, design software, free plans, floor plan app

Renting the thing instead of hiring you

pressure washer rental, rent a pressure washer, truck rental, u-haul, trailer rental, equipment rental, tool rental, buy used, second hand

Free pickup, the dump, and somebody else’s problem

free junk pickup, free mattress removal, free scrap pickup, free appliance pickup, landfill hours, dump hours, transfer station, donation pickup, goodwill drop off, home warranty, insurance covers, landlord responsibility

The career and the paycheck

salary, wage, pay scale, how much do they make, apprenticeship program, trade school, certification exam, license exam, now hiring, job openings, apply online, resume, indeed, ziprecruiter, union local

Definitions and study

stands for, stand for, what does it mean, abbreviation, acronym, meaning of, definition of, wikipedia, diagram, symbols, how does it work

Starting the same business you run

how to start, start a cleaning business, franchise, franchise cost, business plan, license requirements, insurance for contractors, invoice template, pricing guide, marketing for, lead generation

Google Ads for a service business, trade by trade

• Cleaning businesses: one booking can be worth twenty-six. A one-off deep clean and a customer on an every-two-weeks schedule cost you the same click, but the second one pays you twenty-six times a year. If your tracking counts both as “a lead”, Google optimizes toward whichever is easier to win, and that is always the one-off. Bid for the recurring customer, put the schedule in the ad text so one-off shoppers self-select out, and keep move-out and post-construction cleans in their own campaign - they are a different job at a different price with a different urgency.
• Landscapers: your whole year is decided in about eight weeks, and 107,000 people a month are only browsing. “Landscaping ideas” is 90,000 searches a month and “front yard landscaping ideas” another 17,000, and neither is a customer - that is a phone with a Pinterest board open, weeks or months before anyone gets asked for a number. Meanwhile the real demand arrives in a short spring window and again at fall cleanup, and a flat monthly budget spends the same in January as in April. Move the budget with the season, keep design-and-build separate from weekly maintenance, and if you plow snow or clear storm damage, that is its own campaign with its own hours rather than a keyword bolted onto the mowing ads.
• Painters: 108,000 people a month are painting it themselves on Saturday. “How to paint a room” outruns every commercial painting term in the country, so the first thing a painting account needs is the how-to family blocked at campaign level. After that, painting is bought on before-and-after evidence and on whether the crew looks like people you would leave in the house. “Quality workmanship” is a line every competitor copies in ten seconds. Interior and exterior are two businesses with different seasons and different quote sizes, and sharing a campaign lets the cheaper interior clicks eat the exterior budget. Send the click to a page with photographs of finished work near that neighborhood, not to your homepage.
• Moving companies: 80,000 a month want the boxes, not the truck. “Moving boxes” is 80,000 searches a month of people packing themselves, and “truck rental” is the same decision with a bigger price tag. Both of those clicks cost roughly what a real mover click costs, so a mover without those negatives is subsidizing the DIY option out of the ad budget. Then the part nobody warns you about: nobody hires the first mover they click. They collect three or four quotes in an afternoon, so your cost per click only means something held against your win rate. Local and long-distance are separate businesses with separate margins. Ask for the move date on the form - three weeks out is a customer, “sometime this year” is a browser.
• Handyman services: the search term is the whole job. Handyman work has the widest keyword set in home services, because the customer types the task rather than the trade - mount a TV, fix a fence gate, replace a light fixture, hang a door. Some of that list you do profitably and some you would rather decline, and nobody can guess which from the outside. That makes the search terms report matter more here than in any other trade: read it weekly, name the jobs you do not want as negatives, and put your minimum charge in the ad so the fifteen-minute callouts stop ringing. One warning specific to this trade: never block the bare word “jobs”, because “small jobs” is how your best customers describe their punch list.
• HVAC and electrical: two peaks a year, and a vocabulary problem underneath. HVAC demand arrives in a heatwave and again at the first freeze, so a flat monthly budget wastes the shoulder months and runs out on the day the phone will not stop ringing. Its share of the table above is the quiet kind: 21,000 people a month pricing up the career and 11,000 asking what the four letters stand for, all of it cheap and all of it billable to you. Electrical carries the same leak at four times the size, which is why it now has a page of its own: 123,900 searches a month there belong to somebody who wants to become an electrician rather than hire one. Both trades live on answered calls, so check your ad hours against the answered-call log rather than against your intentions.
• Specialty exterior and surface trades: flooring, concrete and epoxy coatings, pressure washing and pool cleaning. These four share one problem, and the table above proves it with the biggest number on the page: “pressure washer” is 186,000 searches a month, and almost every one of those people is comparing machines rather than crews. Epoxy coatings behave the same way - a garage floor installed, or a bucket of resin from a big-box store. Add the retail words before you touch a bid: for sale, kit, rental, home depot, lowes. After that they behave like small construction jobs: high ticket, quote-shopped, one visit, and worth a campaign each, because a weekly pool service and a concrete coating that happens once a decade have nothing in common except your invoice pad.
• Pest control: 114,000 people a month are dealing with the ants themselves. “How to get rid of ants” is the second largest row in the table, and it is the single most expensive habit a pest control account can have, because those clicks are cheap and endless and they will eat Thursday by Tuesday. Real pest control is bought within hours of a sighting, and the searcher names what they saw: bed bugs, termites, roaches, wasps, mice. Those are different jobs at wildly different prices, and one campaign called “pest control” hands the budget to whichever pest is cheapest to click. Termite and bed bug work carries the ticket that pays for the account, so each gets its own ad group, its own ad text and its own bid.
• Junk removal: a lot of that search wants it taken for free. Junk removal splits three ways: people who will pay to have it gone today, people hunting a free pickup or a charity collection, and people looking up the landfill’s opening hours. Only the first group pays you. Block the free-pickup and landfill phrasings by name rather than the word “free” on its own, which would also block “free quote”. Then decide which side of the dumpster question you are on: if a crew loads it, “dumpster rental prices” is somebody who has chosen to load it himself and it belongs in your negatives, but if you drop roll-offs on driveways it is your best search of the week. Put a price marker in the ad, a truckload figure or a minimum load, so the person with one mattress and no budget clicks somebody else’s ad instead of yours.

