A roofing company in Charlotte, North Carolina opens Google's Keyword Planner for the first time, types in "roof repair," "roofer" and "roof replacement," and clicks the button. Keyword Planner comes back with 1,693 keyword ideas. We ran exactly that search on November 12, 2026, and that is the number it returned.
Most owners do one of two things next. They pick the keywords with the biggest search numbers, or they add every idea that looks related and let Google sort it out. Both feel sensible, and both burn money.
The question that decides whether Google Ads pays for itself is not "which keywords get searched the most." It is "which searches come from someone who is about to hire a business like mine." This article is about how to answer that, in order, before you spend a dollar.
Keyword Planner Gave a Charlotte Roofer 1,693 Ideas. A $1,500 Budget Can Afford About Ten.
Keyword Planner is built to show you how big a market is. It is very good at that. What it cannot tell you is which of those 1,693 searches will turn into a signed job, and it was never designed to.
Look at what a click costs in that same Charlotte search. Keyword Planner estimates the top-of-page bid for "roof repair" at $21.78 to $105.73. For "roof repair near me," it is $34.54 to $103.31. For "roofing companies near me," it is $17.92 to $74.14. Those are ranges of what other advertisers have been bidding, not your price, and in our experience real costs land closer to the top of the range than the middle. We explain why in Keyword Planner's Bid Range Isn't What You'll Pay per Click.
Now take a $1,500 monthly budget. Google spreads a monthly budget across 30.4 days, so that is about $49 a day. At $75 to $105 a click, near the top of those ranges, $49 a day buys about half a click. Across a month, that is roughly 14 to 20 clicks in total.
Twenty clicks a month is not enough to spread across 1,693 keywords, or 200, or 50. At best, it starts testing one small, tight group of searches to find out whether it brings in jobs. So the real job of keyword research for a small business is not finding more keywords. It is throwing away almost all of them.
The search volume column will not help you do that. Keyword Planner rounds its monthly searches into broad buckets, so treat its numbers as a way to compare one search with another, not as a count of people. "Roofers near me" at 480 a month and "roof leak repair" at 210 a month tells you the first is the bigger search. It does not tell you that exactly 480 people typed it, and it says nothing about which one is more likely to call you.
"Roof Leak Repair Near Me" Is a Customer. "Types of Shingles" Is a Reader.
Every search is a person in the middle of doing something. Some are about to pick up the phone. Some are learning. Some are not your customer at all. The words they type usually tell you which, and that is what you sort by.
Think of it as a ladder with four rungs.
Rung 1: Ready to hire. The search names your service and a sign that the person wants it done by someone. Location words like "near me" or "Charlotte." Urgency words like "emergency," "same day" or "leak." Hiring words like "company," "contractor," "quote" or "estimate." For our roofer, that is "roof repair near me," "roof leak repair," "emergency roof repair," "roofing companies near me" and "roof replacement Charlotte."
Rung 2: Comparing. The person knows they need a roofer and is deciding which one. "Best roofing company," "best roofers near me," "roofer reviews," "metal vs shingle roof." These are real buyers, just one step earlier, and they usually take longer to call.
Rung 3: Learning. The person is researching, not hiring. "Types of shingles," "how long does a roof last," "how to fix a roof leak," "what does hail damage look like." Some of them will need a roofer someday. Very few need one this week, and a small budget cannot afford to wait for them.
Rung 4: Wrong person. The search contains your service but the person will never pay you for it. "Roofing jobs," "roofer salary," "roofing supply near me," "roof patch Home Depot." Keyword Planner suggested "roofer supply near me" and "home depot roof repair" in our Charlotte search, right alongside the real buyers. Keep them off the list entirely.
On a small budget, you buy rung 1 and nothing else until it is working. Rung 2 comes next, once rung 1 is producing jobs at a price you can live with. Rung 3 is for businesses with money to spare and a long sales cycle to fill. Rung 4 is never.
This is also why the biggest search numbers are a trap. In Charlotte, the generic word "roofer" gets 260 searches a month, and nobody can tell you whether the person typing it wants a roof, a job or a definition. "Roof leak repair" gets fewer searches and leaves no doubt at all.
"Roof Repair Cost" Is Someone With a Leak. "Roof Replacement Cost" Is Someone Collecting Three Quotes.
