How do you identify your real SEO competitors?

In brief: a real SEO competitor is a domain that Google puts in front of you for your queries, not necessarily the one you name in meetings. Start from your real queries, validate them in results captured with constant settings, then sort the domains surfaced by tools rather than accepting them wholesale.

Ask a team who its competitors are and it will name a few without hesitation. Then open Google on the searches that keep the site alive, and you will often find a publication, a comparison site or a directory that nobody had mentioned. This method treats the question as an investigation: you collect clues, cross-check them and rule out false leads.

What is a real SEO competitor, and how does it differ from a business competitor?

An SEO competitor is a domain that ranks for the same organic keywords as yours. It is therefore fighting for a share of the same traffic from Google's natural results. This is the working definition used by Semrush's organic competitor reports, which list the domains ranking for the same keywords as the analysed site.

This definition is deliberately narrow. It measures neither commercial competition, nor margin, nor brand positioning. It simply tells you who shares your ground in search results (the SERPs, that is, Google's results pages).

SEO competitor or business competitor

A business competitor sells an offer comparable to yours to the same audience. An SEO competitor, by contrast, appears for the same searches, whether or not it sells anything similar. The two profiles often overlap, but not always.

A neutral example: a specialist publication, a comparison site or a marketplace can capture your traffic without selling the same offer. For a user typing a broad search, these pages sometimes answer better than a company's page. In terms of organic traffic, they are nonetheless competitors, and their editorial strategy looks nothing like that of a business rival.

Why the list in your head is misleading

The competitors you already know are not necessarily the ones Google pits you against. Your mental list reflects your market history, your trade shows, your tenders. Google, for its part, works query by query and by search intent, with no knowledge of your internal map of the market.

To correct this bias, the method follows four stages, in this order:

1. build a set of queries from real data;

2. identify candidate domains through keyword overlap;

3. check these domains in SERPs captured with comparable settings;

4. prioritise, then sort to exclude false competitors.

Prerequisites: what you need to have to hand

Gather these items before you begin. None is complicated, but missing any one of them undermines everything that follows.

• Access to the Search Console property of the site: this is the necessary condition for using your real queries.

• The target market: country, language, and the devices that matter for your audience.

• An organic competitor report tool (for example a Semrush-type report) or, failing that, manual checks in Google.

• A spreadsheet to record, for each capture, the query, the country or city, the language, the device and the date.

The process can be done free of charge, manually. Tools mainly speed up the calculation of keyword overlap between domains, which is time-consuming to reproduce by hand.

Step 1: build your query set from Search Console

Start by exporting what Google has actually shown you. In the Search Console Performance report, the "queries" dimension gives you the working base: it reflects the impressions and clicks observed for your site, not a theoretical list of every possible term.

Spotting expected and unexpected queries

Go through the list looking for two things. First, expected queries, those that match your offer and confirm your positioning. Second, unexpected queries, which reveal topics on which Google associates you with without your having chosen it.

Pay particular attention to queries with high impressions and a low CTR (click-through rate). They often signal an area where other results attract more users, and therefore ground where competitors deserve to be identified.

Next, group these queries by offer theme and by intent (finding information, comparing, buying). The aim is a strategic list of manageable size, not an export of thousands of rows.

Segmenting by country and device

Apply the Country and Device filters when these dimensions matter for your audience. The country dimension groups data by the country the search came from, and the device dimension distinguishes desktop, tablet and mobile. A site targeting France on mobile does not necessarily have the same rivals as a site aiming at several markets on desktop.

Beware of a common confusion: the country of the search is not the country of the company. A business based in France may draw most of its traffic from Belgium or Switzerland, and that is the market you need to analyse.

Filling in missing queries

Complete the set with the target topics and keywords you already know. Search Console does not tell the whole story, and this gap is documented.

Point to watch

Google excludes anonymised queries from the detailed tables; these are generally issued by only a few dozen users over two to three months (Google Search Central, 2022). They may still be included in chart totals when no query filter is applied. This mainly affects the long tail and small sites: a query's absence from your tables therefore does not prove that it generates no activity.

Step 2: identify candidate domains through keyword overlap

Run an organic competitor report in the tool you have available. It lists the domains ranking for the same organic keywords as your site. At this stage, you get candidates, not yet confirmed competitors.

Competition level or common keywords

Two columns look alike but do not say the same thing. Common keywords count the raw intersection between two portfolios. Semrush's competition level, by contrast, is calculated from the number of keywords each competitor has and the number of common keywords.

The consequence is concrete: two domains can share many queries without having comparable overall organic coverage. A very large generalist site overlaps with almost everyone without resembling you. This calculation remains specific to the tool's methodology, hence the check that follows.

Reading both rankings together

First sort the list by competition level, then by common keywords, and compare the two results. Domains that rank highly in both lists are your strongest candidates. Those that appear in only one call for closer examination: a large number of common keywords can mask a very partial similarity.

Bear in mind that some discovery tools rely on the common keywords found in Google's organic top 20. The scope of the analysis therefore depends on how the tool defines its results window.