What one inquiry is worth in your trade

The gap between these trades is the whole reason this page exists, and it shows up first in the click price. All measured, same pull as the table above: “house cleaning services near me” is 55,000 searches a month at about $2.50, “lawn care near me” 107,000 at about $5.00, “hvac repair near me” 90,000 at about $7.00, “pest control near me” 165,000 at about $9.00, “movers near me” 74,000 at about $10.00 and “garage door repair near me” 99,000 at about $15.00. Six times the price from one trade to the next, on the same platform. The job value and the win rate loaded in are a starting point only, so put your own trade's numbers in. Nothing leaves your browser.

The build, in the order it has to happen

Below is the whole job, written out so you can do it yourself with the Google login you already have. We publish it because an owner who works through it ends up with an account that pays for itself, and because you should be able to check the reasoning line by line instead of taking a monthly report on trust.

Budget a weekend for the setup - realistically two evenings and a Saturday. Then thirty to sixty minutes every other week, permanently. Almost everybody manages the Saturday. The every-other-week half hour is where small accounts quietly die.

The examples move between trades on purpose - HVAC, house cleaning, lawn care, painting, moving, appliance repair, pest control, garage doors - because the gap between an emergency trade and a scheduled one is the real subject here. Same software, different account.

Plumbing has its own version of this guide, with plumbing numbers

1. 1

   Do the money math before you open the account

   Two owners sign up on the same Monday with the same $2,000 a month. One cleans houses. One installs garage doors. Build them the same account and one of them is burning money, and from the outside you can't tell which. The garage door owner can pay $60 for a click and still come out ahead. The cleaner can never pay $12 - but she can lose money at $6 for a year without noticing, because her lead looks tiny on the day and only pays out over eighteen months of visits. Same budget, opposite mistakes. He underspends. She overspends and calls it working.

   You don't start with a keyword. You start with what one lead is worth. Average ticket times the share of quotes you actually win. A landscaper who signs one estimate in three on a $2,400 spring cleanup and install is looking at $800 an estimate request. That's revenue, not profit, so if you want to be strict about it use your gross margin instead. Either way, write the number down. Everything further down this page leans on it, and a number you guessed at in week three is worse than one you worked out in week one.

   Then the fork that splits this whole page: emergency work or scheduled work. A no-heat call at 6am in January and an interior repaint booked for April are both home services, and almost nothing about the account is the same. They don't differ the way people assume, either. The emergency click is not automatically the dear one - inside HVAC, a no-heat click and a furnace-replacement click can both land in the twenties. What differs is everything after the click. Emergency converts inside the hour or not at all, and it converts on a human picking up the phone. Scheduled converts three weeks later, after two other quotes, on the strength of your photos and your written price. So you can't judge the two on the same calendar, and you certainly can't run them on one budget. If you do both, that's two campaigns, and you almost certainly start with one.

   Two more things only you know. When does a human actually answer? Ads run around the clock by default, and a pest control ad serving at 1am into a voicemail box is money handed to the next exterminator on the page. And how far will you really drive? A lawn crew that spends forty minutes each way for a $55 mow lost money before the trailer gate came down. Draw the service area for an average job, not for the one big job you'd cross the county for.

   One more thing before you spend a dollar on any of this. For most of the trades on this page - HVAC, cleaning, pest control, garage doors, moving - Google also runs Local Services Ads: the green-checkmark listings that sit above every ad on a phone, bill you per lead instead of per click, and need a background check, your license and your insurance certificate before they'll run. Verification takes a week or three, so start it today and build the Search account while you wait. Where they're available they usually deserve the first budget you have. They don't replace Search, because they only cover certain job types and they run out of inventory, which is why the rest of this page exists.

   Last, the page the ad lands on, and it is not one page. Emergency work and scheduled work need different ones, and if the second doesn't exist, build it before you buy the clicks. Parking that traffic on the homepage just buys you proof that the page is missing. Then pick one town and one service to start. The account can grow later. A starved account can't.

   • Write down the average ticket for the ONE service you'll advertise first, not the average of everything you do.
   • Multiply it by the share of quotes you win. That's what one lead is worth, and every number below leans on it.
   • Decide emergency or scheduled. Both means two budgets, and you probably start with one.
   • Write the hours a human actually answers. Not the hours painted on the van.
   • Draw the service area for an average job, not for the biggest one you'd ever take.
   • Start your Local Services Ads verification today if your trade is eligible. It takes weeks, and it sits above everything you're about to build.
   • A page ready for each campaign, and they are not the same page.
   • Emergency page: the service and the town in the first heading, tap-to-call in the first screenful, and either what a service call costs or the words “price before we start”.
   • Scheduled page: photos of finished jobs, the brands you install, financing if you offer it, how long the crew is in the house, and a form that asks four questions instead of eleven.