Price searches are the hardest call on the ladder, and the bid ranges show that other advertisers are not sure about them either.
In Charlotte, the top-of-page estimate for "roof repair" is $21.78 to $105.73. For "roof repair cost," it is $4.62 to $18.74. "Roof replacement" runs $19.66 to $66.12, while "roof replacement cost" runs $3.99 to $19.99. Adding the word "cost" cuts what competitors bid by 70% to 80%. The market has decided that people asking about price are worth much less than people asking for the service.
That is sometimes wrong, and the difference comes down to one question: how long can this customer wait?
Someone searching "roof repair cost" usually has water coming through a ceiling. They are not going to spend three weekends researching. They want to know whether this is a $400 problem or a $4,000 problem, and then they want someone on the roof. For an urgent, lower-ticket job, a price search is close to rung 1, and it is cheap because everyone else is ignoring it.
Someone searching "roof replacement cost" is planning a big project. NerdWallet puts the average roof replacement at about $9,500 in 2026. They will get three quotes, ask a neighbor, check reviews, and maybe wait until spring. That search belongs on rung 2 at best. It can still be worth paying for, but only if your landing page actually talks about price ranges. A page that dodges the question is a fast way to pay for a click and lose the visitor.
Apply the same test to your own business. If the job is urgent and the ticket is modest, like a burst pipe, a locked-out car or a broken AC in July, price searches are buyers. If the job is big, planned and easy to postpone, like a remodel, a new roof or a fence, price searches are shoppers. Buy the first kind early. Add the second kind later, with a page built to answer them.
Your Contact Forms and Call Recordings Are a Keyword List You Already Paid For
The words you would use to describe your business are often not the words your customers type. A roofer says "flashing failure." A homeowner types "water stain on ceiling after rain." A roofer says "residential re-roof." A homeowner types "new roof cost." If your keyword list is written in your words, it will miss the searches that matter most.
You already have a record of how customers talk. It is sitting in four places:
- Contact form messages. Read the last 50. Customers describe the problem in their own words, and those words are often exactly what they typed into Google first.
- Call recordings or call notes. Listen to how people open the call. "I've got a leak over my kitchen" is a keyword. "My insurance company told me to get an estimate" is another, and it is one many roofers never think to target.
- Reviews, yours and your competitors'. Reviews tell you what customers were worried about and what they were relieved by. "Came out the same day" tells you "same day roof repair" matters to people.
- The person who answers the phone. Ask them which questions they hear every week. They will give you a better keyword list in five minutes than an afternoon in Keyword Planner.
Pull every phrase that describes a job you want, then sort each one onto the ladder. Keep the rung 1 phrases. You will usually end up with ten to twenty that fall into two or three natural groups, such as "leak repair," "storm damage" and "replacement." That is your starting list.
Then take those groups back to Keyword Planner, not for ideas this time, but to check the price. Keyword Planner is good at telling you which of two searches is bigger and which one competitors fight harder over. It is the wrong place to start your list and a good place to sanity-check it.
Your Budget Sets the Size of Your Keyword List, Not Google
Here is the arithmetic that should decide how many keyword groups you run. It takes two minutes.
- Daily budget. Monthly budget divided by 30.4. For $1,500, that is about $49.
- Likely cost per click. Take the top of Keyword Planner's range for your main rung 1 searches. For our Charlotte roofer, that is roughly $75 to $105.
- Clicks per month. Monthly budget divided by cost per click. $1,500 at $100 a click is 15 clicks.
- Clicks each group needs. A group of keywords has to collect a minimum number of clicks before it can show you whether it brings in calls. If roughly one click in ten turns into a lead, a group needs about 30 clicks to produce three leads, which is the bare minimum before you can judge it. Your own rate will differ, so use it once you have it.
Run those numbers for the Charlotte roofer and the answer is uncomfortable. Fifteen clicks a month cannot even fully test one group. At $1,500, this business should pick its single best rung 1 group, most likely leak repair, put every dollar behind it, and leave replacement and storm damage for later. It takes about $3,000 a month to fully test one group at these prices, and about $6,000 to test two.
The same arithmetic works for any business. A plumber paying $25 a click on $1,500 gets 60 clicks, enough for about two groups. A lawyer paying $150 a click gets ten, which is not enough for even one. The number of keywords you can afford was never set by Google or by how many ideas Keyword Planner hands you. It is set by your budget divided by your cost per click.
