SEO Keywords: How to Find and Use Them Without Wasting Months on the Wrong Ones

TL;DR

• Most keyword research guides still teach volume-first thinking, but 58.5% of US Google searches now end without a click, meaning high search volume doesn’t guarantee high traffic anymore.
• AI Overviews reduce click-through rates for the top-ranking page by 58%, according to an Ahrefs study of 300,000 keywords from December 2025, so the keywords you pick matter more than ever.
• Use the Intent-Revenue Stack (a 3-filter framework) to evaluate keywords by revenue proximity, AI vulnerability, and content format fit before you even glance at search volume.
• Long-tail keywords convert at roughly 2.5x the rate of head terms and are far less likely to trigger AI Overviews that steal your clicks.

I spent the first half of 2024 doing keyword research the “right” way. I built a gorgeous spreadsheet. Thousands of keywords, sorted by monthly search volume, color-coded by keyword difficulty. I followed the playbook every guide teaches: start with seed keywords, expand with a tool, pick the high-volume terms with manageable difficulty, write content.

Six months later? Most of that content was pulling single-digit visits per day. Not because it was bad. Because I’d picked keywords where Google was answering the query itself, right there on the results page, and nobody needed to click through to my site.

That spreadsheet taught me something the hard way: the volume-first approach to keyword research is now a fast path to wasting time. The game has changed, and the standard advice hasn’t caught up. In this article, I’ll show you the framework I use now, one that filters for keywords that actually drive revenue, not just impressions.

Why the Standard Keyword Research Playbook is Broken

Here’s what almost every keyword research guide tells you: find keywords with decent volume and low difficulty. Rinse and repeat. That advice made sense when Google was a list of 10 blue links and every click went somewhere.

That’s not what Google looks like anymore.

A Semrush and Datos study found that 58.5% of US searches and 59.7% of EU searches ended without any click in 2024. Zero clicks. The user typed something, Google answered it, and that was it. And that number keeps climbing as AI Overviews expand to more than 200 countries and 40 languages.

But it gets worse for certain types of keywords. An updated Ahrefs study from February 2026 analyzed 300,000 keywords and found that AI Overviews now reduce the position-one organic CTR by 58%. For every 100 clicks you used to get by ranking first, Google keeps 58 of them.

So when you pick a keyword purely because it has 10,000 monthly searches, you need to ask: how many of those searches will actually result in someone visiting your page? If an AI Overview sits on top of that results page, the answer might be “almost none.”

Watch Out: Search volume numbers in keyword tools represent how many times a query is typed, not how many clicks are available. A keyword showing 5,000 monthly searches might only produce 800 actual clicks to organic results if AI Overviews or other SERP features answer the query directly.

What Are SEO Keywords, Really? (A Definition That Actually Helps)

SEO keywords are the specific words and phrases people type into search engines (or speak to AI assistants) that you can target with content to earn visibility and traffic. That’s the textbook answer.

But here’s a more useful way to think about them. Keywords are evidence of problems. Every keyword is someone articulating a need, a question, a frustration, a desire to buy something. The keyword “best CRM for small teams” isn’t just three words with a search volume attached. It’s a person who runs a small team, is unhappy with how they manage contacts, and is actively shopping for a solution.

When you start seeing keywords as problems instead of search volume numbers, everything about how you pick them changes. You stop asking “which keyword has the most volume?” and start asking “which keyword represents a person I can actually help, and who might actually become a customer?”

Think of it like fishing. Volume-first keyword research is like picking the lake with the most fish, regardless of whether those fish are the species you want. Intent-first research is picking the spot where your exact fish are biting. Smaller pond, but you’re actually catching something.

The Intent-Revenue Stack: A Framework for Picking Keywords That Pay

I call this the Intent-Revenue Stack because it forces you to evaluate every keyword through three filters, in a specific order, before you look at volume. Most people start at the bottom of this stack (volume). You should start at the top.

Filter 1: Revenue Proximity. How close is this keyword to a buying decision? A keyword like “what is project management” is far from revenue. “Best project management software for agencies” is one step from a purchase. “Monday.com vs Asana pricing” is practically standing at the checkout counter. Prioritize keywords where the searcher is closer to spending money.

Filter 2: AI Vulnerability. Will Google’s AI Overview eat your clicks? You can check this manually by searching the keyword and seeing if an AI Overview appears. Or use a tool like Ahrefs’ SERP features filter or Semrush’s AI Overview tracking to check at scale. If a keyword triggers an AI Overview that fully answers the query, that keyword is “AI-vulnerable” and you should deprioritize it unless you have a strong brand-visibility reason to target it.

