How to Find Competitor Keywords: 10 Sources, Including Free Ones

The entire market for this question is owned by companies selling keyword tools, which is why every answer to it recommends a keyword tool. The tools are genuinely useful and you should use one. It is worth being precise about what they know, though, because two very different things get presented in the same table of results.

A platform learns which keywords a competitor ranks for by watching search results and recording what it sees, which is close to observation and reasonably trustworthy. Everything numeric beside that keyword is modelled: search volume from clickstream and planner data, traffic from an assumed click curve, difficulty from a proprietary score. Meanwhile two datasets that measured something real about the terms you and your competitors both compete on are sitting in accounts you already have and pay nothing for.

1. 1Identify your search competitors, who are not your business competitors. The sites taking the traffic you want are frequently review sites, consultancies and publishers rather than the vendors you meet in deals. Enter your five most commercially important terms and write down who actually ranks. That is the list to research.
2. 2Start with the data you measured yourself. Open Search Console and export every query where you already appear. Impressions with a poor average position are terms where the market has decided you are relevant and you are losing. That set is real observed data, not an estimate.
3. 3Pull their organic set and cut it to the money terms. Run each competitor's domain through an SEO platform, then discard everything that is not a buying, comparison or problem query. A competitor ranking for ten thousand terms is usually earning revenue from about forty of them.
4. 4Separate what they buy from what they earn. Paid and organic keywords are two different strategy statements. Terms they pay for are terms they need now; terms they rank for are terms they invested in a year ago. A term appearing in both is one they consider existential.
5. 5Read their site structure directly. Open their sitemap and their navigation. The folders, the clusters and the comparison pages show you the keyword plan as the team drew it, including the sections that are new and not yet ranking, which no tool will show you.
6. 6Test the answer engines with buying prompts. Ask the assistants your buyers use the questions your buyers ask, and record which competitors get named and which sources get cited. Prompts are not keywords, and this layer is invisible to every keyword tool.
7. 7Turn the list into a decision and set a re-check. Pick the handful of terms where the gap is both winnable and commercially meaningful, decide whether each is a content, product or paid response, and re-run the comparison quarterly to see what they added.

No free tool hands you a competitor’s complete keyword list, and anything advertising that it does is showing you a teaser to sell a subscription. What is free is a set of sources that, combined, answer most of the question, and two of them are better than anything you can buy because they are measurements rather than estimates.

1. Your own Search Console

Export the performance report by query for the last twelve months. The rows to read are the ones with substantial impressions and an average position outside the top five: the engine considers you a candidate and users are choosing somebody else. Search each of those terms and record who wins it. You have now built a competitor keyword list from observed data rather than from a vendor’s estimate, scoped precisely to the terms that already touch your business.

2. Your own Auction Insights and search terms report

Auction Insights lists the advertisers appearing in the same auctions as you, with impression share and overlap rate against each. The search terms report gives the actual queries that triggered your ads, which is different from the keywords you targeted and frequently more revealing. Together they tell you exactly who is bidding against you and on what. The budget reading of the same report belongs to a different question and is covered separately.

3. Google Ads Keyword Planner

Enter a competitor’s domain or a landing page and Keyword Planner will suggest the terms Google associates with it, along with volume, competition level and a top-of-page bid range. The bid range is the number to pay attention to: it prices the commercial value of a term far better than any difficulty score. Accounts with no spend see volume as brackets rather than exact figures, which is annoying and rarely fatal for research purposes.

4. Bing Webmaster Tools keyword research

A free keyword research module with search volume, trend lines and related questions, available to anyone with a verified site and requiring no advertising spend at all. Volumes are Bing-scale, so use them for relative comparison rather than as Google forecasts. It is also the only free tool in this list that will happily give you a full keyword universe around a topic without asking for a credit card.

5. Autocomplete, People Also Ask and related searches

These come directly from what people type, which makes them the most accurate phrasing source in existence and the one no vendor can improve on. Type your category, then your category plus a competitor name, then your competitor name plus “vs” and “alternative”. The suggestions that appear are real demand, and the questions in the People Also Ask box are the sub-questions the engine believes belong to the topic. Check in a clean browser session, since results personalise.

6. SEO platforms

Ahrefs, Semrush, SE Ranking and Mangools all do the same core job: enter a domain, get the terms it ranks for and the terms it bids on, then run a gap report against your own. The output is enormous and mostly irrelevant, so the skill is filtering. Keep buying-intent terms, comparison terms, problem statements and anything containing a competitor name. Discard the informational long tail unless you have a specific reason to want it.

7. The competitor’s sitemap and site structure

Request their sitemap and read the URL structure and the navigation. This is the only source that shows you the plan rather than the result: folders reveal how they segmented the market, page clusters reveal which topics they committed real budget to, and brand-new sections reveal where they are heading before anything ranks. A competitor who has just published forty pages about a use case they never mentioned has told you their next move.

8. Public ad libraries

The ad libraries do not disclose keywords, but they disclose the creative running against them, and the creative names the promise, the audience and the objection being pre-empted. Read them alongside a paid keyword export and the pairing becomes far more useful than either alone: you learn both which query they are buying and what they say to the person who typed it.

9. Channel autocomplete on video, app and marketplace search

Each of these platforms runs its own search engine with its own query language, and none of them is covered by the standard keyword tools. Use the platform’s own autocomplete for phrasing, then open a competitor’s presence there, sort by popularity, and read the titles and descriptions of the best performers. Those titles are their keyword targets stated in plain sight.