   What one lead is worth, five trades (revenue, made up, use your own)

   • House cleaning: $160 a clean, one in three callers books. That's $53 a caller, and if one click in eight rings the phone, a caller costs you $40 to $72 against $5-9 clicks - break-even at the bottom of that range and a loss at the top. But biweekly for a year is 26 cleans and $4,160, so the caller who rings you is really worth about $1,387. Cleaning is the trade where you have to believe in the repeat before you spend a cent.
   • Appliance repair: $190 average ticket, 7 of 10 callers book. $133 a call, and at $8 a click you can pay for sixteen clicks before you're behind.
   • Interior painting: a $3,400 job, one estimate in four signs. $850 per estimate request against $6-12 clicks. The widest gap on this page.
   • Local moving: $1,100 for a two-bed across town, one caller in five books. $220 a call, clicks $8-15.
   • HVAC replacement: an $11,000 system, one quote in four turns into an install. $2,750 a quote request. The clicks run $15-30, more than the emergency terms rather than less, because everybody bidding on them is chasing the same $11,000. That's exactly why the expensive click is the cheap one here.
   • If your close rate is a guess, say so out loud and go count last month's quotes. Ten minutes with the invoices beats a benchmark off a blog.
2. 2

   Sort the searchers into buyers and readers

   The twelve searches in the table above are the loud version of a problem every trade has. The woman typing “pressure washer” at nine on a Sunday night is comparing two machines on a big-box site and will never call a crew, and her click costs you exactly what a real one costs. So before you spend anything, sort every term you're considering into buyer or reader, and be ruthless about it. The readers are always cheaper and always more numerous, and they'll empty the budget before a real job gets a shot at it.

   You read intent straight off the words. A town name, “near me”, “today”, “emergency”, “cost”, “quote”, “replacement”, a brand of appliance, a specific symptom - that's a homeowner with a problem and a wallet. “How to”, “DIY”, “ideas”, “salary”, “training”, “for sale” - that's a reader, a shopper or somebody eight weeks from a decision. Watch which side the price words fall on. “How much does an exterminator cost” is a buyer, plainly. Only “how to” is the enemy. Blocking every question is one of the most expensive overcorrections in small accounts.

   Urgency inside the buyer bucket matters as much as the bucket does. “Wasp nest in wall” and “termite inspection cost” are both real customers and they want completely different things from you. One wants a truck this afternoon. The other wants a number and a date. Same trade, different ad, different page - and if they sit in the same ad group, one of them gets a headline written for the other.

   Then the test that saves the most money. Read each term cold, with nobody standing there explaining what you do, and ask whether it could match a search for something you don't sell. Bare “landscaping” pulls in design software, community college programs and 107,000 people a month collecting ideas. Bare “garage door” is mostly people pricing a $900 door at Home Depot, and pricing it is as far as most of them ever get. A one-word term is never the seed of a small campaign, and no match type rescues it. Add words until only your customer would type it.

   Where do the words come from if you have no data yet? Your own phone. The way a caller describes the problem is the way they typed it fifteen seconds earlier. “My washer won't drain.” “The garage door won't go up and it's making a grinding noise.” Write those down word for word. Forty searches a month with buying intent will out-earn forty thousand without, every month, and the long specific ones are also the cheap ones.

   • Read every term cold. If it could mean a machine, a bag of material, a class or a store, it needs another word before it goes in.
   • Price words are buyer words. Keep “cost”, “how much”, “quote”, “price”. Block “how to”.
   • Three words beats one. Always.
   • Split urgent wording from planning wording now, on paper. They'll be different ad groups in step 3.
   • Type your main term into Google on your phone and look at who's advertising. All national brands and big-box stores means you picked the wrong lane.
   • Steal the words off your own voicemail. Customers type what they say.

   Nine terms, six trades, sorted in ten seconds

   • lawn care service near me - BUY. The cheapest real buyer on this page at $4-8 a click.
   • garage door spring replacement - BUY. Nobody types that out of curiosity, and they need it today.
   • how much does an exterminator cost - BUY. A price question is a buying question. Do not block it.
   • appliance repair near me - BUY at $6-10, and worth it if the phone gets answered inside two rings.
   • landscaping ideas - BLOCK. 90,000 a month, and it's a Pinterest afternoon.
   • garage door opener - don't bid on it bare, and don't negative it either. As a negative it takes “garage door opener repair” and “garage door opener installation” down with it, and those are the jobs. Bid on the qualified versions and block the shopping words instead: home depot, lowes, manual, remote battery, programming, chamberlain remote.
   • how to get rid of ants - BLOCK. 114,000 a month, cheap clicks, and they'll have eaten Thursday by Tuesday.
   • pressure washer - BLOCK the shopping family: for sale, best, electric, gas, rental. 186,000 people a month want the machine. Leave the two words alone if washing is what you sell.
   • mulch delivery - BLOCK if you don't deliver mulch. Obvious, and it's the kind of thing that survives in an account for a year.
3. 3

   Give the account a shape it can learn from

   Open a campaign when you have a separate budget to protect and a separate result to buy. That's the whole test, and it's the reason most small accounts carry four campaigns too many. Owners split by keyword, by match type, by device, by zip code, and every split cuts the call data into slices too thin for Google's bidding to learn anything from. The ad group is a smaller question: one theme, one ad that can answer every word in it without lying, one page that delivers what was asked for.

   Which brings the emergency and scheduled split back, now with numbers. An HVAC company's no-heat repair averages $340 in the till and the customer decides in four minutes. Its system replacement averages $11,000 and the customer decides over eleven days, after three quotes. They cannot share a budget, because the emergency terms run $12-25 a click and would eat the whole thing by lunchtime. They cannot share bids either: a replacement quote request is worth $2,750 to you, while a no-heat call - $340 in the till, booked eight times out of ten - is worth about $270. Ten to one. No single bid is right for both, and whatever you land on is wrong for each of them in opposite directions. And they cannot share a landing page, because the repair page needs a phone number above the fold and the replacement page needs financing, equipment brands and photos of a finished install. Two campaigns. Always.