Step 3: check the domains in SERPs with comparable settings

Open Google and capture the results for your strategic queries yourself. This is the step that turns a tool's list into observation. It has value only if the captures are comparable with one another.

Documenting the search context

Google varies its results according to location, language, device type, related results, recent searches and the time of the search. For each capture, record at a minimum:

the exact query, the country or city if relevant, the language, the device and the date. For a local query, tighten the geographic control: a difference of a few kilometres can change the list of players.

Then note, for each query, which domains and which URLs come up. Recording the URL and not just the domain will help you understand which page the competition is played out on.

Viewing a non-personalised version

Google offers a "Try without" link to view a non-personalised version of the results. Use it to limit the effect of your browsing history during the manual check.

Please note

Even without personalised recommendations, Search continues to use, among other things, location, language and device type. This check reduces one bias, but it does not produce a universal SERP. Personalisation does not necessarily affect every query or every result, either.

Step 4: prioritise domains and exclude false competitors

Now sort your candidates using your captures. Do not automatically keep every domain a tool surfaces: sorting is the part that calls for judgement, and it is what makes what follows reliable.

Ranking by recurrence

Rank the domains by how often they recur across your strategic queries. A domain present on almost all your high-value queries carries more weight than one that is highly visible on a single query. Then examine the detail of the common keywords and the URLs involved, query by query, to understand what the real meeting point is.

Setting aside off-topic domains

Exclude domains that are off-topic for your offer, the search intent or the target area, even with a high technical overlap. A site that shares your keywords but answers a different intent, or targets another country, is not taking the traffic that matters from you.

Finally, segment the final list into three families:

• direct competitors for the offer;

• editorial competitors or aggregators (publications, comparison sites, useful directories);

• very generalist domains, to be handled separately.

You end up with a shortlist that is justified and ready for in-depth analysis.

Worked example: an online accounting training site looking for its real competitors

This case is fictional and purely educational: no figure or domain below comes from real data. It serves to show how the four steps fit together.

From the query set to the final list

The site offers distance-learning accounting courses for a French audience. In Search Console, we extract the queries, group them by intent (learning about a profession, comparing courses, enrolling), then segment by France and mobile, the audience's dominant usage in this scenario. The unexpected queries reveal searches for qualifications the site had not targeted.

The organic competitor report puts a generalist directory at the top of the overlap list. It shares an enormous number of keywords with the site, but on simple information searches. Conversely, a training provider that is less visible in the tool appears on all the commercial-intent queries when the SERPs are captured, with the settings recorded in the spreadsheet.

The sorting therefore keeps the less visible provider as the priority competitor and demotes the directory. A raw ranking by volume of common keywords would have done the opposite.

Final decision table

Key takeaways

The aim is not the longest list, but the most reliable one. A few domains backed by dated records are worth more than a multi-page export that nobody has checked.

Special cases: sites with no history and Google's AI surfaces

A new site with no rankings

A new site has no ranking data yet, so Search Console cannot feed step 1. Start instead from a list of target keywords that represent your offer, your intents and the market you are actually targeting. Semrush states that its Position Tracking Competitors Discovery report can identify the main competitors from such a list.

This list must then be validated in the SERPs, as for any other site. It remains a hypothesis until you have observed who Google actually displays.

Competitors cited in AI answers

If Google's AI features are part of what you track, analyse the domains and pages cited in AI Overviews and AI Mode separately. Google states that these features may use "query fan-out": several related searches run across subtopics to identify more supporting pages while the answer is being generated (Google Search Central documentation, updated 15 May 2026). The set of links displayed may therefore differ from that of a standard web search.

This analysis complements the analysis of organic competitors without automatically replacing it. It only makes sense if these features actually appear in the market you are studying. Once your list is reliable, you can move on to a broader approach to competitive analysis SEO.

Common mistakes to avoid

• Limiting yourself to competitors you already know, or only to commercial competitors. You miss the media outlets and aggregators that capture your traffic.

• Keeping every domain a tool suggests without checking the SERPs or the intent. A tool calculates overlap, not relevance.

• Comparing SERPs captured in different contexts: different countries, devices, dates or levels of personalisation make the differences impossible to interpret.

• Confusing a large number of shared keywords with genuine proximity. Always read this number alongside the level of competition and the intent.

• Believing that a query missing from Search Console has no potential. Anonymised queries remain invisible in the detailed tables.

Practical tips for keeping a reliable list over time

Keep a dated spreadsheet of your records. Without a history, you won't know whether a domain is rising, disappearing or was never there: only comparing one period with another can tell you.

Review the list when your offer, your market or your strategic queries change. A new service, a geographical expansion or a shift in the dominant intent is enough to change your competitive landscape.

Start with 10 to 20 priority queries before expanding. It is better to have a controlled, well-documented base than a scope so broad that the records become incomparable.

Finally, distinguish between priority competitors and domains to monitor in the final list. This separation avoids spreading the analysis across secondary players, while keeping an eye on those that could rise. A solid competitor list is one that can be justified, domain by domain, with precise records and a clearly noted search context.