keyword research

Most “keyword research” fails before you even open a tool. I've seen teams produce over 100 pages in a quarter, confidently publish them, and then monitor Search Console for months while impressions barely shift. Not because the writing was terrible, but because the keywords were wrong: mismatched intent, fake volume, or a pile of near-duplicates that cannibalized each other (depending on your setup). Google doesn’t grade you on effort. It evaluates you based on how well you addressed the correct question compared to the other alternatives.

Keyword research isn’t a spreadsheet exercise; it’s how you decide what your site should be about in a way search engines can understand and users actually want. When executed properly, it influences your content strategy, internal links, product pages, and even the names you assign to features. Speaking from experience, done poorly, it turns into busywork: chasing head terms you’ll never win, ignoring the long-tail queries that convert, or picking keywords that look nice in a report but bring the wrong traffic.

I treat keyword research as a series of bets. You’re betting on demand (real people searching), competitiveness, and payoff. But tools are certainly helpful—Ahrefs, Semrush, GSC, and others—but they don't make the decision. The decision is derived from analyzing the SERP, identifying trends, and realistically assessing what you can publish that's genuinely more useful or thorough.

When you nail this part, everything that follows becomes simpler: briefs become clearer, content clusters are logical, and rankings seem less like chance. Get it wrong and you’re just producing pages.

What keyword research is actually for (and what it isn’t)

A lot of teams treat keyword research like a sorting contest: export 10,000 keywords, filter by volume, pick the top 50, and call it a plan. That doesn’t work because volume isn’t the thing you’re really buying. Or you’re buying fit:

• Fit with what the searcher wants right now (intent)
• Fit with what your site can credibly answer
• Fit with how Google is right now choosing to rank results
• Fit with what you can convert

If those don’t line up, you can “rank” and still lose. You’ll get traffic that bounces, traffic that never converts, or traffic that only reveals up for weird variations that don’t matter.

Keyword research also isn’t only for blog posts. The best keyword work usually changes your site structure:

• You realize a “feature” should be its own landing page because people search for it directly.
• You realize two “different” pages are fighting for the same query and need to be merged.
• You realize your pricing page is the real entry point for a category of searches, so it needs better copy and internal links.

This is the part most people skip because it’s not as simple as “write another article.” But it’s where the compounding effect comes from.

The three bets: demand, competition, payoff

If you want a practical way to think about keyword research, keep it to three bets. So every keyword you pick is you saying “I think this is worth doing.” Here’s what you’re betting on.

1) Demand: are real people searching, and is it the right kind of demand?

Tools give you volume, but volume is messy. It can be inflated by:

• Bots and noise
• Ambiguous terms with mixed intent
• A “topic” that’s seasonal or spiky
• People searching for definitions, not solutions

The fix isn’t complicated: you sanity-check demand with the SERP. If the top results are surprisingly all dictionary-style definitions and you’re trying to sell software, you can still write the page—but you should expect low conversion. That’s not a moral failure; it’s just what that query is.

A quick route to judge demand quality is to look at the modifiers people use. Queries with words like:

• “go-to”, “top”, “software”, “tool”, “platform”
• “pricing”, “cost”, “alternative”
• “template”, “checklist”, “example”
• “how to”, “guide”, “step by step”

…tend to signal a person trying to do something, not just learn trivia. That usually means better business value, even when the volume is lower.

2) Competition: who already owns the SERP, and why?

Keyword difficulty scores are fine as a rough filter, but they’re not a decision. The decision comes from looking at what ranks and asking: why do these pages deserve to be here?

When you open the top results, you’re usually seeing one of these situations:

• Authority lock: massive sites dominate because the topic is broad and they’ve earned trust. You can still compete, but you need a sharper angle or a stronger site.
• Format lock: Google clearly wants a specific format. If you publish a different format, you’re swimming upstream.
• Intent split: the SERP mixes informational and commercial pages. That’s a sign the query is ambiguous, and you’ll need to choose which intent you’re serving.
• Weak results: the pages ranking are thin, outdated, or not actually answering the query well. These are your go-to opportunities, even if the keyword tool says it’s “hard.”

Competition isn’t just “how many backlinks do the top pages have.” It’s also: how clearly do they satisfy the searcher? If the current results are good, you’ll need to be meaningfully better. “A bit longer” or “more headings” isn’t better. Or better means clearer, more complete, more current, more usable, or more trustworthy.

