How to Find the Keywords Your Competitors Rank For

A keyword gap analysis sounds simple. List every keyword a competitor ranks for, subtract the ones you already rank for, and what is left is your content plan. Every tutorial says the same thing: enter your competitor's domain.

Then you actually look at who ranks for your money queries. It is not the shop across town. It is Angi, Yelp, Thumbtack, Houzz, a Reddit thread, and a 'best plumbers in Austin' listicle from a local news site. Feed those domains into a gap tool and you get a keyword list you cannot act on, because the pages that rank for them are directories and you are not a directory.

The version that works for a service business has four steps: separate SERP competitors from business competitors, run the gap only against operators in your weight class, delete the keywords aggregators own, then filter the survivors for commercial intent. Everything below is the detail.

Who is actually your competitor in Google — and who just looks like one?

Your Google competitor is any domain sitting in a result you want, and in a service SERP most of those domains are not businesses like yours. Ahrefs looked at the top 10 results for 1.3 million US keywords in 2025 and found 72.9% of top-10 pages were more than three years old, with the average #1 page five years old. A service SERP is a queue of aged incumbents, and the incumbents are mostly platforms.

So before you type a single domain into a gap tool, sort every result on page one into a bucket. The bucket decides whether the keyword is even reachable.

The last row is your business competitor: the company that takes a call that would otherwise have been yours. Everyone above it is a SERP competitor — an obstacle, not a benchmark. Gap analysis only produces usable output when you run it against the last row.

Why does running a gap analysis against Angi waste your afternoon?

Because the export comes back unwinnable, and the math is brutal: Ahrefs tracked 1 million random URLs for a year and found only 1.74% of newly published pages reached Google's top 10 — in a separate 2-million-URL sample limited to non-empty English pages, 6.11% did. Your odds against an aggregator's aged, heavily linked city page are worse than that baseline, not better.

But the link gap is not even the real problem. Intent is. When Google ranks four directories for 'plumbers in Austin', it is telling you the searcher wants a list of ten providers to compare — not one provider's homepage. You cannot honestly publish a 'best plumbers in Austin' page about yourself. The keyword is not hard for you. It is structurally wrong for you.

Run the gap against Angi and you also inherit their page count. Their keyword list is one row per service, per city, times thousands of cities. Copying that means mass-producing thin location pages — the exact pattern that gets a small site's index quality flagged.

How do you find the local operators you can realistically outrank?

Build the list by hand from live SERPs, then keep only the domains whose referring-domain count is inside roughly 2x yours — most gap tools cap you at three or four competitor domains anyway, so you get 3–4 slots and should spend them well.

• Search 8–10 of your real money queries (service + city, 'emergency X near me', 'X repair cost [city]') in an incognito window set to your city.
• Ignore ads. Ignore the map pack. Write down only the organic non-directory domains that appear more than once.
• Pull the domain rating and referring-domain count for each. Drop anyone with more than about twice your referring domains — you will not close that gap inside a year.
• Keep the 3–4 closest to your own profile. Those are your comparables.
• If nobody local has a real website, widen the net: find an operator in a same-size metro who does. They are not competing for your calls, but their keyword footprint is a valid model for yours.

That last case is common. In a lot of trades, the whole local field is running a five-page site from 2016. That is good news — it means the gap you should be running is against a comparable operator in a different city, and the entire keyword set is open in yours. Get the fundamentals right first; our technical SEO service exists because a fast, crawlable site is what makes those easy keywords actually land.

How do you run the gap and read the output?

Enter your domain plus your 3–4 comparables, then filter hard before you export. Ahrefs' own content gap workflow recommends excluding competitor brand terms, requiring at least two competitors in the top 10, setting a minimum search volume of 20, and capping Keyword Difficulty at 30; it also warns that a list over 100,000 keywords is too big to work with. If your export is that large, your filters are wrong.

Semrush's Keyword Gap tool splits the output into tabs, and it recommends a KD filter of 0–49 if your domain has not built much authority yet. Read the tabs in this order:

Verdict: work Weak first. Every keyword in that tab is a page that already exists, already has some equity, and needs an edit rather than a new asset. Missing is second. Untapped is a maybe-pile, not a plan.

Ahrefs also splits gaps into two kinds worth naming: domain-level gaps (they cover a topic you have no page for) and page-level gaps (you both have a page, theirs ranks for more keywords because it is more complete). Domain-level gaps get a brief. Page-level gaps get a section added to the page you already published. Do not create a second page for a topic you already cover — you will just cannibalize yourself.

What do you do with the keywords the directories own?

You delete most of them, and the test takes ten seconds: open the live SERP, and if four or more of the top five results are directories or listicles, Google has decided that query wants a list. One operator page will not win it. But deleting is not your only option — a directory-owned keyword can still be monetized three other ways.

• Buy it instead of earning it. If 'plumbers in [city]' is owned by aggregators forever, that is a paid search line item, not a content brief. Rent the position you cannot own.
• Get onto the page that beats you. If a local news listicle ranks #1, the winning move is being in it. That is outreach, not writing — see link building for service businesses.
• Reframe to the sub-question. The directory page ranks for 'best water heater installers in Austin' but answers 'how much does water heater replacement cost in Austin' badly, in two sentences. That question is yours to take.

That third move is where most of your winnable volume actually lives. Aggregator pages are wide and shallow by design. They list twenty companies and dedicate a paragraph to the actual question. A specific, honest, numbers-first answer from an operator beats that paragraph — and it is a page you can genuinely write.

How do you filter the survivors for commercial intent?

Keep the queries with a wallet on the other end and cut the rest — and commercial queries have a second advantage right now: they are the ones AI Overviews intercept least. In Seer Interactive's February 2026 analysis of 53 brand accounts and roughly 49,000 tracked queries, AI Overviews appeared on 36% of informational queries but only 8% of commercial and 5% of transactional ones.

So the buyer-intent keyword you keep is also the keyword least likely to have its click eaten by a summary box. Run every survivor through three questions:

• Does the keyword name a service, a symptom, a price, or a place? 'AC not cooling' and 'crown cost without insurance' pass. 'What is HVAC' does not.
• Would you pay for the click in Google Ads? If you would not bid on it, it is not worth a page either.
• Does the searcher want to hire someone, or do it themselves? DIY tutorials pull traffic that will never call you. Skip them unless they are a genuine first step on a path you can walk a buyer down.

A keyword with 40 searches a month and hiring intent is worth more than one with 2,000 and none. That is the same rule that governs keyword research for a service business generally, and it is the rule that survives every algorithm update: rank for the query a person types when they are about to spend money.

Do you need a paid tool to do this?

No — the free path gets you the same answer, it just costs a day instead of an hour, and for a single-location operator with fewer than about 30 target keywords the free path is genuinely fine.

Verdict: buy the tool if you manage more than one location or more than roughly 50 keywords; otherwise do it by hand once, and you will learn your SERP better than any export would teach you. If you are picking a platform anyway, we broke down Semrush vs Ahrefs vs Moz for service businesses.

One warning that applies to every tool: their position data is an estimate from a crawled index, not what a customer in your city sees. Before you commit budget to your top 10 gap keywords, open each one in a live, incognito, city-set search and confirm the SERP looks like the tool says it does. It often does not.

If you would rather see the finished list than build it, that is essentially what we hand over in an SEO engagement: your real competitors, the keywords they own that you can take, and the ones you should stop chasing. Get my free audit and we will map the gap for your market.