Competitive keyword research is the process of finding which searches other sites already win, then deciding which of those searches your site can realistically compete for. The second half is where most people go wrong. A list of competitor keywords is easy to collect. Knowing which ones to pursue, and which to leave alone, is the actual skill.
This guide walks through a method you can run with free tools: identify the right competitors, collect candidate searches, read the results pages, and turn what you find into a short, ranked list of articles to write. The method is a recommendation, not an industry standard, and every judgement call in it is something you can check against your own Search Console data.
Start with search competitors, not business competitors
The companies you compete with for customers are not always the sites you compete with in Google. A local dental clinic may lose sales to the clinic down the road, but the pages outranking it for "how long does teeth whitening last" may be a health publisher, a dental supply brand, or a forum thread. Those are your search competitors for that query.
Build two lists:
- Business competitors: the companies a customer would compare you with when buying.
- Search competitors: the sites that actually appear when you search the questions your customers ask.
Business competitors give you a starting vocabulary: how they name their services, which problems their blog covers, how they structure their categories. Search competitors tell you what you are up against for each individual query. Keep the lists separate, because a keyword can be worth targeting against one group and hopeless against the other.
Collect candidate searches with free tools
You do not need a paid platform to get a workable first list. Several free sources reveal what competitors target and what people type.
Competitor site structure. Open a competitor's blog index, resource section, and navigation. Page titles and headings reveal the searches they are aiming at. If the site publishes an XML sitemap (often at /sitemap.xml, though the location varies), you can scan every URL slug in one pass. Slugs are often close to the target keyword.
Search results for your own seed terms. Type your core service or product terms into Google and note which sites appear repeatedly. Repeated appearance across several related searches is a useful signal that a site covers the topic broadly.
Autocomplete and related searches. Typing a seed term into the search box and reading the suggestions shows how people phrase the question. The "related searches" block at the bottom of results and the "People also ask" box (where it appears) extend the list. These are phrases real searchers use, which makes them good candidates even before you know anything about difficulty.
Your own Search Console. If your site already has impressions, look at the queries it appears for. For each one, search the phrase yourself and note which competitors share the results page with you. That gives you a starting list built from your own data rather than from guesswork. If you use any third-party tool that reports competitor traffic, check how the tool describes where its data comes from before you rely on it.
Put everything into one spreadsheet with a column for the phrase, where you found it, and which competitor (if any) appears for it. Do not filter yet.
Read the results page before you judge the keyword
A keyword is only competitive or easy relative to the pages currently ranking for it. A difficulty score from any tool compresses that picture into a single number. Looking at the results yourself takes little time per query and shows you what a score hides. For each candidate phrase, search it and note:
- What kind of page ranks. Are they blog articles, product pages, category pages, tool pages, videos, forum threads, or directories? If every result is a product page and you plan an article, the intent probably does not match.
- Who owns the results. Large publishers, marketplaces and well-known brands, or small niche sites like yours? A results page made entirely of very established sites is a harder target for a new website than one that includes a few small, specialised sites.
- How well the top pages answer the question. Are they thorough and current, or thin, dated, or only loosely on topic? Weak answers at the top are an opening.
- Whether the result is specific to your situation. Check whether searches that include a location, a customer type, a product variant, or a precise problem show different, weaker or fewer strong pages than the broad head term.
- What else occupies the page. Ads, maps, answer boxes and video carousels can push organic results down. A query can look attractive and still send little traffic to the pages listed.
The point of this inspection is intent fit and realism. The sections below turn it into a decision.
A simple triage for each keyword
The following is an author-proposed framework, not an industry rule. Sort each candidate into one of three groups after you have looked at the results page.
| What you see in the results | Reasonable decision |
|---|---|
| Page types match what you would write, and some results come from small or specialised sites | Write it; likely worth testing |
| Page types match, but the results are dominated by very established sites with comprehensive pages | Postpone; look for a narrower version of the same topic |
| Page types do not match your planned content (for example, all product pages or tools) | Skip, or build the page type that Google is rewarding if it fits your business |
Narrowing is the most useful move. If "dental implants" is dominated by large health sites, a phrase like "dental implants cost: what's included in the price" asks a more specific question and may have a different mix of competitors. Whether it does is something you confirm by searching, not assume.
Check whether competitors are covering topics you have not
Once you have the list, compare it against your own site. The question is simple: which questions do competitors and search results answer that you do not?
Do this by listing your published pages and matching each to a phrase in your sheet. Phrases with no matching page are your gaps. Then ask of each gap:
- Is this a question our customers actually ask before buying?
- Can we answer it from facts we know to be true about our business?
- Does it lead naturally toward something we sell?
A gap that fails the third question may still be worth writing for awareness, but rank it below those that do. A gap that fails the second question is a risk: if you cannot back up the claims, a competitor's article is not a template to copy.
Use competitors as input, not a blueprint
Competitive research goes wrong in a few predictable ways.
Copying structure and claims. Reading a ranking article to understand what searchers expect is reasonable. Rewording its points, statistics or prices into your own post is not. You inherit its errors, add nothing new, and any figures you repeat should be traceable to a primary source you have checked yourself.
Assuming a competitor's ranking page is working for them. A page may rank and still bring in few customers. You cannot see their conversions. Judge keywords by whether your own audience would act on the answer.
Chasing the highest-volume phrases. If you search a broad head term and the results page is dominated by established sites, test a group of specific questions you can answer well instead. Then check in Search Console which of your published pages gain impressions first, and let that guide where you invest next.
Ignoring intent differences within one topic. "Invisalign vs braces for adults" and "Invisalign cost" share a subject but serve different decisions. Each deserves its own page and its own look at the results.
Researching once. Results pages change as competitors publish and update. Revisit your shortlist on a regular schedule and look at whether the ranking page types or sites have shifted.
Turn the research into a short plan
At the end you should have a ranked shortlist rather than a sprawling spreadsheet. A practical register has these columns:
- The target phrase
- The intent you think it serves (learn, compare, buy, fix a problem)
- What ranks today (page types and site types)
- Whether you have a page for it
- The facts from your own business you can use to answer it
- A decision: write, refresh an existing page, postpone, or skip
Here is a hypothetical example for a small accounting practice, included only to show the shape. The phrase "how to register as self-employed" shows mostly government and large guide sites; decision: postpone. The phrase "self-employed bookkeeping for freelance designers" shows a mix of small sites and one large one; the practice serves freelance designers; decision: write. None of these details are real results. They illustrate how the results page changes the decision.
Before you publish, set a baseline. Note the page's current impressions, clicks and average position in Search Console (zero is a valid baseline for a new page) so that you can tell later whether the competitor analysis paid off. Review after a period you choose in advance, and compare against your baseline rather than against a competitor's visible position.
Where an automated workflow fits
Manual competitive research is worth doing once to understand your market, but it is slow to repeat for every article. If your bottleneck is turning researched searches into published pages, a workflow tool can take over part of the loop.
Pentra reads your website, drafts a business profile that you confirm, and researches keywords people search for. It writes articles from those confirmed facts, runs an independent fact check, and publishes to WordPress, Shopify, Webflow, Ghost or your GitHub site on the schedule you choose. You pick a pace from 1 to 21 articles a week, or approve every article yourself. It also opens each published page to confirm the article is live, improves pages that sit close to page one, and reports clicks and positions from your Google Search Console. It does not replace the judgement in this guide about which competitors matter to your business; it handles the research-to-publication work once you are clear on your market. Pentra has a free plan with 3 articles a month and no credit card required.
Try Pentra to see how researched topics become fact-checked articles on your site.