Keyword Research for Contractors: Finding the Terms That Book Jobs

Why most contractor keyword lists are built backwards

The usual mistake is starting with volume. An owner opens a tool, sees "plumber" gets 90,000 searches a month, and decides that is the word to chase. Then he spends a year losing to national directories and wondering why the phone stays quiet. "Plumber" is a research term, not a hiring term. The person typing it might be a homeowner, a student, a job-seeker, or a competitor. Volume without intent is just noise you pay to rank for.

The job-booking terms sit lower on the volume chart and higher on the intent chart. "Water heater replacement cost Fort Myers" gets a fraction of the searches, but the person typing it has a broken tank, a service area, and money in mind. That is the whole game: fewer searches, warmer buyers, jobs that actually close.

Build the list from the bottom up instead. Start with the jobs you want more of. Write down what a homeowner would actually type the night their unit quits or their gutter falls. Then group those phrases by intent, not by search count. The volume number is a tiebreaker, never the starting point. A term with 40 searches a month where every searcher needs your exact service in your exact town is worth more than a term with 4,000 where nine in ten are just curious.

There is a second trap hiding inside the volume trap: seasonality. Contractor demand swings hard by month. "AC repair" barely registers in a Florida January and spikes in July. "Gutter cleaning" climbs in fall. A tool's annual-average number can make a strong seasonal term look weak, or a weak one look steady. When you read volume, read it as a curve, not a single figure, and build the pages for your busy season before the season, not during it.

This is also why generic agencies underdeliver here. They pull a keyword report, sort by volume, and hand you 500 terms nobody in your ZIP will ever type. Since 2008 we have worked one lane, home-service contractors across 20 trades, and the pattern holds in every one: the money words are specific, local, and boring to a marketer but obvious to a homeowner in trouble. The owner who writes his own first list from his own call notes almost always beats the vendor report, because he knows what people actually ask for.

The four buckets every contractor keyword falls into

Sort every phrase you find into one of four buckets. This one habit does more for a contractor site than any tool subscription, because it tells you which words deserve a page and which are a waste of a build.

Service-plus-city terms are the ones that pay the mortgage. Each real service you offer, crossed with each real town you serve, is a candidate page. A roofer covering four towns with three core services has a dozen honest money pages before he writes a single blog post.

Problem-and-symptom terms catch the homeowner earlier, when they are diagnosing, not yet hiring. Answer the question plainly, then point them to the service page. Cost and research terms work the same way: give a real range, be honest about what moves the price, and hand off to a quote. The broad bucket is where budgets go to die. Skip it until everything else ranks.

Watch for terms that look like one bucket but belong in another. "Roof inspection" reads like a service term, but half the people typing it want a free real-estate inspection they will never pay for, while the other half have a leak. The tell is the modifier: "roof inspection for insurance claim" or "storm damage roof inspection" is a buyer; bare "roof inspection" is a coin flip. When a phrase straddles two buckets, split it into its clearer variants and target those, not the ambiguous root.

Where to actually find the terms (free tools you already have)

You do not need to buy anything to build a strong first list. The best sources are the ones Google gives away and the ones sitting in your own business.

• Google autocomplete. Type your service and pause. Add "in [your town]," add "cost," add "near me," add "emergency." Every suggestion is a phrase real people finish typing.
• People also ask. Run a search for your main service. That accordion of questions is a ready-made list of the things homeowners want answered before they call.
• Related searches. Scroll to the bottom of the results page. Those eight links are Google telling you what else this searcher looks for.
• Your own call log. The exact words a customer used on the phone ("I need someone to fix a leak under my sink") are keywords. Write down how callers describe their problem for two weeks.
• Your reviews and your competitors' reviews. Customers name the exact jobs and towns in their own words. Mine both.

Google Search Console is the strongest free source once your site has any history. It shows the queries that already put your site in front of people, including ones you never targeted. Sort by impressions with low clicks: those are terms you rank for badly and could win with a dedicated page.

One more free source people overlook: the autocomplete on YouTube and on Amazon-style sites, and the questions on forums like Reddit and Quora where homeowners describe problems in full. People are blunter when they think a marketer is not listening. "Why is my ceiling wet after it rains but no visible leak" is a real forum thread and a real page you could own. Read how homeowners talk when they are frustrated, then write the heading in their words.

Paid tools (the ones that show volume and difficulty) are useful for prioritizing, not discovering. They confirm whether "gutter guard installation Sarasota" is searched a lot or a little, and difficulty scores hint at how crowded a term is. Handy, but you can build a booking list without them. Spend the tool money on one thing if you spend it at all: seeing which terms your competitors already rank for, so you know the field before you pick your battles. Sort their ranked terms by traffic and you get a shortlist someone already validated with real money.

Reading intent: which terms bring buyers, which bring browsers

Intent is the difference between a phone call and a bounce. Two phrases can have the same volume and completely different value. Learning to read intent by eye is the single most useful skill in contractor keyword research, and it takes about five minutes to learn.