3) Payoff: if you rank, what happens next?

This is where keyword research stops being a content task and becomes a business task.

A keyword can be easy to rank for and still be a waste of time if it doesn’t lead anywhere. Payoff can mean:

• Direct conversion
• Assisted conversion
• Brand authority
• Internal link value

The practical move is to decide up front what success looks like for each keyword group. For example:

• A “how to” guide might be judged on email signups or time-on-page, not demos.
• A “best X software” page might be judged on demo clicks and comparison-table engagement.
• A “pricing” query might be judged on assisted conversions and sales-qualified leads.

If you don’t pick the payoff metric, you’ll end up judging everything by raw traffic, which is absolutely how teams get stuck publishing content that looks good in reports and does nothing else.

Intent is the middle step people skip

You can’t pick keywords well if you don’t understand intent. Intent is just: what job is the searcher hiring this query to do?

In practice, most queries fall into a few buckets:

• Informational: “what's…”, “how to…”, “examples of…”
• Commercial research: “best…”, “top…”, “X vs Y”, “alternatives”
• Transactional: “buy…”, “pricing”, “free trial”
• Navigational: brand or product names

Intent matters because Google is picky about mismatches. If the SERP is full of “top tools” listicles and you publish a 4,000-word tutorial, you might write a great piece and still not rank. Google will assume you didn’t answer the question people are asking.

A simple workflow that works:

1. Search the query in an incognito window.
2. Write down what types of pages show up: guides, product pages, templates, videos, forum threads, tools.
3. Note the angle: beginner vs advanced, B2B vs consumer, “definition” vs “step-by-step.”
4. Decide whether you can credibly match that intent and bring something better.

If you can’t match intent, don’t force it. Pick a different query or a different angle that fits your site.

Keywords are not pages: group first, then map

A classic mistake is treating every keyword as a separate page. That’s how you get:

• 12 posts that all target the same query with slightly separate wording
• a “definition” page fighting a “how to” page for the same SERP
• internal links that don’t make sense because the topic boundaries are fuzzy

Instead, group keywords by meaning and intent. One page should usually target a cluster of close variants, not one exact phrase.

Here’s a practical way to group:

• Start with one main query.
• List the close variants that mean the same thing.
• Separate anything that tweaks intent into a fresh group.

Example :

• “keyword research” (main)
• “how to do keyword research”
• “keyword research tool”
• “keyword research template”
• “keyword research for ecommerce” (could be the same page with a section, or its own page if the SERP is clearly different)

Mapping groups to pages is how you avoid cannibalization. It also makes your internal linking obvious: the template page links to the main guide; the tools page links to the guide; the guide links back to both.

Tooling: what each tool is good for

Tools are helpful, but each one has blind spots. I like to be blunt about this because teams waste weeks arguing over which tool is “right” instead of using them for what they’re good at.

Google Search Console (GSC)

Because it’s your actual impressions and clicks, gSC is the closest factor you have to truth. It’s also limited:

• Queries are sampled and sometimes grouped weirdly.
• You won’t see everything.
• It won’t help much for brand-new sites with no data.

What it’s great for:

• Finding keywords you’re already ranking for.
• Spotting pages that rank for unexpected queries.
• Catching cannibalization.

Ahrefs / Semrush

These are great for expansion and competitor research. They aren't a source of truth for exact volume.

What they’re great for:

• Keyword expansion from seed terms
• SERP previews and competitive analysis
• Finding competitor pages that bring traffic
• Rough difficulty filtering

Where they mislead:

• Volume can be off
• Difficulty scores can hide “weak SERPs”
• Click potential is often ignored

SERP inspection

This is the underrated one. It costs time, but it saves you from bad bets.

What it’s great for:

• Seeing intent and format
• Spotting SERP features that reduce clicks
• Noticing patterns

If you only do one “tool step,” do this one.

A practical workflow that doesn’t turn into busywork

You don’t need a 40-tab process. You need a repeatable loop that gets you to publishable decisions.

Step 1: Start with real inputs, not tool outputs

Seed topics should come from:

• sales calls
• support tickets
• onboarding questions
• product naming
• competitor positioning

If you start with a tool export, you’ll end up writing content for the tool, not for your market.

Step 2: Expand seeds into keyword groups

Use a keyword tool to expand, but keep it tight:

• Pull 50–200 related queries, not 10,000.
• Keep the list messy at first.
• Group by intent as you go.

