Competitor SERP Analysis: Who Really Ranks — and Which Competitors to Pick

Full transcript

1,261 words · script draft, under review.

Intro

Ask a business owner who their competitors are, and you'll get a confident list. The shop across town. The big brand everyone knows.

Then you look at the search results for the keywords they care about, and those names barely show up. Instead there's a niche review site, a marketplace and a forum, in the top ten again and again.

And if you then feed the biggest of those into a gap analysis, you get forty thousand keywords you "should" rank for. Almost none of them are yours to win.

In this episode, we fix both problems. First we find who really ranks across your keywords. Then we narrow that down to three to five competitors that are actually worth analysing.

Business competitors versus search competitors

There are two kinds of competitors, and they overlap less than you'd think.

Business competitors sell what you sell. Search competitors take the clicks you want. A review site that sells nothing is a search competitor. A forum is a search competitor. A big brand that never bothers with content might not be one at all.

For SEO analysis, the second list is the one that counts. And the only honest way to build it is to look at who actually ranks, across many keywords, not just the three you check by hand.

Share of top ten and visibility

Start with your keyword set. Ideally the clustered list from module C, the keywords that passed the gates in the last episode, plus the ones you already rank for.

Pull the top ten for each keyword, or reuse the snapshots you already have. One row per result: keyword, position, domain.

Now count. For each domain, in how many of your keywords does it appear in the top ten? Divide by the number of keywords. That's its share of top ten. A domain that appears for sixty of your two hundred keywords has a share of thirty percent.

Share is blunt. It treats position ten the same as position one, and a tiny keyword the same as a big one. So I add a second number.

Visibility. For each appearance, multiply the keyword's search volume by an expected click-through rate for that position, and add it up. Use your own CTR curve from episode eleven if you have one. Rough is fine. We use it to rank domains, not to forecast traffic.

Worked example: the long list

Here's our example shop, which sells coffee equipment. Made-up numbers, for illustration. Two hundred keywords.

At the top by share is a big marketplace, in the top ten for about forty percent of the keywords. Next, competitor-a, a niche review site, around thirty percent. Then a coffee forum, at about twenty. Competitor-b, a shop very much like ours, at fifteen. Competitor-d, another specialist shop, at ten. And competitor-c, the brand the owner was most worried about, at five percent.

Now visibility. The order changes. Competitor-a jumps to first place, because it ranks in the top three on the bigger keywords. The marketplace drops, because many of its appearances are at positions eight to ten. The forum drops a lot, because it mostly ranks on small questions.

Share tells you who is everywhere. Visibility tells you who is actually taking the clicks.

Why they win

A list of names is only half the job. For the top domains, write a short profile: why do they win?

Page types: reviews, product pages, comparison lists, tools. Use the domain types from episode twenty-nine and the page types from episode thirty. Coverage: one strong page ranking for many keywords is a different competitor than a hundred pages each ranking for one. Structure: do their ranking URLs follow patterns, like a page per brand or per use case? Authority: does this domain rank on almost anything, or because its pages fit the searches so well? Episode forty-two looks at that properly. And freshness: recent dates, new URLs appearing in your snapshots.

Keep the top fifteen or so domains as your long list, with the giants and aggregators flagged, not deleted.

Who ranks most is not who to copy

That long list answers the question "who is in the way". It doesn't answer "who should I learn from".

In a gap analysis you compare your site with theirs. If theirs is a hundred times bigger and covers twenty topics you'll never touch, most of its keywords are noise. And giants win for reasons you can't copy: brand searches, millions of links, years of history. Benchmark against them, and every gap looks like an authority problem.

A good competitor for gap analysis looks like a better version of you.

Four filters

So run every domain on the long list through four filters.

Filter one: niche relevance. You already have the data. How many of your keywords does the domain appear for, and how broad is the rest of the site? A site that shows up for a third of your keywords and does little else is a niche competitor. A site where your topic is one section among fifty is not.

Filter two: business model. A shop, a publisher, a marketplace and a forum win with different pages. Keep at least two that share your model.

Filter three: size. Near your size, or one step ahead. Not ten steps. A rough band is enough: smaller, similar, larger, giant.

Filter four is the exception rule for giants. Sometimes the giant really is your competitor. The test: does it show up on most of your keywords, with a page type you could build? A marketplace ranking with category pages across your set is a real competitor for a shop. A giant ranking with one random thread or one encyclopedia article is just in the way.

Worked example: the short list

Back to our coffee shop, still with made-up numbers. Six domains from the long list.

The big marketplace ranks with category pages on about forty percent of our keywords. It passes the giant exception. Keep it, labelled "page-type reference only".

The coffee forum ranks on many questions, but with threads. That's a weak-SERP signal from episode thirty-two, not a gap competitor. Out.

Competitor-b and competitor-d are shops like ours, a step ahead in size. Keep both. These are the core.

Competitor-a sells nothing, but wins the bigger keywords with comparison pages. Keep it as the content benchmark.

And competitor-c, the brand the owner fears, barely ranks. Keep it, but only as a business competitor, so it doesn't distort the numbers.

Notice what decided it. Not who was biggest, and not who appeared most. The forum appeared more often than shop D, and still didn't make the list.

The scorecard and the pitfalls

The output is a one-page scorecard. Each competitor gets a role, a model, a size band, and one sentence explaining why it's there. That sentence is your answer when someone asks "why these sites?"

The role decides how each competitor is used. Core competitors go into the keyword gap. Page-type references go into the page-type gap. Content benchmarks go into both, with lower weight.

Three pitfalls. Your keyword list decides your competitors, so make sure it covers what you actually want to rank for. Sometimes only one section of a large site competes with you, so look at the URL paths, not just the domain. And keep the list small. Three to five is enough. Every extra competitor adds export cost and noise.

Wrap-up

So: find your competitors from the SERP, measure share and visibility, profile why they win, then filter down to a short list with a role for each. The competitor kit, with the script and the scorecard, is linked in the description. Next episode starts the gap analysis module: content, page-type and structural gaps, and how the whole process fits together. See you there.