   Do the arithmetic out loud before you save any budget, because this step decides whether the account was ever going to work. $60 a day for a local mover at about $12 a click is 5 clicks a day, 150 a month. If one in eight of those calls you, that's roughly 19 calls, and at one in five booking, about four moves. Four moves at $1,100 is $4,400 of revenue against $1,824 of spend - and before you celebrate, on a mover's margin (two guys, a truck, fuel, insurance) that $4,400 leaves you somewhere near $1,700. Month one is break-even. Break-even in month one, with nineteen calls and two weeks of search terms to read, is a good result. Month two is where the money is: same budget, better keywords, the waste blocked. What you must not do is call it a loss and switch it off in week three. And the same $60 spread across four campaigns to “test” is one click a day each - thirty days later you have nothing to read at all. One campaign you can read beats four you can't.

   Some budgets are simply the floor, and you should say so before you launch instead of finding out in week five. An interior painter at $45 a day and $9 a click gets 5 clicks a day, and estimates are slow to convert. That account won't tell you anything real for six weeks, and the discipline that pays is not touching it in week two. On the emergency side the floor sits much higher: HVAC at a blended $18 needs $180 a day to buy the ten clicks a day that make a campaign readable quickly. If you can't fund that, run the scheduled side first and come back to emergency once the account is earning.

   Match types: start everything as phrase. Later, when a specific search has actually produced calls, add it again as exact so it can have its own ad and its own page. Leave broad off. Broad only behaves when three things are true at once - conversion tracking that counts real jobs, a conversion-based bid strategy, and a negative list that somebody maintains - and under about thirty conversions a month it'll spend your month proving it can't tell a buyer from a reader. And don't let anyone tell you exact match means exactly that. It hasn't since 2018, and there's no setting to change it.

   Three things on the build screen arrive switched the wrong way, and all three cost money. First, “Include Google Display Network” is ticked when the page loads. Untick it. Every time. It's the most common reason a first campaign spends four hundred dollars in a week and produces no calls, and you can't see it in the search terms report, because Display clicks never appear there. Search partners you can leave on and check after two weeks by segmenting the report by network. Second, open Locations, then the small “Location options” link underneath, and set it to “Presence: people in or regularly in your included locations”. The default is “Presence or interest”, which is how a Toledo furnace company ends up paying for a click out of Phoenix. Third, the bidding. On day one you have no conversion history, so Maximize Conversions has nothing to maximize and will spend two weeks paying whatever it likes to find out. Start on Maximize Clicks with a maximum CPC cap near the number you worked out above - $25 for the HVAC emergency group, $9 for the painter - so nothing can quietly bid forty dollars on your behalf. Once you're clearing thirty conversions in thirty days, move to Maximize Conversions and then leave it completely alone for two weeks while it learns.

   One more shape rule, and it's the one that hides the most damage. Your own business name goes in its own campaign. Brand clicks are cheap and they convert better than anything else you run, so blended into the main campaign they pull the average down and cover up the fact that the work you're actually trying to buy is losing money underneath.

   • One campaign per budget-and-goal. Never one per keyword, per match type, per device, or per zip code.
   • One ad group per theme, and one page per ad group. If one ad can't honestly answer every keyword in the group, split it.
   • Divide the daily budget by the cost per click before you save it. Under 5 clicks a day the campaign can't be read at all, however long you leave it running.
   • Emergency and scheduled work never share a campaign, a bid, or a landing page.
   • Turn the Display Network off before you save the campaign. It's on by default and it never shows up in your search terms report.
   • Set location options to “Presence”, not the default “Presence or interest”. Two clicks, and Google hides them under a link.
   • Start on Maximize Clicks with a CPC cap. Move to conversion bidding at about thirty conversions a month, then don't touch it for two weeks.
   • Everything starts phrase match. Exact is earned by searches that already produced calls.
   • Broad stays off until tracking is real and a human maintains the negative list.
   • Your business name gets its own campaign, and only if somebody else is bidding on it.

   The click arithmetic for three different budgets

   • Local moving, $12 a click, $60/day: 5 clicks a day, 150 a month. About 19 calls, about 4 booked moves, $4,400 of revenue against $1,824 of spend - and roughly $1,700 of that revenue is actually yours.
   • Interior painting, $9 a click, $45/day: 5 clicks a day. That's the floor. Six weeks before the numbers mean anything, and leave it alone in week two.
   • HVAC emergency, $18 blended, $120/day: 6.7 clicks a day. Above the floor, under the ten a day that gives you a fast read.
   • To buy those ten emergency clicks a day at $18, the budget is $180/day - a $5,472 monthly ceiling. Decide that before you launch, not in week three.
   • Google can spend up to twice your daily budget on a single day, but never more than 30.4 times it in a month. $60/day is an $1,824 ceiling, not a promise of exactly $60.
   • The same $60 split across four campaigns is one click a day each. Thirty days, thirty clicks, no verdict.

4. None of this is difficult. It is a long run of small decisions, and each one quietly costs money when it goes the wrong way. The agent makes them for you and shows you every single one before it touches the account.

   Start free - no credit cardFree for 7 days. No card, no call, cancel whenever.
5. 4

   Twelve headlines, and the three that decide the ad

   A responsive search ad takes up to 15 headlines at 30 characters and up to 4 descriptions at 90 characters. Ship 12 headlines and at least two real descriptions. Count the characters yourself, spaces included: 31 characters gets rejected, and Google doesn't round. Only write genuinely different ones, too. A near-duplicate is worse than an empty slot, because it burns a combination Google could have tested and teaches it nothing.

