How to analyse a competitor without copying them

The first mistake in competitor analysis is choosing the competitors. Most teams list the companies they lose deals to. In search you are not competing with those companies — you are competing with whatever URL currently occupies the position you want, and that is frequently a magazine, a marketplace listing, a forum thread or a comparison site that does not sell anything you sell.

Step 1: find who you actually compete with

Take fifteen to twenty queries that matter to you — the ones where a ranking would produce a customer, not just a visitor. Search each one and record the domains in the top ten. Then count. The domains that appear repeatedly across your query set are your search competitors, in order of how much of your space they occupy.

Two things usually fall out of this immediately. A company you consider your main rival turns out to be invisible in search, which means their traffic comes from somewhere else and copying their site will teach you nothing. And a site you had never considered a competitor turns out to hold six of your twenty queries.

Step 2: analyse per query, not per company

A domain-level score tells you who is bigger. It does not tell you anything actionable, because the same competitor can be unbeatable on one topic and asleep on another. Work query by query, and for each one answer three questions about the page that ranks first:

1. What kind of page is it — a guide, a comparison, a category, a product, a forum thread? That is the intent shape the engine has decided this query wants.
2. What does it actually cover, section by section? List the h2s. That is the coverage bar you have to clear.
3. Why is it winning — the content, or the domain? A thin page from a strong domain is beatable with a better page. A genuinely excellent page from a strong domain is not, this year.

That last distinction is the whole point of the exercise. It is the difference between a query worth a week of work and one that will absorb a month and produce nothing.

Step 3: sort the gaps into three buckets

Bucket A — pages they have that you do not

The classic content gap. Useful, but treat it with suspicion: the fact that a competitor has a page does not prove the page works. Before adding it to the list, check that the page actually ranks for something. Copying a competitor's failures is a common and expensive way to look busy.

Bucket B — pages you both have, where they win

Compare the two pages directly against the query. Usually one of four things is true: they answer the question directly and you answer it in passing; they cover three subtopics you omit; their page is a different type (a comparison where yours is an essay); or their page is simply older and better linked. The first three are fixable this week. The fourth is a link problem, not a content problem, and no rewrite will solve it.

Bucket C — queries nobody answers well

This is the bucket that matters for a small site, and it is the one almost every analysis skips because tools do not surface it. You find it by reading result pages: queries where the top results are forums, aggregators, four-year-old posts, or broad pages that mention the topic in one paragraph. An unsatisfied query is worth more than a gap, because you are not fighting authority — you are filling a vacuum.

The query-selection method for small sites is mostly a method for finding bucket C systematically.

Step 4: check the structural reasons, not just the words

Sometimes the competitor is not writing better — they are structured better. Worth checking on the pages that beat you:

• How many internal links point at their winning page, and from where? A page linked from the homepage and six related articles carries authority a buried page does not.
• Do they have a hub page that links to a cluster of related pages? Topic clusters are the most commonly copied structural advantage, and the easiest to replicate.
• What structured data do they emit? An FAQ or product result takes clicks from positions above it.
• Is their page faster, or simply usable on a phone where yours is not?
• How recently was it updated, and did the content actually change or just the date?

Step 5: turn it into a plan you will actually execute

The output of a competitor analysis should be a short ordered list of pages to create or fix, each with a reason. If a row does not say why you can win it, it does not belong on the list.

1. Fix first: pages where you rank 8–20 and the winner covers three things you do not. Fastest movement available.
2. Then: bucket C queries, where the current results are weak. Slower, but genuinely winnable.
3. Then: bucket A pages that demonstrably work for a competitor and fit something you can credibly claim.
4. Not yet: anything where the winner is a strong page on a strong domain. Note it, revisit in six months.

The line between analysis and copying

Copy the structure: the intent shape, the subtopics a good answer covers, the fact that a comparison table belongs on the page. That is not plagiarism, it is reading what the query demands.

Never copy the content. Beyond the obvious reason, a page assembled from what competitors already say has no reason to outrank them — it is by construction a worse version of a page that already exists, and it fails exactly the test that decides whether AI-assisted content ranks: does this page contain anything that could only come from you?

How often to redo this

Quarterly for the full pass, and continuously for the ten or twenty queries you care most about — those you should be tracking anyway, because a competitor overtaking you on a money query is worth knowing about in the week it happens rather than at the next quarterly review.

Frequently asked questions