You’re not trying to build a master database. You’re trying to pick your next 5–10 pages with confidence.

Step 3: Sanity-check with the SERP

For each keyword group, check:

• What formats are ranking?
• Are the results fresh or stale?
• Are there obvious gaps?
• Are there big brands you can’t realistically outcompete right now?

If it’s a “now” keyword or a “later” keyword, this is where you decide.

Step 4: Map to a page type and a conversion path

Pick the page type that matches intent:

• Guide / tutorial
• Comparison
• Alternatives
• Template
• Glossary / definition
• Product / feature landing page

Then decide the next step you want the reader to take. Don’t overthink it, but do pick one:

• internal link to a product page
• email signup
• free trial
• demo request
• download a template

If the page has no next step, you’re relying on luck.

Step 5: Publish, then feed results back into research

Keyword research isn’t finished when you publish. The feedback loop is:

• What queries did the page actually get impressions for?
• Which sections got engagement?
• Did it rank for the first intent or drift into something else?
• Did it cannibalize an existing page?

That feedback tells you whether to expand, split, merge, or build supporting pages.

Common failure modes

A lot of keyword research advice is “do more research.” That’s not the issue. The issue is doing the wrong kind of research and then committing to it for months.

Chasing head terms too early

Head terms are tempting because they look like “big wins.” They’re also where competition is highest and intent is often broad.

If your site is newer or your category is crowded, head terms are usually a “later” play. You can still target them, but you should expect a longer timeline and more supporting content.

A better early approach is mid-tail and long-tail queries that:

• have clear intent
• map to a specific page
• have SERPs with weaker results
• align with your product or service

Treating volume as the goal

Volume isn't the goal. Qualified traffic is the goal.

If you publish ten pages that each get 200 visits/month and convert, you’ll beat one page that gets 20,000 visits/month and does nothing. This sounds obvious, but teams still get pulled into volume because it’s easy to report.

Creating near-duplicate pages “to cover more keywords”

That doesn’t work. Google usually picks one and ignores the rest, or rotates them, or splits signals so none of them win.

If two pages have the same intent, merge them. If they have unique intent, separate them clearly and make the difference obvious in:

• title and intro
• headings
• examples
• internal links
• CTA

Ignoring SERP features and click loss

Some SERPs are click-starved. If the query triggers:

• heavy ads
• a hefty featured snippet that answers the question fully
• a local pack
• a shopping carousel

…your “volume” might not translate into clicks. That doesn’t mean you can’t target it, but you should adjust expectations and payoff metrics.

Not aligning with what you can credibly say

If you can’t back up claims with experience, data, screenshots, or clear explanation, your page will read like a rewrite. That’s a problem because the SERP is already full of rewrites.

The fix is to pick keyword groups where you can add something real: a process, a checklist, a template, a worked example, a decision framework.

Building a balanced keyword set

A solid keyword set usually has a mix: a few head terms to anchor your main pages, plenty of mid-tail terms that map cleanly to specific use cases, and long-tail queries that show obvious intent and are easier to win. Pair that with a quick SERP check, and you’ll know whether you’re writing a guide, a comparison, a product page, or something else entirely. That’s the difference between “we published a lot” and “we published the right things.”

Here’s how I like to balance it in practice:

• Head terms : These inform your main guides and core landing pages. They’re harder, slower, but they define what you’re “about.”
• Mid-tail : These are the workhorses. They tend to have clearer intent and better conversion alignment.
• Long-tail : These are often the fastest wins. They also tell you what people actually care about, which can reshape your product messaging.

The mistake is treating these as separate strategies. They should feed each other. Long-tail queries often reveal the sections your head-term guide is missing. Mid-tail pages often become the internal link bridges that help your head-term pages rank.

Brand-new: Keyword research deliverables that teams actually use

Keyword research “deliverables” usually fail because they’re not built for decisions. A 15-tab spreadsheet isn't a deliverable if nobody knows what to do with it.

If you want outputs that get used, keep it to a few things:

1) A keyword-to-page map

This is the backbone. For each planned page, include:

• Top keyword group
• Intent
• Page type
• Target URL
• Notes on SERP format
• Internal links
• Success metric

This forces clarity. Because you can see the whole site plan at once, it also prevents duplicate pages.

2) A short “SERP notes” section per page

Two to five bullets is kind of enough:

• What’s ranking and why
• What’s missing
• What angle you’ll take
• What you need to include to be competitive

This makes briefs faster and keeps writers from guessing.