Buyer-intent signals are words that mean action is close: "replacement," "installation," "repair," "cost," "quote," "near me," "emergency," and any town name. Browser-intent signals mean the person is still reading: "how does," "what is," "DIY," "vs," "types of." Both matter, but they need different pages. Buyer terms get service pages built to convert. Browser terms get guides built to inform, then hand the reader to a service page.

The fastest way to check intent you are unsure about is to look at what already ranks. Search the phrase. If the first page is full of service providers with quote buttons, it is a buyer term. If it is full of articles, videos, and "ultimate guide" posts, it is a research term. Google has already sorted the intent for you; match your page type to what it rewards.

A short worked example. "Standing seam metal roof" looks like a buyer term but usually ranks articles: people are learning the style. "Metal roof installers near me" ranks providers: pure buyer. "Metal roof cost per square foot" ranks calculators and cost guides: research with money on the mind. Same trade, three different page types, one honest guide that hands off to a service page. Get this sorting right and you stop building pages that rank for words nobody hires on.

Urgency is its own intent signal worth calling out, because it is the warmest of all. "Emergency," "same day," "24 hour," "after hours," and "burst" or "flooding" mean the homeowner will call the first credible result and not shop around. These terms convert at a rate the volume never predicts, and they are usually under-defended because competitors chase the bigger numbers. If your trade has emergency work, an honest same-day page targeting that language is often the fastest phone-ringer on the whole site.

One intent, one page: mapping keywords to a site structure

Here is the rule that keeps a contractor site from cannibalizing itself: one clear buyer intent gets exactly one page. When you build three near-identical "roof repair" pages hoping to catch more traffic, they compete with each other, split their ranking strength, and confuse Google about which one to show. One strong page beats three thin ones every time.

Group your sorted keywords into clusters where every phrase means the same job. "AC repair," "air conditioner not cooling," "AC fixing service," and "cooling system repair" all point at one intent: fix my AC. That is one page, targeting the strongest phrase in the URL and title, with the variations woven naturally into the copy. You are not writing for the robot; you are covering the ways a real homeowner might phrase the same need.

Then layer geography. Each core service crossed with each real service town becomes its own page, and only for towns you genuinely serve. Do not build "roofing Miami" if you have never crossed the county line. Fake location pages read as thin and rank poorly, and they burn trust with the one homeowner who does call. A page per service, a page per service-plus-town you truly cover, and a guide per research question is a clean, honest map.

Link the map together so it works as one machine. Service pages link up to a hub page for the trade; town pages link to the parent service; guides link down to the service page they answer for. This internal linking is not decoration. It tells Google how the pages relate and it passes ranking strength from your strongest pages to your newest ones, which is how a fresh town page climbs faster than it would alone.

A cluster this size (94-plus pages is typical for a multi-service, multi-town contractor) is not padding. It is one honest answer for each way a buyer might search, all linked so ranking strength flows between them. Every page earns its place by matching a real intent to a real page; the day you cannot say what job a page books, it should not exist. For how those pages get built, ranked, and measured over time, that is the job of the wider SEO plan; here the point is the map, not the construction.

How AI answers are changing which keywords matter

Keyword research used to end at Google's blue links. It does not anymore. When a homeowner asks ChatGPT or Google's AI overview "who does emergency roof repair near me," the answer is assembled from pages the model trusts, and being in that answer is the new front page. This shifts what you research and how you phrase pages, though the underlying discipline is the same: know the exact question, answer it cleanly.

The practical change is that full-sentence, question-shaped queries matter more than they did. People type differently when they expect a conversation than when they expect ten links. "What does it cost to replace a 2000 square foot roof in Southwest Florida" is a real query now, and the page that answers it in plain language with a real range is the page that gets cited. Your keyword list should include these longer, spoken-style questions, not just the clipped two-word terms.

Structure earns the citation. Pages that answer the question up top in two or three sentences, use plain headings that match how people ask, and give honest specifics (ranges, timelines, what is and is not included) are the ones AI answers pull from. This is where keyword research and page structure meet: the question becomes the heading, the answer sits directly beneath it.

A concrete way to find these queries: type your service into an AI chat and watch how it asks follow-up questions, or read the "People also ask" box, which increasingly mirrors how the AI models phrase things. Note the full sentences, the ones with a subject and a verb, not just the keyword fragment. Then check which of your buyer intents already have a clean question form and add the missing ones. Cost, timeline, "is it worth it," and "who does this near me" are the four questions almost every homeowner asks before hiring, in every trade.

AI-answer optimization is its own discipline and its own conversation; the point for keyword research is narrow. Add the question-shaped, conversational phrases to your list, and write pages that answer them in the first breath. Do that and the same page can win a blue link and a spot in the AI answer. That overlap is the highest-value real estate in contractor search right now, and it costs you nothing extra: you are already researching the questions, you are just writing them down the way a homeowner would say them out loud.