How to Analyze Competitor Keywords: A Step-by-Step SEO Guide

Last month a reader sent me a screenshot of her analytics. Traffic was up 40%. Revenue was flat. "I'm ranking for everything my competitors rank for," she wrote, "so why isn't it working?"

The answer sat in her keyword list. Of the 300 terms she'd pulled from three rival sites, roughly two-thirds were informational queries with zero commercial intent—people looking for definitions, not solutions. She had spent four months chasing traffic that would never convert. That's the trap with competitor keyword analysis: it's easy to collect data, brutally hard to interpret it.

This is the process I actually use, including the parts the tool vendors won't tell you about.

How to analyze competitor keywords: the short version

You're looking for three things. Keywords your competitors rank for that you don't. Keywords where you both rank but they consistently sit above you. And keywords where the whole niche is under-served—thin content, outdated pages, no clear answer.

That third category is where the money is, and almost nobody talks about it because it doesn't show up in a gap-analysis report. Tools compare your profile to a rival's profile. They don't tell you that your competitor's top-ranking page for a high-value term is a 900-word blog post from years ago that nobody has updated.

Why most people get this backwards

The instinct is to start with volume. Big numbers feel like opportunity. But a keyword list sorted by volume is a list of the hardest fights in your market, ranked by how badly everyone else wants them.

I made this mistake on a B2B project. I built a content plan around 18 high-volume head terms—things like "project management software" and "team collaboration tools." Eleven months later, two of those pages were ranking. Total conversions: four. Meanwhile, a single low-volume page targeting a specific integration query brought in 31 signups in six weeks. I'd almost skipped it because the tool showed 90 monthly searches.

Volume is a tiebreaker. Intent and gap are the main event.

The step-by-step method I actually run

Here's the sequence. It takes about two hours per competitor once you know the tools, and you can do most of it on free plans.

1. Pick the right competitors—not the obvious ones

Everyone picks the market leader. Wrong move. The leader's keyword profile is massive and mostly irrelevant to you: brand terms, PR-driven placements, pages that rank on domain authority alone rather than content quality.

What you want:

• Two direct competitors of similar size to you
• One larger player, to see where the category is heading
• One or two adjacent players—companies that solve a neighboring problem and occasionally cross into your space

That last group is underrated. Adjacent competitors reveal keywords you'd never think to target because they're slightly outside your mental model of your own product.

2. Pull their ranking keywords—and filter hard

Any free competitor analysis tool will show you a rival's keyword list. Some show a capped number, which is fine—you don't need everything.

Set your filters to: position 1-10, and a minimum volume that makes sense for your niche. In B2B, that might be 50 searches. In consumer, 500.

Then export. Now comes the part that matters: sort by intent, not volume.

Intent type	What it looks like	Conversion potential	Should you target it?
Informational	"what is X", "how does X work"	Low	Only if it feeds a funnel or builds topical authority
Commercial investigation	"best X for Y", "X vs Z"	High	Yes—this is usually your best hunting ground
Transactional	"X pricing", "buy X", "X alternatives"	Very high	Yes, but expect the hardest competition
Navigational	competitor brand names	N/A	Skip—you'll rarely win these

In my experience, commercial investigation queries convert 3-5x better than informational ones for most products. That ratio holds whether you sell software, services, or physical goods. It's not a law of nature, just a pattern that keeps showing up.

3. Run the overlap analysis—and read it honestly

Take your list and mark every keyword where you already rank somewhere. Then split into three buckets:

1. You rank 1-10, they rank above you. These are refinement jobs. Look at their page: is it more recent, better structured, does it answer the query faster? Sometimes you just need to update a page and add a section.
2. You rank 11-30. These are your cheapest wins. You're already visible; a content refresh or a few internal links can push you into the top 10.
3. You don't rank at all. Real gaps. These need new content, and they need prioritization.

Bucket two is where I'd put my first hour of work every single time. It's boring. It's not glamorous. It works.

Can you check competitors' Google Ads keywords?

Not directly. Google doesn't publish which keywords an advertiser bids on, and no tool has access to that data. Anyone claiming otherwise is guessing.

What you can do is infer. Look at which queries trigger a competitor's ad when you search them. Note the ad copy—it often names the exact pain point they're bidding against. And check their landing pages, because paid traffic almost always routes to a page built around a specific keyword cluster.

That's inference, not extraction. Treat it accordingly, and don't build a strategy on a guess you can't verify.

Prioritizing what you actually found

A raw gap list for even one competitor can run to hundreds of rows. Nobody executes a hundred-page plan. You need a filter that reduces the list to something a human can finish.

My rule: score each keyword on three factors, roughly equally weighted.

