Pull "bathroom remodel" in Keyword Planner for our example market, Raleigh-Durham, and the top-of-page range reads $14.65 to $60.00 a click. For "home addition contractor," the same range runs $3.28 to $13.43.
Now look at the jobs behind those searches. Zonda's 2025 Cost vs. Value report puts a midrange bathroom remodel at $26,138. Its midrange additions run from $60,645 for a bathroom addition to $170,517 for a primary suite. The bigger job has the cheaper click.
That one fact breaks the way most contractor accounts get set up: one bid, typed in once, applied to everything. Set it high enough to show on bathrooms and you overpay on every addition search. Set it for additions and your bathroom ads never make the first page. The pattern isn't a local quirk, either. Nationally, Keyword Planner shows "bathroom remodel" at $9.04 to $52.17 and "home addition contractor" at $4.59 to $28.14.
This chapter is about setting bids by hand, one project at a time, reading the columns that tell you whether those bids are working, and turning what you learn into the number Google needs before broad match takes over.
Bathroom Searches Top Out Near $60 a Click and Addition Searches Near $13. One Account-Wide Bid Fits Neither.
The campaign we launch first is the one Chapter 6 called Isolation (our name, not a Google setting): every keyword on exact match, every bid set by a person. That second part is manual cost-per-click bidding. You write down a ceiling per click. Google usually charges less than that ceiling: per its manual CPC page, the charge is whatever clears the minimum Ad Rank and edges out the next advertiser down.
Google lets you set that maximum in two places. One ceiling can sit on the whole ad group, so every bathroom keyword inherits it. Or a single keyword can carry its own, and that number wins for that keyword. We set bids at the ad group level, one per project, using the project ad groups Chapter 5 built: baths, kitchens, additions, basements, whole-home, outdoor living. A keyword gets its own bid only when its numbers earn one.
Each ad group's opening bid comes from two numbers you already have:
- The market's range. Google builds Keyword Planner's two top-of-page numbers from history: roughly the 20th and the 80th percentile of what advertisers paid to show up there. It tells you roughly where the auction sits.
- Your ceiling. Chapter 3 worked backward from the job to the most a click can cost. With its illustrative rates, the $26,138 bathroom could afford about $3.07 a click and the $82,793 kitchen about $9.72. Run the same chain on the $60,645 bathroom addition and it comes out near $7.12.
Put those side by side and the account sorts itself. The addition ceiling sits inside its range, so additions can compete from day one. The kitchen ceiling sits near the bottom of its range. The bathroom ceiling sits below the bottom of its range at those rates. Your own margin, close rate and booking rate will move every one of those numbers, which is why Chapter 3 told you to pull them first.
One more thing about week one. A brand-new keyword shows a dash where its Quality Score should be, until enough searches exactly match it for Google to score it. In our experience, a new account should open toward the top of what each project can afford, not the bottom. The early clicks are how the account earns the quality that lowers prices later. When the dashes turn into scores and the top-of-page numbers hold, ease each bid down in small steps.
The Searches That Book Estimates Should Show Above the Results 75-90% of the Time.
A bid is a setting. Whether it works shows up in the columns. In the accounts we build, five columns go on the ad group and keyword screens before the first kitchen click comes in:
- How often you're on top. Google's name is Search top impression share. Take every kitchen search your ad qualified to show above the listings for; this is the slice where it actually did.
- How often you're first. Search absolute top impression share: the same slice, counting only the first ad on the page.
- Why you missed, part one. Search lost top impression share (rank). A rival's kitchen ad outranked yours. Raise the bid, sharpen the ad, or fix the kitchen page it lands on.
- Why you missed, part two. Search lost top impression share (budget). The money ran out.
- Quality Score, keyword by keyword, with its three parts.
There's a trap in the menu next to them. Google also offers top impression rate columns, and its own help page warns against bidding toward those: a higher bid can pull you into tougher auctions where you land lower, so the rate can drop while you're winning more. The share columns are the ones Google says you can bid toward.
Our rule is 75-90% top impression share on the searches that book estimates. Below 75%, too many homeowners who typed exactly what you sell are seeing someone else's ad first. Above 90%, in our experience, you're paying the steepest prices in the auction for the last few appearances, and that money usually books more estimates somewhere else.
Those two lost columns point at different fixes. Losing to rank means the bid is too low for your quality, or the quality is too low for your bid. Chapter 4's ad group and landing page work is the cheaper fix. A budget loss says the bids would win but the day's money can run out early, and the homeowner searching that evening never sees you.