Filter 3: Content Format Fit. Can you create something genuinely better than what exists? Look at the top 5 results. Are they thin listicles you can outdo with real depth? Or are they 5,000-word monsters from sites with massive domain authority? If you can’t realistically create a page that’s meaningfully different (not just longer, but different), skip the keyword.

Only after a keyword passes all three filters do you check volume. And here’s the uncomfortable truth: volume becomes the least important metric.

Filter	Question to Ask	Red Flag	Green Flag
Revenue Proximity	How close is this to a purchase?	Pure informational, no product tie-in	Commercial or transactional intent, clear buyer signal
AI Vulnerability	Does an AI Overview answer this fully?	Full AI Overview with citations, query fully resolved	No AI Overview, or AI Overview is vague and incomplete
Content Format Fit	Can I create something meaningfully different?	Top results are from Wikipedia, major publications, or massive sites	Top results are thin, outdated, or missing a specific angle
Volume (last!)	Is there enough demand?	Under 50 searches/month with no related cluster	Any volume above 50 if the keyword passed the first three filters

How to Actually Find Keywords Worth Targeting

Enough theory. Here’s the process I use now, step by step.

1. Start with your customer’s problems, not seed keywords. Talk to your sales team. Read your support tickets. Look at the questions people ask in your industry’s Reddit communities. Write down 15 to 20 actual problems your customers face. These problems become your keyword seeds, and they’re far better seeds than generic terms you’d pull from a tool.

2. Expand each problem into keyword clusters using a research tool. Plug each problem phrase into Semrush’s Keyword Magic Tool, Ahrefs Keywords Explorer, or even Google’s free Keyword Planner. But instead of sorting by volume, sort by intent. Group commercial and transactional keywords together first. These are your priority targets.

3. Run every priority keyword through the AI Vulnerability check. Search each keyword on Google. Note whether an AI Overview appears and how thoroughly it answers the query. Keywords where the AI Overview is incomplete, wrong, or absent are gold. Keywords where the AI Overview perfectly answers the question are probably not worth a dedicated page.

4. Check for content format fit. For each keyword, examine what ranks on page one. Ask yourself: can I bring something to this topic that doesn’t exist yet? Original data, a unique framework, a specific audience angle, first-hand testing results? If the answer is “I’d basically be rewriting what’s already there,” move on.

5. Build topic clusters, not keyword lists. Group related keywords into clusters where one pillar page covers the broad topic and supporting pages cover specific subtopics. This builds what Google calls topical authority, which matters more than targeting individual keywords in isolation.

“I’m guessing a lot of folks are throwing away time, money, and SEO expertise chasing traffic for terms that are never coming back. But, as marketers always say, ‘it depends!’ Check your results.”

— Rand Fishkin, CEO of SparkToro (via Search Engine Land)

Fishkin’s point is worth sitting with. The keywords that worked for you in 2023 might be dead today, and the tools won’t always tell you that. You have to look at actual clicks in Google Search Console, not just impressions or estimated volume.

Where and How to Place Keywords (Without Making It Weird)

Finding good keywords is half the job. Using them well is the other half. And “using them well” doesn’t mean stuffing your target phrase into every paragraph. It means placing keywords where they do the most work for search engines while keeping the content natural for humans.

Here’s where your primary keyword should appear (at minimum): in the title tag, the H1, the first 100 words of your page, and the meta description. Your secondary and related keywords should appear in H2 and H3 subheadings, in body paragraphs where they fit naturally, and in image alt text.

But here’s what most guides skip. With AI search engines like ChatGPT, Perplexity, and Google’s own AI Overviews now pulling specific passages from pages to generate answers, how you structure those passages matters as much as where you place keywords.

Each section of your content should be self-contained, meaning a single paragraph or section should make complete sense if someone (or some AI) extracts it without any surrounding context. Lily Ray, VP of SEO Strategy and Research at Amsive, has emphasized that what we used to call “chunking content” for AI is really just good formatting with clear headlines, something SEOs should have been doing all along.

This means: don’t bury your keyword-rich answer four paragraphs into a section after a long windup. Lead with the answer. Expand after.

Pro Tip: Write every key paragraph as if it could be the featured snippet or the AI-extracted answer. Front-load the definition or answer in the first sentence, then add context. This pattern works for both Google’s traditional featured snippets and AI Overview citations.

Long-Tail Keywords: Where the Real Money Hides

Everyone says “target long-tail keywords.” Few people explain why they’re actually more valuable now than they were three years ago.

Long-tail keywords convert at roughly 2.5x the rate of head terms, according to multiple conversion studies. That’s always been true. What’s new is that long-tail keywords are also less likely to trigger AI Overviews that steal your clicks. Why? Because they’re specific enough that Google’s AI often can’t generate a comprehensive answer for them. The query is too niche, too context-dependent.