   Here's the part owners skip. Twelve headlines is not what you shipped. Google assembles them three at a time, in an order nobody shows you, so what you actually put into the auction was 220 different trios - 1,320 once you count the orders they can run in - and your card gets charged for whichever ones it feels like serving. Some of those are good. Go find the worst one: take the three that name the job least, read them as a single ad, and see whether a stranger could tell you what's for sale. “Quality You Can Trust” next to “We Go The Extra Mile” next to “Book Online Today” is a complete, finished ad for a dentist, a dog groomer or a used-car lot. It says nothing about anything.

   The rule that fixes it: the subject repeats, the claim never does. Every headline names the trade or the specific job. No two headlines make the same claim about it. “Broken Spring? Fixed Today”, “Garage Door Repair Denver” and “Same Tech, Start To Finish” all say garage doors and duplicate nothing, because one is the symptom, one is the search term and one is the promise. What kills a set is five headlines making the same trust claim with the words shuffled. Licensed, insured, trusted, family-owned and reliable are one claim in five costumes, and it gets one slot.

   Put a real number in the text. A price is the cheapest negative keyword you'll ever write: people who were never going to pay a service call fee read “$129 Service Call” and don't click, and you don't pay for them. If you quote on site instead, say that, in those words. Same effect, no commitment to a number you can't hold.

   The ad changes shape between emergency and scheduled work, and this is where owners get it wrong. An appliance repair ad for a dead fridge sells the hour: today, tonight, one visit, we carry the part. A painting ad sells the absence of risk: a free in-home estimate, a written price, drop cloths and a finish date, photos of the last twelve jobs. Speed words in a painting ad read as pushy. Reassurance words in an emergency ad read as slow. Write the ad to the group's intent, never to the company in general.

   Vary the lengths on purpose. Twelve headlines all sitting at 29 characters is a set Google can only fit two of.

   • Count every headline by hand, spaces included. 31 gets rejected.
   • Twelve headlines, twelve different jobs. Name the trade or the actual job in eleven of them.
   • Read the three that name the job least as one ad. If a stranger can't say what's for sale, rewrite one of them.
   • One trust claim, one slot - and you already used it.
   • Put a real price or a real promise in the copy. It filters out people who were never going to pay.
   • Write two descriptions minimum. You usually get one, sometimes both, so the second has to carry something the first doesn't.
   • Include a few headlines at 18-22 characters. All-30s means only two ever show.

   Four headlines that failed, and their fixes, counted

   • WEAK: Quality You Can Trust (21) - a car dealer could run this tomorrow.
   • FIXED: House Cleaning, $160 Flat (25)
   • WEAK: We Go The Extra Mile (20) - which mile? Nothing in it says anything.
   • FIXED: Written Price, Fixed Date (25)
   • WEAK: Fast, Local, Reliable (21) - three words, one claim, zero subject.
   • FIXED: Broken Spring? Fixed Today (26)
   • WEAK: Book Your Service Today (23) - service of what?
   • FIXED: Washer Not Draining? (20)
   • Three of those four fixes came out four or five characters longer than the line they replaced: 21 to 25, 20 to 25, 21 to 26. That's the whole cost of naming the job, and you were spending those characters on the word “Quality” anyway.
6. 5

   Block the wrong searches before the first impression

   Negatives are not a tidy-up job for later. They load before the campaign serves its first impression, because an unprotected trade campaign starts buying how-to videos, product listings and directions to the supply house inside the first hour, and you don't get that hour back. Put them at campaign level. And put the biggest one in first: every service you don't offer. That's the single largest source of waste in small accounts, and it's the one item nobody else can write for you.

   Be specific, because “things I don't do” runs longer than owners think. A pest control company that handles ants, roaches and rodents but not wildlife, not termites and not fumigation needs raccoon, squirrel, bat, bird, termite, tent, fumigation and the plural of each. A lawn crew that mows and edges but won't touch trees needs tree removal, stump grinding, arborist, irrigation, sprinkler repair. Every one of those is a real search with a real cost, and none of them will ever become a job.

   Then the part that makes owners swear negatives are broken. Regular keywords stretch. Negatives don't. Yours block the exact word you typed and nothing standing next to it - no plural, no synonym. “Job” sails straight past “jobs”. “Rental” never sees “rent”. “Cheap” will sit there and watch “inexpensive” spend all month. Type both forms of every word, then read the list back and circle anything that only shows up once.

   Pick the narrowest form that stops the bleeding. A negative broad blocks any search containing all your words in any order, which is blunt work: “free” also kills “free quote”, and “free quote” is a homeowner about to book you. Negative phrase blocks those words in that order, which is usually what you want. Be careful with “hire”, too. “Hire movers” is a customer with boxes stacked in the hallway. “Moving jobs hiring” is not. Block the phrases and leave the bare word alone.

   The product family deserves its own paragraph, because it is the biggest single thing on this page. Pressure washers, moving boxes, garage door openers, bagged mulch - 486,000 searches a month between the twelve rows in the table, all of them people buying a thing instead of a crew, and all of them clicking at roughly what a real customer clicks at. A mover without “truck rental”, “u-haul” and “moving boxes” in the list is paying to help somebody move themselves. But block the shopping wording, not the noun, or you take your own repair and delivery work down with it.