3) A prioritized backlog with a simple scoring model

Don’t pretend scoring is science, but do it anyway. Even a basic 1–5 score for:

• intent fit
• competition realism
• payoff

…is enough to rank your backlog and stop debates from dragging on.

New: How keyword research changes as your site grows

Treat keyword research as a living input, not a one-time task. As search demand evolves, competitors change positions, and your site gains authority over time, the opportunities available to you also transform.

What tweaks over time:

• Your baseline ability to rank improves. You can move from long-tail to mid-tail, and later to head terms.
• Your internal link graph gets stronger. New pages rank faster because you can support them straight away.
• You get real query data. GSC starts telling you what Google already associates with your brand and pages.
• Your content gaps become obvious. You’ll see clusters where you have one strong page but no supporting pages, or vice versa.

A practical cadence that works for many teams:

• Monthly: review GSC for quick wins.
• Quarterly: refresh your keyword map and reprioritize based on what worked.
• Ongoing: add “fresh” keywords from sales/support/product conversations.

If you only do keyword research once a year, you’ll keep making bets with old information.

FAQ

What's keyword research?

Keyword research is the process of finding and prioritizing the search terms your audience actually types into Google when they’re trying to solve a problem you can help with. It’s not just a list of phrases—it’s a decision tool. You use it to pick topics, shape pages, and set expectations for traffic and conversions. Done well, it preserves you from writing “cool” content nobody searches for, and it prevents you from chasing terms that look major but don’t match what you sell.

A valuable way to think about it: keyword research is where you decide what promises your site is going to make. Each page is a promise to answer a specific question better than the alternatives. If you pick the wrong questions, you can write great pages and still get ignored.

How does keyword research work?

It does the job by turning a messy set of ideas into a ranked plan. You start with a few “seed” topics, expand them into real queries using tools and SERP clues, then group those queries by intent. After that, you sanity-check difficulty by looking at who already ranks and what formats Google is rewarding. The final step is mapping groups to pages so you’re not creating duplicates or forcing one page to cover five different intents.

In practice, the “magic” step is the SERP check. Tools help you find candidates. The SERP tells you what Google thinks the query boils down to, what kind of page is expected, and whether there’s space for you to win.

What are the benefits of keyword research?

The biggest win is focus: you stop guessing what to publish and start choosing based on demand and intent. Keyword research helps you spot quick wins, avoid cannibalization, and set realistic goals. And yes, it saves time. Or writing less content, but with clearer purpose, usually beats publishing more.

It also improves alignment across teams—SEO, content, paid search, and product can talk about the same terms instead of arguing from opinions.

One more benefit that doesn’t get enough attention: it makes your site easier to understand. Clear topic clusters and cleaner page intent make internal linking simpler and help search engines connect the dots.

How do I get the ball rolling with keyword research?

Start tiny and keep it concrete. Pick one product, feature, or pain point and list the questions you hear from customers, sales calls, support tickets, and demos. Plug those into a keyword tool to expand and get rough demand. Then open the current top results and note what intent they’re serving and what’s missing. Build a short backlog: 5–10 keyword groups, each mapped to one page type. Ship one page, learn, repeat.

If you want a simple first milestone: publish one page that targets one obvious keyword group, then use GSC data from that page to pick the next two. That loop builds momentum fast.

Keyword research isn’t about stuffing a page with the same phrase 20 times. The goal is to ensure that you're creating the appropriate content for the right audience at the optimal time. The top workflows start wide, then get picky, and finally get practical. Neglecting that middle aspect—intent—will lead you to waste time pursuing keywords that appear favorable in a tool but fail to convert.

A solid keyword set usually has a mix: a few head terms to anchor your main pages, plenty of mid-tail terms that map cleanly to particular use cases, and long-tail queries that show obvious intent and are easier to win. Pair that with a quick SERP check, and you’ll know whether you’re writing a guide, a comparison, a product page, or something else entirely. That’s the difference between “we published a lot” and “we published the right things.”

Treat keyword research as a living input, not a one-time task. As search demand evolves, competitors change positions, and your site gains authority over time, the opportunities available to you also transform. For what it's worth, choose five key topics, associate them with particular pages, and begin producing content that you'd be proud to see rank in the next quarter.

Ready to take action? Vistrify can help you get started. Try it free and see the difference for yourself.