10. AI assistant prompt testing

A growing share of evaluation now begins with a question typed into an assistant rather than a keyword typed into a search box, and no keyword tool covers it because prompts are not keywords. The method is manual and it works: keep a fixed list of buying questions, run them monthly, and record which vendors are named and which pages are cited. The guide to competitive analysis with AI covers how to structure that as a repeatable exercise.

This is the part that separates competitive research from SEO reporting. A keyword export is a list of decisions somebody made and paid for. Four readings turn it into intelligence.

• New topic clusters show the roadmap. A set of pages appearing around a use case, an industry or an integration they have never mentioned is a competitor investing ahead of a launch. Content teams get briefed before announcements go out, so this is one of the earliest external signals of a product direction.
• Comparison and alternatives terms show who they fear. Publishing a page targeting your name is a deliberate act aimed at buyers already considering you. Note which competitors they built these for and in what order, because that is their own ranking of the threat, written by their own team.
• Brand bidding shows where the pressure is. Paying for their own brand name means somebody is bidding on it. Paying for yours means they have decided your buyers are worth intercepting. Both are current decisions with a daily cost attached, which makes them a far more reliable signal than anything on their website.
• The vocabulary shift shows the segment move. Watch the modifiers. A competitor whose terms drift from “small business” toward “enterprise”, or from one industry to another, is repositioning, and the keyword set usually moves a quarter or two before the homepage does.

Those four readings belong in your competitive positioning work rather than in a search report, because each one changes what you say rather than what you publish.

1. 1Search the term yourself. Tool rankings are sampled and location-dependent. Before acting on “they rank first for this”, run the query in a clean session in the market you care about and look at what is actually on the page, including the AI answer above it.
2. 2Never compare numbers across two tools. Volume, traffic and difficulty are each vendor’s own model. Comparing a Semrush figure with an Ahrefs figure measures the difference between two models, not a change in the world. Pick one platform and stay inside it.
3. 3Check the intent behind the keyword. A high-volume term whose results are all definitions, courses or job listings is not a term your buyer types on the way to a purchase. Read the top ten before you value the keyword.
4. 4Cross-check paid claims against the ad libraries. An estimated paid keyword set is one of the least reliable outputs these tools produce. If a competitor is genuinely running a campaign, the creative is sitting in a public ad library and you can confirm it in a minute.
5. 5Confirm against your own impressions. If a tool says a competitor dominates a term that matters to you, your own Search Console should show you appearing on it too. When it shows nothing at all, you are looking at a term that lives in a different market from the one you sell into.

Researching which terms a competitor targets raises no legal question at all: search results are public and the tools read what anyone can see. Acting on the research is where the boundaries sit, and they fall almost entirely on the advertising side.

• Buying a competitor’s brand as a keyword is generally permitted. Google’s own policy allows it, and courts on both sides of the Atlantic have broadly accepted it. The Court of Justice of the European Union addressed the platform’s position in the Google France cases in 2010, holding that the search engine does not itself infringe by selling the keyword.
• The trademark in your ad text is the real exposure. In Interflora v Marks & Spencer in 2011 the same court set the European test: the ad must let a reasonably well-informed and reasonably observant internet user tell that the advertiser is a third party unconnected with the trademark owner. Ambiguous ad copy can lose a case in Europe that would have survived in the United States. Platform policy restricts trademarks in ad text independently of the law.
• Comparison pages are safer than comparison ads. Naming a competitor on a page you own, with accurate and sourced claims, is ordinary comparative marketing and is protected in most markets provided it is truthful and not misleading. The same claim compressed into an ad headline loses the context that makes it defensible.
• Rules differ by country, and by platform. Trademark treatment of keyword advertising varies across jurisdictions, and platforms apply different policies in different regions. Check locally before rolling out brand bidding at scale, particularly outside the US and the EU.

• Their actual keyword list. What a tool shows is an inferred set built from observed rankings and detected ads, not an export from their account. Proxy: their sitemap and page titles, which show what they built on purpose, and Auction Insights for the paid terms you share.
• Which keywords make them money. Rankings are not revenue, and a competitor can rank beautifully for terms that convert at nothing. Proxy: the terms they pay for, and the terms whose landing pages carry a demo or trial form rather than a newsletter signup.
• Real search volume. Every published volume is modelled, including Google’s own rounded brackets. Proxy: your own Search Console impressions for the same query, which are counted rather than estimated.
• Their bids and budgets. Auction bids are private. Proxy: the top-of-page bid range in Keyword Planner as a market price, and your own impression share against them as a measure of relative pressure.
• Which prompts surface them in AI answers. Answers are generated per conversation and no library of them exists. Proxy: a fixed set of buying prompts you run yourself on a schedule, logged so the trend is visible.

Organic keyword sets move slowly, so a full comparison once a quarter is enough and anything more frequent produces noise. Paid keywords move weekly, cost money every day they run, and are therefore the half worth watching closely. Set a monthly check on the terms where you and a competitor overlap, and an immediate check whenever you lose a run of deals to the same name.

What is worth recording is not the ranking table but the change and what you decided about it: a new comparison page aimed at you, a new cluster around a use case, a shift in the modifiers they use. Those belong beside your positioning work rather than in a search report. Flares watches competitor pages and messaging continuously, which matters here because a new comparison page targeting your brand is worth knowing about the week it goes live rather than the quarter it starts ranking.