You are the person who signs off the competitor list, so you never open the page behind it. Four of those ten domains cannot be outranked by any article. How many of this quarter's briefs point at a slot that was never available?
The page above your article looks like a roster of competitors. On the query I pulled this week, four of the ten entries were not competitors at all: a video, two forum threads, a course catalog. You can write the best page in your category and none of them will move, because you cannot outrank them. The usual competitive analysis in SEO starts from a domain export that treats all ten as the same kind of thing, which is exactly where the budget goes wrong. Six of the remaining slots are winnable, and they are worth less than they were eighteen months ago. So which of the things standing on your page can actually move, and which ones have to be worked around?
I have built pages this way myself. A tool exported the column, I trusted the column, and the brief that came out of it was clean, specific, and aimed at a slot nobody could take. Four months later the piece was live, it was good, and it sat behind a Quora thread that had held the position since 2024.
Your SEO Competitor List Is Probably Wrong
The tool exported a column of domains. Everything in it ranks for something you rank for, and that is the only thing the entries share. It is a list of co-occupants, not a list of rivals. For the query "competitive analysis in seo" I pulled the live page on 27 October 2026 and counted what was standing on it.
Behind the AI Overview, the top ten organic results held six pages you could outrank and four you cannot touch: YouTube at 4, Quora at 5, a Coursera course listing at 6, and a second Quora thread at 7. Across all 19 organic entries in a depth-20 pull, the split is 13 to 6 in the same direction, so the unwinnable slots cluster at the top rather than spreading evenly. That is 40% of the visible top ten on one keyword, on one day. Treat it as a dated count for the query you are looking at, not as a law of the industry.
The page is also less stable than it looks. Seven of the ten top-ten slots on this read differed from a snapshot taken the day before, and neilpatel.com moved from 4 to 10. Size is not what defends this page one either. newtarget.com holds position 2 with 2,761 organic keywords and an estimated traffic value of $5,478, while quora.com holds position 5 with 31,046,101 keywords. Roughly 11,000 times the keyword footprint sits three slots lower.
So the column of domains resolves into four kinds of occupant, and only the first is a rival in any sense your competitor tool understands.
Occupant type | Can you displace it? | Correct action |
|---|---|---|
Rival page (agency, vendor or publisher blog, competitor product page) | Yes | Run the keyword and content gap work, then price the slot before you commit |
Marketplace (Amazon category, G2 or Capterra category, app store listing) | No | Get listed and optimize the listing; do not build a category page of your own |
Forum or UGC platform (Reddit thread, Quora answer, LinkedIn Pulse) | No | Answer as a participant and mine the thread for intent |
Directory or aggregator (industry directory, review site, another company's "best of" list) | No, but you can occupy it | Claim and complete the listing, then treat it as distribution |
Nobody needs to be told the first half of this. The AI Overview sitting above all of it already says the domains that rank "may differ from your direct business rivals", and the freshest guide on the sibling query says the same thing in its first step. Announcing the distinction changes nothing. Sorting each occupant into a class and attaching a different action to each class is where every one of them stops, including the guide published four days ago. That sorting is the part no occupant of this page does, and it is where a useful SEO competitive analysis diverges from a keyword-gap report.
Occupant Type 1: The Rival You Can Actually Outrank
Six of the ten slots here belong to pages a brand can beat. That is the class your competitor tool describes, and inside it the normal competitor analysis work applies: which rivals cover the query, what they cover that you do not, who links to them. Run that work on the other three classes and you will spend a quarter building a category page a marketplace has already won.
The freshest guide in this competitive set is also the most useful comparison. On the sibling query "seo competitor analysis", a guide from olostep.com published on 23 October 2026 sits at 7. It gets further than anything else on the page. It splits competitors into realistic, aspirational and giants to exclude, and it tells you to note where Wikipedia and Reddit appear but leave them out of direct benchmarking. Then it moves on.