   • Write out every service you don't do, then add the plural of each. That list is the biggest negative you own.
   • Add both forms of everything. “Job” does not block “jobs” and “rental” does not reach “rent”.
   • Use phrase negatives by default. Broad negatives block any word order and take buyers with them.
   • Never block bare “free”, bare “hire”, bare “repair”, bare “jobs” or bare “cost”. All five are how customers talk.
   • Load them at campaign level before the campaign is enabled, not after the first bill.
   • Re-read the list every time you read search terms. It is never finished.

   The day-one list for a lawn and landscape crew

   • how to, how do i, diy, do it yourself, yourself
   • video, youtube, tutorial, guide, pictures, ideas, landscaping ideas, front yard ideas, design software
   • dirt cheap, free diy, cheapest way to do it yourself
   • Not bare “free”, and not bare “cheap” either. “Free estimate” and “free quote” are two of the best searches a lawn crew or a painter ever gets, and one word kills both - phrase or broad makes no difference to a single word. “Cheap lawn service near me” is still somebody who wants the lawn cut. If price shoppers are the problem, block “cheapest” and put your minimum job size in the ad instead.
   • salary, wages, now hiring, job openings, apprenticeship program, courses, training, certification exam
   • rental, rentals, rent, renting, for sale, used, second hand
   • home depot, lowes, menards, tractor supply, mower parts, replacement parts, mower blades, mower repair, bagged mulch, mulch calculator
   • Everything you don't do: tree removal, stump grinding, arborist, irrigation, sprinkler repair, hardscape, retaining wall, pool
   • Same list for a moving company, different teeth: truck rental, u-haul, moving truck rental, trailer rental, moving boxes, packing boxes, box store, mover jobs, movers hiring. Leave “hire movers” running, that's a customer with boxes in the hallway. And leave storage alone if you offer it - “movers with storage” is one of the better searches you'll see.

7. Everything past this point is the work that never finishes: reading search terms, adding negatives, checking what the budget did overnight. That is exactly the part the agent takes off your hands.

   Start free - no credit cardFree for 7 days. No card, no call, cancel whenever.
8. 6

   Measure the phone, and price a lead

   Most trade accounts measure nothing and then get judged on clicks, which tell you exactly nothing about whether a truck moved. If your work arrives as a phone call, Google will build the tracking for you: it shows a forwarding number in the ad and counts any call over a set length as a lead. Ten minutes of setup, and the phone starts telling you which search paid for it.

   The default length is 60 seconds. That's a judgment somebody at Google made, and you're allowed to overrule it. An appliance repair booking is often done in twenty-five seconds - “my dryer's not heating, can you come Thursday, what's the fee” - and at 60 seconds you'd throw away half your real leads. A pest control inspection that only counts once it's actually scheduled might need 120 seconds. Set it to how your calls really go, then listen to a few and adjust.

   If work arrives as a form instead, you need a conversion on the page people land on after they hit send, which means two snippets of code Google hands you. Two things to check before you go near a web developer. First, auto-tagging. With it off, the click ID never travels with the visitor, so anything that has to find its way back - conversions imported from Analytics, offline conversions you upload when a quote turns into a job, the report that tells you which keyword paid for the booked work - can't get home. Healthy campaigns look dead and get switched off on the strength of it. Second, look for conversions Google already built out of your Analytics and left switched off. Switching those on takes a minute, and they're sitting there in a startling number of accounts.

   Then the number only you have. What's one lead worth? A garage door company installing at $1,400 with one quote in three signing puts $467 in that box. $467. Their number, off their own invoices. Without it, Google treats a $190 dryer repair and a $2,750 replacement quote as the same event and bids the same for both. And be honest about what it can't see: a pest control plan at $420 a year renews for a decade, and Google will only ever know about the first one.

   Last, the hours. Ads run 24/7 unless you say otherwise. Pull the report after two weeks and look at the overnight rows. Nineteen clicks between 11pm and 6am at $8 a click is $152, and if the phone went to voicemail on all nineteen, you paid $152 to introduce nineteen people to whoever answered next. Schedule the ads to the hours a human picks up. The exception proves the rule: an HVAC no-heat campaign at 2am in January is some of the best money in the trade, but only if someone is really awake to take it.

   • Turn auto-tagging on if it's off. One switch, no downside.
   • Look for conversions imported from Analytics and left disabled. Switch them on.
   • Create a call conversion and set the length to how your calls really go, not to Google's 60 seconds.
   • Put a dollar value on a lead: average ticket times close rate. Your numbers, not a benchmark.
   • Set an ad schedule to the hours a human picks up, unless someone really is awake at 3am.
   • Check the overnight rows after two weeks. Clicks with no calls at 2am is the easiest money you'll save all year.

   Numbers to put in the boxes, and one Google will never see

   • Pest control quarterly plan: $105 a visit, four visits a year, and the average customer stays three years - $1,260 before they leave. Half of callers sign, so a call is worth $630. Put $210 in Google's box if you only count year one, and know you're a third of the way to the truth.
   • Moving: $1,100 a job, one caller in five books, so $220 - but a Friday-in-June lead and a Tuesday-in-February lead are not both worth $220, and the schedule report will show you which is which.
   • Landscaping: a $2,400 spring package, one estimate in three signs. $800, and most of them renew next March without costing you another click.
   • Call length by trade: an appliance repair booking is done in 25 seconds, an HVAC replacement quote takes four minutes. One threshold can't serve both, which is one more reason they're separate campaigns.
   • Google counts from the day you switch it on and never fills in the past. You'll know it works when a booked job turns up in the report with a keyword attached to it.
9. 7

   The first month, when it looks broken

   The first forty-eight hours you're not optimizing anything, you're checking for a pulse. Has money left the account, and did the ad clear review? If day two shows $0.00, that isn't a slow start, it's a fault, and it'll be one of four things: the ad got disapproved, the billing method never went through, the CPC cap sits below what the auction costs, or the location radius got drawn around the wrong town. Go find out which. A campaign that physically cannot show looks exactly like one that's warming up, and the keywords always take the blame.

