You open Ahrefs, type your competitor's domain into Site Explorer, and fifteen minutes later you're staring at 12,400 organic keywords. Now what? Most people export the list, sort by volume, and start writing blog posts about whatever sits at the top. That's not analysis. That's guessing with extra steps.
A real competitor keyword analysis guide has to answer one question your competitors can't: where is the gap between what they rank for and what they actually deserve to rank for? Because those two things are rarely the same. In my own projects, roughly a third of the keywords a competitor "owns" are held by pages that don't really serve the search intent — they rank on domain authority alone. Those are the openings.
Key takeaways
- Keyword overlap is meaningless without intent mapping. Two sites can target the same phrase and want completely different visitors.
- Score every opportunity using volume, difficulty, intent match, and your current position — not volume alone.
- AI search surfaces (ChatGPT, Perplexity, AI Overviews) are now a keyword terrain that most guides ignore. Prompts that trigger competitor citations are as valuable as a ranking keyword.
- The output of an analysis is a decision, not a spreadsheet. If you can't name the three pages you'll build next week, you haven't finished.
- Free tools get you 80% of the way. The paid tier buys speed and historical data, not a different answer.
What competitor keyword analysis actually tells you (and what it doesn't)
Search for any "what is competitor analysis" definition and you'll get the same line: it's the process of identifying what your rivals rank for so you can find content gaps. True, and useless on its own.
Here's the part nobody spells out. A keyword list from a competitor tells you what worked for them at some point in the past. It doesn't tell you whether that content is still converting, whether the page is thin, or whether they've let it rot. I've audited sites where the top-performing page hadn't been touched in four years and was quietly losing positions every month. Copying that page is copying a dying asset.
What the data reliably reveals
- Which topics your competitor considers core to their business (the clusters they've invested in repeatedly)
- How broadly they've covered a subject — one page or a full hub
- Where they're weak: keywords with high volume but a generic, mismatched landing page
- Which search intents they're ignoring entirely
What it hides
Conversion data. Backlink quality behind each ranking. Whether the visitor actually stays. You'll never see a competitor's bounce rate in any tool, and that's exactly the number that would tell you whether a keyword is worth chasing.
So treat the export as a map of territory, not a list of targets. The targets you pick yourself.
The step-by-step workflow I use, from export to decision
Four steps. Each one filters the list down. If you skip one, you end up with a spreadsheet nobody acts on.
Step 1: Pull the right competitors, not the obvious ones
Your direct business rivals are the worst place to start. They usually sell the same thing to the same people, so they cover the same ground you already cover. Look instead at:
- Adjacent players — companies that solve a neighbouring problem for your audience
- The informational publishers in your niche, the blogs that rank for every "how to" question
- One large player outside your league, to see the ceiling of what's possible in the topic
I once built a content plan for a B2B tool by analyzing two competitors and one industry publication. The publication gave me 40 informational keywords the competitors had completely ignored. Six of those pages ended up driving more trial signups than the product pages. Informational traffic does convert — just not on the timeline people expect.
Step 2: Cluster before you evaluate
Raw keyword lists are noise. Group them by topic and by intent before you judge a single one. A keyword with 8,000 monthly searches split across four different intents is worth less than four keywords of 1,500 each with clean intent.
Tools like Ahrefs Keywords Explorer, Semrush, and SE Ranking all offer automatic clustering now. I still do a manual pass on the top 50 keywords, because automated clustering merges things that shouldn't be merged. It'll happily put "best CRM for small business" and "what is a CRM" in the same bucket. Those are two different pages.
Step 3: Score every opportunity
This is where most guides wave their hands. They say "prioritize" and move on. Here's the formula I've used on client projects for the past couple of years — it's not perfect, but it forces a decision:
Opportunity score = (search volume × intent match) ÷ (keyword difficulty × current position gap)
Intent match is a 1–3 multiplier: 3 if the keyword exactly matches a page you already have or plan to build, 1 if it's tangential. The position gap is how far you are from page one — if you're at position 40 and the top result is at 1, that gap is large (and it should lower the score, since it's a long climb).