Exclusion is a spreadsheet decision, and the tiering stops there. Classification asks a second question it never reaches: can this slot be displaced at all, and if it cannot, what do you do instead of benchmarking it? Leaving a forum thread out of your gap analysis is correct and incomplete. The thread still took a slot in the ten results your buyer scans.
Here the class boundary matters, because the price of getting it wrong now has a number attached. Ahrefs measured click-through at desktop position 1 with an AI Overview present as a decline of roughly 58% in 2026, against 34.5% in April 2025. If the slot you are chasing returns about half of what it returned last year, "can I outrank this" becomes a budget question rather than a reflex. The framework layer for deciding which rival is worth that spend is the competitor analysis framework, and it only starts once you know you are looking at a rival.
Occupant Type 2: The Marketplace That Will Not Move
The Coursera listing at 6 is a catalog page. Nobody wrote it to answer your buyer's question; it exists to sell seats in a course, and it ranks because thousands of course pages inside the same catalog point at it. Amazon category pages, G2 and Capterra category pages, and app store listings behave the same way. Their rank is fed by the inventory they contain.
You cannot outrank that by adding one more page to the inventory. A blog post does not feed the taxonomy, does not add a listing, and does not inherit the internal links that hold the category in place.
The layer is also consolidating, which changes what competing would even mean. G2 announced the acquisition of Capterra, Software Advice and GetApp from Gartner on 29 January 2026, closed it on 5 February 2026, and Gartner's 10-K filed on 12 February 2026 disclosed consideration of about $110 million. The combined network claims 6 million verified reviews, more than 200 million annual software buyers, 10,000+ vendors and 2,000+ categories. You do not outrank a $110 million consolidation. You get inside it, complete the listing, and make sure the fields a buyer filters on are accurate.
Handing over the listing work is a different job from keyword work. It belongs to whoever owns product marketing or partnerships, and the handoff is the split described in SEO and content strategy. Getting that ownership wrong is how a marketplace slot stays empty for two years while your blog keeps publishing at it.
Occupant Type 3: The Forum Is a Signal, Not a Competitor
Three Quora slots sat on this page: 5, 7 and 15. LinkedIn Pulse held 14, and that one is worth pausing on, because you do not outrank LinkedIn. You publish into it, and the post inherits the platform's domain rather than your own.
The structural version of this is Reddit. SE Ranking's analysis of the May 2026 core update put Reddit's share of top-3 positions at 10.24%, up from 8.56% after the March update and 9.19% after the one in December 2025. The same analysis notes the figure counts regular top-3 results only, and real presence is larger once Discussions and Forums features are included. The gains were not uniform either: smallest in real estate, healthcare and news, which is the part worth remembering if you work in a category Google treats as sensitive.
Read the three Quora slots as three different questions rather than one competitor. Each thread exists because someone typed a question the page one was not answering well, which makes the forum the cheapest intent research on the page. Answer there as a named practitioner, and take the phrasing back to your briefs.
Occupant Type 4: The Directory You Should Be Inside
The last class is the one teams treat as a consolation prize: industry directories, review aggregators and the "best of" listicles that somebody else maintains and ranks. On this page it is thin, because this is a methodology query rather than a buying query. On the category query you actually sell into, it is often half the page.
The action is to claim and complete the listing, then keep it current. That sounds like admin work until you look at where the shortlist actually forms. TrustRadius reported in 2026 that 74% of buyers consulted user reviews, that 83% shortlisted three or fewer products at an average of 2.7, and that 79% already knew about the product they eventually bought before they started researching. G2's July 2026 survey of 1,038 buyers ranked review sites first for shortlist influence at 38%, one point ahead of AI chatbots at 37%.
That one point is not a trend, and G2 says so itself: a single percentage point does not show a shift, and the chatbot figure measures a different behavior. Read it as a ranking of influence, because the ranking is what tells you where to spend the listing effort. By the time a buyer reaches your blog post, the shortlist may already be closed, and the list that closed it was somebody else's directory.
The Imitation Trap: Three Models That Work Elsewhere
A non-displaceable occupant is still a model you can copy by accident, and copying it is expensive. Three show up in most content plans I audit.