   Day seven you open the search terms report and do exactly one thing: add negatives. Leave bids, budgets and targets alone. Don't harvest yet. Don't draw a conclusion, because week one is noise and a bid strategy that's still learning will punish you for fiddling with it. Day fourteen to thirty is the real read - cost per call by ad group, then the search terms again, this time for waste and for gold. Any search that produced a call goes back in as an exact match keyword, in the ad group it already belongs to, not in a new one. Adding the keyword is the harvest. Building it a private ad group is how you end up with the single-keyword structure step 3 told you not to build, and how a dozen calls a month get split four more ways until none of the piles says anything. A term earns its own group when that one term is producing five or six calls a month on its own. For most of you that happens once, sometime in your second season.

   Season is the trap this page exists to warn you about, because most trade advice pretends it isn't there. A landscaping account judged over four weeks in February is being judged on nothing. Pest control before the first warm spell looks dead, then triples in nine days when the ants come inside. HVAC in a mild October is a flat line, and then the first hard freeze arrives and one week of it outperforms the whole quarter. Judge a seasonal trade against the same weeks last year, or hold your verdict. What you must not do is panic-cut a campaign in the trough and have nothing running when the phone was about to go.

   Then month two, one decision per campaign, said out loud: feed it, fix it, or stop it. A campaign that spent three times your target cost per job over a full month, with working tracking, in season, has finished learning. Say so and act. And know where the real cost sits. The setup is cheap. What's expensive is the two weeks you skip because a rooftop unit job ran over and the crew was short on Friday. The search terms keep arriving whether you read them or not.

   • Day two: is it spending, and is the ad approved? Zero spend means something is broken - go find it.
   • Day seven: search terms only. Add negatives. Touch nothing else.
   • Day fourteen to thirty: cost per call by ad group, and harvest every search that produced a call as an exact keyword in the group it already belongs to.
   • Judge a seasonal trade against the same weeks last year, or don't judge it yet.
   • Month two: feed it, fix it, or stop it. One decision per campaign.
   • Put the every-other-week half hour in the calendar like a job, because it is one.

   Two weeks of search terms, made up, from an appliance repair shop in Wichita

   • “dryer not heating repair near me” - 7 clicks, 2 calls. Add as exact, in the repair group it already belongs to.
   • “samsung fridge repair wichita” - 5 clicks, 2 calls. Add as exact. A brand turning up twice is worth keeping an eye on.
   • “lg washer error code oe” - 11 clicks, 0 calls. Negative phrase “error code”. Your day-one list blocked “how to” and never saw this wording coming.
   • “appliance repairman wanted” - 4 clicks, 0 calls. Negative phrase “wanted”. You had now hiring and job openings on the list. Nobody thinks to add “wanted”.
   • “scratch and dent appliances wichita” - 6 clicks, 0 calls. Negatives “scratch and dent” and “outlet”. You blocked “used” and “for sale”, and neither one reaches this.
   • “dishwasher won't drain reset” - 3 clicks, 0 calls. Negative phrase “reset”. It isn't “how to” and it isn't “diy”, so your list waved it straight through.
   • Every one of those got past a list that already blocked how-to, DIY, salary, training, hiring, used and parts. That's what the every-other-week read is for.
   • And this report is not the whole story either. Low-volume searches sit below a privacy threshold, so the rows never add up to the full spend. Never claim you've accounted for every dollar.

A worked ad: emergency heating and cooling, counted by hand

This is one ad, for one ad group - emergency heating and cooling repair in Toledo - written for an HVAC company that doesn't exist. HVAC is only the example; swap in your own trade's fast job and the counting does not change. Twelve headlines, every one counted by hand, and one description. The card draws the first three; the other nine sit underneath as the pool Google picks from. Google's floor is two descriptions and its ceiling is four, but you rarely see more than one on a phone, so what's below is the stronger of the two. Google also grabs any three of the headlines and serves them in an order nobody chooses, which means each one has to survive being read next to any other two.