• Relevance. Would someone searching this plausibly buy what you sell? If the honest answer is "maybe in six months with a lot of nurturing," drop it.
• Difficulty against your actual site. Not the tool's abstract score—your realistic chance. A site with thin authority shouldn't be targeting the terms the market leader owns.
• Content gap. Does the current top result do a bad job? Are the top 10 full of forum threads and outdated posts? That's a signal the query is winnable.

Three criteria, ranked. That third one is the tiebreaker, and it's the one most people skip entirely.

The AI search layer nobody is checking

Here's the piece that's new, and it changes the calculus.

When someone asks an AI assistant "what's the best tool for X," the answer cites sources. Those sources get visibility your analytics may never attribute properly, because the click often doesn't happen. The user reads the summary and moves on.

So run a simple check. Take your competitors' best-performing queries—the ones driving their organic traffic—and ask an AI assistant those same questions. See who gets cited. See whether your brand appears at all.

On a project last year, a rival was cited in AI answers for eight of the ten queries that mattered most in our category. We appeared in one. That single data point reshaped our entire content plan for the next two quarters—we stopped optimizing for a ranking position and started optimizing for being quotable: clear definitions, specific numbers, direct answers near the top of the page.

This isn't a replacement for traditional keyword analysis. It's an overlay. But ignoring it in 2026 is like ignoring mobile in 2012.

Free tools and what they actually give you

You don't need a paid subscription to start. Here's the honest breakdown.

What free gets you

Most keyword tools offer a limited free tier—a capped number of searches per day, or a capped keyword list per domain. That's usually enough to analyze three or four competitors if you pace yourself.

SpyFu is the one people mention most for competitor research specifically, largely because its whole positioning is built around competitor spying rather than your own site's performance. It shows historical data, which is genuinely useful for spotting whether a competitor is gaining or losing ground on a term. The free version limits how much you see, but the direction of travel is often enough to be informative.

Google Search Console—your own data, not competitors'—is free and permanently underused. It shows you queries where you rank 11-20, which is exactly the bucket-two gold I mentioned earlier. I've met people paying for tools while ignoring this.

Where paid earns its keep

Paid plans matter when you need bulk exports, historical trend data, or the ability to track a large competitor set over time. If you're running this analysis once and acting on it, free tiers are fine. If you're doing it monthly across ten rivals, you'll want the paid version and you'll resent the price less than you expect.

The tool matters far less than the filtering discipline. I've seen better strategic decisions come out of a spreadsheet built on free-tier exports than out of a $200/month subscription nobody knew how to read.

A real example, start to finish

Let me make this concrete, because abstract advice about keyword research is nearly useless.

A client sold scheduling software to small clinics. Their three competitors were pulling steady organic traffic. We exported:

• Competitor A: 1,400 ranking keywords in the top 10
• Competitor B: 620
• Competitor C: 2,100

After filtering to commercial-investigation intent with realistic difficulty, the combined list dropped to 47 terms. Of those, 19 were keywords the client ranked for in positions 11-25. Nine were complete gaps. The rest were navigational or irrelevant.

We spent the first month on the 19 near-misses—refreshing pages, adding comparison sections, tightening internal links. Eleven of them moved into the top 10. The remaining eight we revisited the following quarter.

Then we built nine new pages for the gaps. Four of them performed within three months. Three took longer. Two still haven't moved, and I'll be honest: I don't fully know why. Sometimes a page just doesn't land, and no amount of analysis explains it.

Net result over two quarters: organic sessions up 62%, and—more importantly—demo requests from organic up from roughly 8 per month to 23.

Mistakes I made so you don't have to

I've botched this enough times to have opinions about where it goes wrong.

Chasing a competitor's whole profile. Early on I'd try to match a rival keyword for keyword. That's how you end up writing about topics you have no business writing about, and diluting the parts of your site that actually convert. You're not trying to become your competitor. You're trying to find the specific terms where you can win.

Trusting the difficulty score as gospel. Those scores are modeled, not measured. They're directionally useful and frequently wrong at the margins. The only real test is publishing and watching.

Ignoring the pages that already exist. New content is more fun than fixing old content. It's also usually the worse investment.

Forgetting that competitors can read your results too. The moment you start ranking, you become someone else's competitor to analyze. That's fine—just don't build a strategy that collapses the second a rival notices.

Where this leaves you

The tools have gotten better every year. The lists have gotten longer. What hasn't changed is that the value isn't in the data—it's in the filtering, the prioritization, and the willingness to spend an hour improving a page that ranks eleventh instead of writing a shiny new one that ranks nowhere.

If you take one thing from this: your best competitor keywords aren't the ones your rivals are winning. They're the ones your rivals are winning badly—thin pages, stale answers, no clear response to what the searcher actually wants. Those are sitting in your export right now, probably sorted somewhere near the bottom, ignored because the volume looks small.

Start there. The numbers will follow.