Read the numbers by the week, not by the day. Google calls impression share an estimate, and its help page says small swings don't necessarily mean anything needs to change.
Auction Insights Shows Who You're Paying to Outbid, and When to Leave the Auction.
The impression share columns tell you how often you won. The auction insights report tells you who you were up against. Open it for the bathroom ad group alone and every rival in those auctions gets a row of six numbers. Together they say how much of the bathroom traffic each rival reached, how often the two of you appeared on one results page, who placed higher when you did, and how often each of you held the top of the page or the first ad slot.
Two things to know before you open it. Google leaves the report blank for any ad group reaching fewer than one in ten of its possible appearances, so a bathroom group bid too low to compete can't see who it's losing to. And it names domains, not budgets. You see who's there, not what they're spending.
In our example market, you know some of the names before you run it. The bath searches are where the national bath brands advertise. West Shore Home had a text ad on a Raleigh bathroom search the day we checked. Five Star Bath Solutions had ads running from 14 different franchise accounts in Google's Ads Transparency Center, and Re-Bath and Bath Fitter both had current ads there. A general contractor bidding on bathrooms is in the same room as companies built around bathrooms.
That's exactly what the report is for: deciding whether a fight is worth it. When an ad group's lost-to-rank column won't come down without pushing its bid past the ceiling the job can afford, you have two honest choices. Lower the bid and accept a smaller share of that auction, or pause the ad group and move its money to the projects where you win at a price that pays. For a lot of general contractors, that means bathrooms take a back seat to kitchens, additions and whole-home work. A bath specialist with better margins may run the same math and stay.
Before you trust any cost from these campaigns, listen to the calls. Once a week, pull a sample of the calls and forms each ad group produced and sort them: a homeowner who booked an estimate, a homeowner who didn't, a vendor, a job seeker, a wrong number. Chapter 1 set the booked estimate as the conversion. If a pitch for roofing leads or a subcontractor looking for work is being counted, every cost per estimate in the account is lower than the truth, and every decision in the next section starts from a number that's too kind.
Thirty Booked Estimates Turn Your Hand-Set Bids Into Broad Match's Target.
Here's where the manual work pays off twice. Chapter 6 launched Exploration (our other name: broad match, with Google setting the bids toward booked estimates) once the account reaches roughly 30 quality conversions a month. Months 2-3 is the usual pace. A bigger account with more budget and a busier phone gets there sooner.
Exploration needs one number from you: its Target CPA, the average you're willing to pay for a booked estimate. (Google renamed the strategy in June 2026; what used to read "Maximize conversions with a Target CPA" now reads "Target CPA.") You don't guess it. Isolation has been measuring it for months.
Take Isolation's last 30 days. Divide what it spent by the booked estimates it produced. Say $10,500 bought 30 booked estimates: that's $350 each, and $350 becomes Exploration's starting target. That's no coincidence with Chapter 3, either. Its illustrative bathroom could afford about $353 per booked estimate, and keeping that number in view is how you stay honest about what the jobs can pay.
Google's own method lands in the same place. Google's Target CPA page says that once a campaign has conversions on record, its suggested target is simply the past month's average cost per conversion, with an allowance for conversions still on their way in. That adjustment matters for contractors. A homeowner who clicks on Tuesday and books the estimate the following week is still being counted, and a target set the day after a slow week can read lower than it really is. Google also suggests judging the target only after 30 or more conversions. Its help page also advises against separate targets for each ad group, because they restrict Smart Bidding, so Exploration runs on one campaign target while Isolation keeps its bids by project.
The temptation is to shave the number. Don't. A target set below your real cost tells Google to skip clicks it could have turned into estimates, and a new broad match campaign has the least room for that. Launch at what Isolation actually paid, then lower it in small steps once Exploration has a month of its own results.
Isolation doesn't stop when Exploration starts. Its manual bids keep protecting the searches you named in Chapter 5, its columns keep showing whether you hold the top of the page, and its cost per estimate keeps checking the number Exploration is working toward. Chapter 6 covered how the two campaigns trade off over time. This chapter gives them a shared measure.
If you'd rather have someone set and read these bids every week, our Google Ads management runs every contractor account this way.
Next: The Ads That Win These Auctions Answer the Homeowner's First Worry.
Chapter 8 writes the ads themselves, starting from what remodeling clients complain about most in their reviews and turning the answer into a headline.
This chapter is part of Google Ads for Contractors, our sixteen-chapter guide for residential general contractors and construction companies anywhere in the US.