Here’s an example. “CRM software” triggers a massive AI Overview with comparisons, features, pricing, the works. But “CRM software for real estate teams under 10 people” probably shows regular organic results because there isn’t enough consolidated content for Google to synthesize an AI answer.

And that niche query? The person typing it knows exactly what they want. They’re not browsing. They’re buying.

Meanwhile, Ahrefs’ widely-cited study found that 96.55% of all web pages get zero traffic from Google. Most of those pages are targeting broad, competitive head terms where they never had a realistic shot at ranking. The math favors going narrow.

The Keyword Mistake That Wastes the Most Time

I want to address one specific mistake because I see it constantly, and it’s not the obvious ones like keyword stuffing or ignoring search intent. Those are table stakes.

The biggest time-waster is treating keyword research as a one-time project instead of a continuous feedback loop.

Here’s what happens. A team spends two weeks building a keyword list, writes 20 pieces of content over three months, then never revisits the keyword strategy. Six months later, half those keywords trigger AI Overviews that didn’t exist when the research was done. Three of the keywords have shifted intent (Google now shows product pages instead of blog posts). Two keywords have new competitors that didn’t exist before.

Keyword research isn’t a phase. It’s a practice. I revisit my keyword targets monthly, checking three things: Are my targeted keywords still producing clicks (not just impressions) in Google Search Console? Have AI Overviews appeared for keywords where they didn’t exist before? Has the intent of the search results changed for any of my target keywords?

If any of those answers are “yes,” I adjust. Sometimes that means updating existing content. Sometimes it means abandoning a keyword entirely and redirecting that page toward a more profitable term. The teams that treat keyword research as “done” after the initial sprint are the teams wondering why their traffic is declining despite publishing consistently.

Frequently Asked Questions About SEO Keywords

What are SEO keywords and why do they still matter?

SEO keywords are the words and phrases people type into Google, Bing, or AI tools like ChatGPT when looking for information, products, or services. SEO keywords still matter because they’re how search engines and AI systems understand what your content is about and whether to show it to users. Even as AI Overviews grow, ranking well in traditional search increases the chance of being cited in AI-generated answers, making keyword targeting more important, not less.

How many SEO keywords should I target per page?

Each page should target one primary keyword and two to five closely related secondary keywords that share the same search intent. Trying to target unrelated keywords on a single page dilutes the page’s relevance for all of them. The secondary keywords should be natural variations, synonyms, or long-tail versions of your primary keyword, not entirely different topics.

Do free keyword research tools work, or do I need paid ones?

Free tools like Google Keyword Planner, Google Search Console, and Google Trends provide useful starting data, especially search volume estimates and the actual queries driving traffic to your site. Paid tools like Semrush, Ahrefs, and Moz add competitive analysis, keyword difficulty scoring, and SERP feature tracking (including AI Overview detection) that free tools don’t offer. For serious keyword research at scale, paid tools save significant time and surface opportunities that free tools miss.

How has AI search changed keyword research strategy?

AI search has shifted keyword research from a volume-first to an intent-first discipline. Because AI Overviews answer many informational queries directly on the search results page (reducing organic clicks by up to 58% for position-one content, per Ahrefs’ December 2025 data), keyword researchers now need to evaluate whether a keyword is “AI-vulnerable” before investing in content. Keywords with commercial or transactional intent, and long-tail keywords with specific contexts, tend to retain more organic click-through potential.

What’s the difference between SEO keywords and the keywords AI tools use?

SEO keywords target traditional search engine results pages. AI tools like ChatGPT and Perplexity use retrieval-augmented generation (RAG) to pull information from web pages that already rank well in traditional search. So the same keywords matter for both, but AI tools also respond to conversational, multi-part queries that traditional keyword research tools don’t always capture. The practical takeaway: optimizing for traditional SEO keywords also improves your visibility in AI-generated answers.

Stop Researching, Start Filtering

The best keyword research I’ve done wasn’t the most thorough. It was the most ruthless about cutting keywords that looked good on paper but wouldn’t drive revenue. If I could go back and talk to 2024 me, staring at that beautiful, useless spreadsheet, I’d say: “Run everything through the Intent-Revenue Stack before you write a single word.”

Volume is a vanity metric when 58% of searches don’t produce a click. Difficulty scores are misleading when they don’t account for AI Overviews. The keywords that matter are the ones where a real human with a real problem will actually land on your page and do something.

If building that kind of keyword strategy sounds like more than your team can handle right now, LoudScale specializes in exactly this sort of intent-first SEO work, from keyword research through content execution.

But whether you do it yourself or get help, start with the customer’s problem. Filter hard. Write less, but write for the right queries. That’s the whole game now.