If the math says your budget cannot fully test even one group, that is a real answer, not a reason to spread thinner. We walk through what to do about it, including shrinking the area or the service you advertise, in Google Has No Minimum Spend. Your Market Does.
Google Matches Meaning Now, So Ten Spellings of One Search Buy You Nothing
Years ago, a good keyword list was long. You added "roof repair," "repair roof," "roof repairs," "roofing repair" and "repair my roof" as separate keywords, because Google only showed your ad on the words you listed. That is no longer how it works.
Google's own help page now describes exact match as showing ads on "searches that have the same meaning or same intent as the keyword." Phrase match goes wider, to any search that contains that meaning, even when the words are different. The words matter less than what they mean. The same page says that very similar broad match keywords, like "red car" and "car red," are recognized as duplicates, and only the one with the higher Ad Rank is used. You get one bid, not five.
Keyword Planner shows the same thing from the other side. In our Charlotte search, "roofing companies near me," "roofing contractors near me," "near me roofing company" and "roofing business near me" each came back with exactly 1,000 monthly searches and exactly the same bid range. Those are not four searches. They are one search wearing four outfits, and Google prices them that way.
So a strong list is short and grouped by meaning. For each rung 1 group, pick a handful of keywords that each say something different: the core service, the service with a location, the urgent version, the problem in the customer's words. "Roof leak repair," "roof repair near me," "emergency roof repair," "water stain on ceiling." Four keywords that mean four different things are worth more than forty that mean one.
Each group then gets its own ad, written to that group, and its own page. Someone searching "emergency roof repair" should see an ad that says you come out the same day and land on a page about emergency repairs, not your homepage. How to structure those groups so each has enough data to learn is its own subject, covered in You Organized Your Google Ads Account. You Starved It. And because looser matching means your ads will show on searches you never typed, start a new account on phrase and exact match, not broad. Broad Match Will Waste Your Money. Unless You Do This. explains when broad match earns its place.
Your First Month of Search Terms Proves Which Keywords Were Worth the Money
Everything up to now has been a well-informed guess. After the first month, you stop guessing.
Open the search terms report. It lists the actual searches that triggered your ads, which is a different list from the keywords you chose. Judge each group by one number first: how many calls, forms or booked jobs did it produce, and what did each one cost? Clicks and click-through rate are interesting. They do not pay the bills.
Three things usually happen.
Some searches you never thought of are bringing in leads. Add them as keywords in the group they belong to. These are often the customer-language searches from the earlier section, like a specific problem or a neighborhood name, and they are some of the most valuable keywords you will ever find.
Some searches are spending money and producing nothing. If they are on the wrong rung, block them. We cover exactly how to do that, and which words almost every business should block on day one, in You're Paying Google for Searches You'd Never Choose.
One group is clearly working. That is your signal to grow, and growth happens one rung at a time. Raise the budget on the group that works before you add a new one. When that group is producing jobs at a cost you are happy with, add the next rung 1 group. When rung 1 is fully funded, add rung 2. Never jump from a single working group straight to a long list of learning searches, because you will hand back every dollar the working group earned.
Most of the advice out there treats keyword research as something you do once, before launch. It is really a monthly habit. The list you start with is the list you could afford to guess. The list you have six months later is the one your customers wrote for you.
Ten Ready-to-Hire Keywords Will Out-Earn 1,693 Keyword Planner Ideas
Choosing keywords for Google Ads comes down to a few decisions, made in order:
- Sort every idea by what the searcher is about to do, and buy only the ready-to-hire searches at first.
- Decide whether price searches are buyers or shoppers in your business by asking how long your customer can wait.
- Build the starting list from your customers' own words, then check the price in Keyword Planner.
- Let your budget divided by your cost per click decide how many groups you run.
- Keep each group short and grouped by meaning, with its own ad and page.
- Let the first month's search terms tell you what to add, what to block and where to grow.
None of this needs a big budget. It needs the discipline to spend a small one only on the people who are ready to call.
If you would rather hand this to people who sort search terms for a living, that is a big part of what our Google Ads management team does every week. And if you are doing it yourself, start with your last 50 contact form messages. Your first ten keywords are probably already in there.
