Say you run content for a mid-size fintech tool. You pull up your three named business competitors, list every article on their blogs, and build a spreadsheet of topics you haven't covered yet.

Three weeks later you've got forty new topic ideas and no obvious way to rank them. Half turn out to be pages one competitor wrote and quietly stopped updating in 2022. Another chunk are topics your own site already covers, just under different titles, so the "gaps" aren't gaps at all.

This is the failure mode almost every competitor content analysis falls into. It treats the exercise as cataloguing rather than diagnosis. The team lists who they compete with for revenue, not who actually owns the answers their audience is searching for, and those are frequently different lists. A community thread can outrank every vendor in the category. A media site with no product at all can be the de facto top result for the exact question a prospect asks before they ever consider buying anything. For that fintech tool, the real content competitor on "how to reconcile invoices automatically" might be a bookkeeping subreddit and an accounting-software blog three tiers removed from the battlecard, not the two named rivals sales tracks.

Cataloguing what exists also tells you nothing about whether it's worth beating. A topic with five competing articles might mean five thin, dated, badly structured pages that any decent piece could outrank in a month. A topic with zero competing articles might have zero because there's no search demand for it at all. Coverage on its own answers the wrong question. What actually matters is how deep the existing content goes, what format it takes, how fresh it is, and what's actually ranking or getting cited right now. Skip any one of those and the analysis produces a list that looks thorough and leads nowhere.

What the spreadsheet misses

A spreadsheet of competitor topics tells you what exists. It doesn't tell you what's worth beating, who actually owns the answer, or whether the gap is even real. Those are three separate questions, and most teams only ever answer the first one.

Take the fintech example again. On "PCI compliance checklist for small merchants," the top organic result is a payments-industry blog with no product at all, and AI answer engines cite a regulatory body's PDF before either named competitor's page. Neither rival is the content competitor there. The publisher and the regulator are, and no amount of staring at the battlecard would have surfaced that.

There's a modern wrinkle worth naming here too. Semrush's content gap research found that nearly 90% of webpages cited by ChatGPT sat outside Google's top 20 organic results for the related queries. If your analysis only checks who ranks in Google, you're missing a large share of what's actually getting surfaced to people asking these questions through an AI system instead of a search box.

The deeper issue is that a "gap" isn't always a missing topic. Ahrefs describes page-level gaps as situations where you and a competitor both cover the same subject, but their page ranks for far more keywords than yours does, which means the fix is expansion or a rewrite, not a brand-new article about something you've technically already published. A spreadsheet that only tracks presence or absence can't tell the difference between a genuine hole and a page that just needs to be better.

None of this is solved by collecting more topics. It's solved by scoring the ones you've already found against the right criteria, which is the harder and more useful half of the exercise.

Choose the right field

Who counts as a content competitor

A content competitor is whoever owns the answer your audience actually finds, not whoever shows up on your battlecard. Business competitors sell what you sell. Content competitors rank or get cited for the questions your audience is typing into a search bar or asking an AI model, and the overlap between those two groups is often partial.

Work backward from the questions instead of forward from the market map. Take the handful of queries that show up in sales calls and support tickets, and see who occupies the results and the AI-generated answers for them. A large publisher can own an informational query without selling anything close to your product. Semrush's guide to SEO competitor analysis makes a related point worth borrowing directly: the goal of finding these gaps is to build a content calendar from them, not a reference document nobody opens again.

More competitors in the field isn't automatically better. Filtering to the topics that show up across several genuine content competitors, rather than chasing every rival's one-off experiment, narrows a long list down to the ideas with real confidence behind them. Five to ten sites or sources that genuinely occupy the territory is usually enough. Past that point you're re-confirming coverage you already understand, not learning anything new.

It's also worth being honest about how research gets done now. HubSpot's 2025 State of Marketing report found that 33% of marketers use AI for research, the single most common AI use case in that survey. That's a reasonable way to speed up the first pass. It's a poor substitute for checking the actual SERPs, the actual content inventory, and the actual performance data behind whatever a model suggests. Treat AI-assisted research as a faster first draft of the field, not the final list.

Audit the gap

What should we score first

Score coverage, depth, format, freshness, and performance together, in the same pass, because any one of them alone gives you the wrong answer. A topic where everyone has published something looks crowded until you notice all of it is two years old. At that point it's a freshness gap wearing a crowded topic's disguise.

Dimension	What it checks
Coverage	Which subtopics competitors address at all
Depth	Whether the treatment is real or a passing sentence
Format	Article, calculator, template, or short video
Freshness	When the page was last meaningfully updated
Performance	What ranks, and what an AI engine actually cites

A lot of "coverage" in a spreadsheet is really just a mention, not a treatment. HubSpot's competitive-analysis guide recommends tracking indexed pages, estimated traffic, top-performing content, freshness, and keyword overlap as a scorecard, and that combination is doing real work: each signal catches something the others miss. HubSpot also points to Backlinko's analysis of WordStream, which held the number one spot for "keyword research" and drew an estimated 99,000 monthly visits by pairing a free tool with genuinely educational content. That's a format gap wearing a topic gap's disguise, the kind of thing coverage alone would never surface.

Sometimes the actual opening isn't a missing topic. It's a missing format, or a missing angle nobody bothered to bring to a subject everyone already writes about. Research that differentiates a piece, an original data point, a first-hand detail, an angle the rest of the field skipped, tends to matter more than whether the topic itself is technically uncovered. And working out where the genuine holes sit, as opposed to the ones a keyword tool flags automatically, is covered in more depth in finding content gaps than there's room for here.

Make the call

When should we skip a topic

Skip a topic when it's crowded with strong, current, well-structured pages and you have nothing genuinely original to add. Reading the field means separating four situations that all look, at first glance, like "a competitor covers this."

Crowded topics have multiple strong, current pages already fighting for the position. The honest move is usually to pass unless you're bringing something the field doesn't have, a dataset, a proprietary angle, a first-hand detail nobody else can claim. Thin topics have coverage but no real depth, several competitors touching the subject shallowly and nobody owning it, which is one of the more winnable positions in the whole audit because the bar to clear is low. Dated topics have a good page that's simply aged out, the field having moved since the last update, and a current, accurate rewrite can displace a stale incumbent faster than almost any other move available. Intent-misaligned topics are covered, sometimes covered well, just not for your reader's actual situation, like a regulated-industry buyer stuck reading a generic compliance overview written for nobody in particular.

Each of those four calls for a different action, and mixing them up is how teams end up publishing a fourth generic overview of a topic that already has three.

Once the field is sorted, sequence by value and winnability together rather than by whichever gap looks most uncontested on paper. Score each topic from 1 to 5 on both. A 1 for value means almost nobody searches for it or asks an AI model about it; a 5 means it comes up constantly in the query data behind your priority terms. A 1 for winnability means the field is dominated by deep, current, well-cited pages with no realistic angle against them; a 5 means the strongest existing page is thin, dated, or clearly written for the wrong reader. A topic that scores high on value and high on winnability, a thin or dated page sitting on a subject people genuinely search for, belongs at the top of the list, well ahead of a topic that scores high on value alone but is already dominated by three exhaustive, well-cited competitors. A wide-open gap with no real audience behind it isn't worth the hours it takes to write, however tempting the blank space looks.

It's worth resisting the pull to just copy whichever page is winning. Avoiding me-too content covers why matching a competitor's structure point for point tends to produce a page that ranks below the original, since neither search engines nor AI systems have much reason to promote a copy over the source they already trust.

How to Do a Competitor Content Analysis