The first is the category taxonomy clone. It works for a marketplace because thousands of product pages sit underneath it and feed it links and inventory. On a brand blog you get an empty hub with twelve child pages and nothing to justify any of them.
The second is the annual listicle farm. It works for a publisher or an aggregator because the list is the product: it earns links, it updates cheaply, and it can rank for hundreds of "best X" variants at once. On a brand site the same list competes with your own products, and every update hands a rival a fresh mention. I have built this one myself, and it took two quarters to see that the traffic went to the comparison, not to us.
The third is the thin Q&A page written to intercept forum queries. It works on a UGC platform because the answers are the content and the platform absorbs the quality variance. Rebuilt on a domain with no community, it reads as filler, and the forum thread outranks yours anyway. Which is why answering inside the thread beats rebuilding the thread.
The bill for guessing shows up at portfolio level. Ahrefs' Content Explorer study by Tim Soulo, published 1 December 2023, found that 96.55% of pages in an index of roughly 14 billion received zero organic traffic from Google, and a further 1.94% received between one and ten visits a month. I cannot tell you how many were aimed at a competitor that could never be displaced. But a page built to imitate a marketplace is dead on arrival, and nobody finds out for a year.
Triage Your Keyword List in One Competitive Analysis Pass
The pass is mechanical once the four classes exist, and it is where a competitive analysis in SEO stops being a research exercise. For each keyword you care about, pull the live top ten and label every slot with one occupant type. Do not read the pages yet. The labeling alone tells you which pile the keyword belongs in, and that answer is most of what people are looking for when they search how to do seo competitor analysis. In Allable the live SERP pull sits in the same project as the keyword list the briefs come from, so the labeled page does not need its own export.
Then three piles, not four. Keywords whose top ten is mostly class one go into the normal content queue. Keywords held by a marketplace or a directory go to whoever owns the listings, with the query attached so nobody writes a page for them. Keywords where a forum or a video occupies the top three go to long-tail search, because the occupant will not appear on a more specific phrasing of the same question. Everything else gets dropped, and dropping is the part that saves the budget.
The budget logic has moved since this pass was last fashionable. SparkToro and Similarweb measured 68.01% of US Google searches ending without a click in the first four months of 2026, up from 60.45% in 2024. For commercial investigation queries specifically, the zero-click rate went from 42% to 51%, and AI Overviews trigger on 67% of them, the highest share of any query type. A displaceable slot is worth entering only when the query survives the overview. If the answer is already fully rendered above your result, you are buying the remainder.
Which route to take comes down to one condition. If the non-displaceable occupant sits above position three and the intent is informational, write for a different query, because the slot is not coming back. If the occupant is where the shortlist forms, buy nothing and get listed instead. That is the same split a B2B SEO strategy has to make every quarter, and it is much cheaper when your list arrives already labeled.
One more version of the same decision catches teams who did everything else right. A keyword can have a displaceable top three and still be a bad target, because the overview answers the question before the click happens. Pull the overview before the brief, not after the draft. If your own content is already summarized above the first result, you are competing for a citation rather than a visit, and those are two different budgets.
What changes in your roadmap is smaller than it sounds. Fewer new articles aimed at head terms a marketplace owns. An owner and a deadline for the listings nobody has touched. A content queue that only contains queries where the top ten is winnable. Watching competitor moves still matters, on a different cadence, covered in competitive intelligence analysis.
The page that started this competitive analysis in SEO was a mix of an agency post from 2022, a guide four days old, a course catalog and three forum threads. Nothing on it was defending the query with authority. What kept the top of it stable was category rather than strength: four slots nobody could enter. So which of your keywords is being held by something you cannot outrank, and how long have you been writing at it anyway?
Label the Page Before You Brief It
Pull one live top ten for a keyword you are about to brief and label every slot with its occupant type before anyone writes an outline. Allable keeps the SERP read, the keyword list and the brief in one project, so the label survives the handover and the query nobody can win leaves the queue early. The 7-day trial is at studio.allable.ai.