Sound arbitrary? It is, a bit. But a rough number beats a vibe. When I first started scoring this way, I discovered that 60% of what I'd been planning to write had a score below 8 — meaning it would take months to rank and wouldn't match any page I actually wanted to build. Cutting that work freed up three weeks of editorial time.
Step 4: Check the AI surfaces
This step didn't exist five years ago. Today, a meaningful slice of your audience never sees a blue link — they get an AI-generated answer that cites three sources. If your competitor is one of those three and you're not, you've lost the click before the query even finishes.
Run the top 20 questions from your clustered list through ChatGPT and Perplexity. Note which competitors get named, and in what context. Then ask yourself: what content would I need to publish for an AI system to consider me an authority on this specific question? Usually it's not a longer post. It's a clearer, more structured one with a definite answer near the top.
Free tools vs paid tools: an honest comparison
You don't need a $200/month subscription to run a competitor keyword analysis. You need it if you're doing this weekly across dozens of domains. Here's how the options actually break down.
| Tool / method | Best for | Real limitation | Cost |
|---|---|---|---|
| Google Keyword Planner | Volume ranges and geo data | Bundles keywords into vague ranges; no competitor-level data | Free |
| Ahrefs / Semrush (paid) | Full competitor keyword exports, historical trends | Overkill for a single site; steep learning curve | ~$100–200/mo |
| Free Chrome extensions | Quick spot-checks on individual pages | Inaccurate difficulty scores; rate-limited | Free |
| Manual SERP analysis | Intent verification, spotting thin content | Slow; doesn't scale past ~30 keywords | Your time |
| AI prompt testing | Finding citation opportunities | No reliable volume data yet | Free–$20/mo |
My honest take: start free for one month. If you're running analysis for more than one site, the paid tools pay for themselves in saved hours — I tracked it once and a single afternoon of manual SERP checks replaced about four hours of paid-tool work. That math only makes sense if your time is cheap. Mine isn't.
Three mistakes that quietly ruin the analysis
Chasing volume instead of fit
A 50,000-search keyword that has nothing to do with what you sell will bring 50,000 people who leave in eight seconds. I've watched a client insist on ranking for a broad industry term for six months. It eventually hit position 4. Traffic went up 40%. Signups didn't move at all. The page was answering a question their buyers weren't asking.
Ignoring branded keywords
Your competitor ranks for their own name. You will never take that. Filter branded terms out of every export before you do anything else, or your list is inflated by 10–30% with pure noise.
Treating the export as the deliverable
The spreadsheet is not the work. The work is the decision that follows: which three pages you build, which two you update, which ten keywords you deliberately ignore. If your analysis ends with a file and no calendar entry, you've done research, not strategy.
A competitor keyword analysis template you can actually use
Here's the structure I put in front of clients. It fits on one page and forces the conversation.
- Competitor set: name, domain, why they're included
- Top 10 keywords they own, with intent labelled
- Keyword gaps: phrases with real volume where no page of theirs exists
- Weak spots: high-volume terms served by a generic or outdated page
- AI citation check: which prompts surface their content, and which don't
- Scored shortlist: your top 10 opportunities using the formula above
- And one line at the bottom: what we build first, and why
That last line is the only part anyone reads twice.
Where this is going next
The keyword list as a concept is slowly losing its edges. Search engines now render an answer, not a list of ten links, and the answer cites sources regardless of whether those sources targeted a specific phrase. That doesn't kill competitor keyword analysis — it changes what you're competing for. You're no longer fighting for a slot in a ranked list. You're fighting to be the source an AI system reaches for when someone asks a question you happen to answer better than anyone else.
Which means the old instinct — find the keyword, write the post, repeat — is becoming the slow way to lose. The faster path is to figure out which questions your competitors answer badly, and answer them so clearly that both a reader and a language model can quote you without hesitation. Start with the three highest-scored opportunities from your list. Ship them. Then look at what changed — not in your rankings, but in who's citing you.