• The trade name goes in three or four headlines. Not twelve.. Google bolds the words in your ad that match what a homeowner typed, wherever that headline happens to land, which is the whole reason “Emergency HVAC Repair Toledo” is written in exactly the words somebody with a cold house uses. Four is as far as it should go. HVAC appears in four of these twelve. Six of the other eight name the equipment instead - furnace, AC, heat pump - one leans on the $129, and exactly one carries no subject at all. That spread is also what Google's own guidance asks for: at least three headlines with no keyword in them.
• The weakest trio still sells heating and cooling.. Don't sample three at random, you'll pick three that pass. Find the three that name the job least. Here that's “Price Before We Start”, “$129 Service Call, Flat” and “No After-Hours Furnace Fee”. Read them as one ad: somebody who comes out to your furnace for a flat $129 with no night surcharge. A stranger knows exactly what's for sale, and not one of the three says HVAC. Furnace carries it on its own, which is the point - the acronym isn't what's doing the work. Exactly one headline in twelve has no subject noun at all, “Price Before We Start”, and one is the ceiling.
• Twelve headlines, twelve separate jobs - and the speed claim gets two of them, not five.. The search term, the winter symptom, the 24-hour promise, the summer symptom, the $129, price certainty, 3,400 jobs as proof, the license, one-visit fixes, heat pumps, no after-hours surcharge, and parts on the truck. “Tonight” and “24 hour” are one promise in two costumes, and by the rule three paragraphs up, one promise takes one slot - it gets two here and that's already generous. What's missing is any headline that only makes you sound nice.
• $129 in a headline is cheaper than any negative keyword you'll write.. A price in the ad text is the cheapest filter in the account. People who were never paying a service call fee read it and scroll past, and you don't pay for that click. Use a real number you'll honor. If you'd rather quote on site, say so instead - which is what “Price Before We Start” is there for.
• You write two descriptions. Google shows one.. It won't save the ad with fewer than two, won't take more than four, and then it picks. On a phone you usually see one. On a desktop you can get both, run together as a pair, so write the second to read properly after the first rather than as a rerun of it. The one above carries the emergency, the license and the price, at 80 characters. The second carries the equipment and the money: “Furnace, AC or heat pump. Flat $129 service call, price agreed before we start.” That's 79. A second description that says the first one again in different clothes is a twin Google can pick instead of the good one, on your money.
• Lengths run 19 to 28 characters on purpose.. Google shows up to three headlines, but only when they physically fit. A set where everything sits at 29 or 30 gets squeezed down to two, and you've lost a third of the ad you wrote. “Heat Pumps Repaired” is 19 characters and it still names the job, which is why it gets shown on the days the long ones don't fit.
• This ad would be wrong for the replacement campaign.. Everything in it sells the hour: tonight, 24 hour, one visit, parts on the truck. Somebody pricing an $11,000 system doesn't want a truck tonight. They want three quotes, financing terms, the brands you install and how long the crew will be in the house. Different ad, different page, different budget - which is why they're separate campaigns below.

One worked example: an HVAC company on $185 a day

HVAC is the example because it is the hardest case on this page - two peaks a year, expensive clicks, and a $340 job sharing an account with an $11,000 one. Swap in your own trade's fast job and slow job and the shape does not change. Two working campaigns and a small defensive one. Three ad groups in total, each narrow enough that one ad answers every keyword in it and points at one page - and no narrower, because at this budget a fourth group would see too few calls a month to tell you anything. Every keyword starts as phrase match. Exact is earned later by the searches that produce calls.

Why it's shaped like this. The split has nothing to do with how the keywords look. The two halves of the business behave differently in every way that costs money: a $340 ticket against $11,000, a decision made in four minutes against one made over eleven days and three quotes, a page that needs a phone number above the fold against a page that needs financing and photos of a finished install. On one budget the emergency terms would eat the replacement money before noon, every day, and you'd never watch it happen. Meanwhile the no-heat and no-cooling terms sit together in one ad group rather than two, because they want the same ad and the same page, and splitting a dozen calls a month across two groups leaves neither of them saying anything. The bare symptom terms in that group will pull some DIY with them - “furnace not working how to reset” is the shape of it - which is exactly what the “how to” and “reset” negatives are there to catch. Your own name always sits alone: brand clicks are cheap and they convert several times better than anything else, so blended into the main campaign they drag the average down and cover up the fact that the work you're trying to buy is losing money underneath. That's the most common way a small account looks healthy while it isn't. Before you even enable the brand campaign, search your own name and see whether anyone is advertising on it - if you're already top of the free results and nobody's bidding, you're mostly buying clicks you get for nothing. Maintenance plans and duct cleaning are missing on purpose. They're a different buyer with a different pitch, and they don't earn a campaign of their own until both campaigns above are paying and total spend is past $10,000 a month. At $5,624 you'd be taking the money out of the two that work to fund a third that hasn't proved anything. Read the whole shape in your own trade: a cleaner runs it as recurring against move-out, a mover as local against long distance, a painter as interior against exterior. The names change and the split does not.

Google Ads management for home services, shown as cards you approve one at a time.

What you actually pay a trades marketing agency for is the hands-on part. Splitting the campaigns by the service that actually makes you money, digging through search terms for leaks, setting the service radius honestly, writing ads that get the phone ringing. AdPilot’s agent does that job and explains itself every time.

Examples of what the cards look like - not results from your account

A quiet phone does not wait for the monthly report.

You keep the account, and the decisions stay with the person paying for them.

Most owners never learn how their own Google Ads account works. Nobody explains it, and an agency has little reason to start. AdPilot is arranged the other way. Every card tells you:

• What it changes - for example, pause a keyword that spent $240 and produced no calls.
• Why it helps - the reasoning, not just the instruction.
• What it is based on - the numbers in your own account.

Approve one and a snapshot is saved first. Then a before and after dashboard shows what happened to cost per booked job, calls and spend, in words that do not require a marketing course. After a month or two you know which levers actually keep the crews busy.

AdPilot is new - and we’d rather earn your trust than fake it.

We could plaster this page with five-star quotes and a “trusted by 10,000 businesses” banner. We won’t - because AdPilot is new, and inventing proof would be lying to you. Instead, here’s what actually protects you:

• You risk nothing to try it. 7 days free, no credit card. If it isn’t useful, you walk away having spent nothing.
• Your account is never at risk. Nothing is ever applied without your one-tap approval, a snapshot is saved before every edit, every change has one-click revert, and a full audit log records it. The worst case is a suggestion you don’t like - so you just don’t approve it.
• Guardrails you can’t trip by accident. AdPilot never deletes campaigns and never touches shared budgets. Some doors simply don’t exist.
• Built by an operator, not a VC-funded black box. Software made to give small businesses the Google Ads work an agency charges a retainer for - without agency prices or the loss of control.

Judge us on the work, the math, and the safeguards - on your own numbers, not our adjectives.