Keyword research is the process of finding the exact phrases your audience types into Google and AI engines, sorting them by intent and difficulty, and matching each one to a page you can realistically rank. Do it in five moves: build seed terms from real customer language, expand them with a tool, group by search intent, score each keyword for effort versus reward, then map the winners to a publishing calendar so the work compounds instead of resetting every quarter.
Most guides stop at "find high-volume keywords." After eight years running SEO campaigns for service businesses, local brands, and eCommerce stores across the USA, Canada, Australia, and India, I can tell you the volume number is the least useful column in your spreadsheet. What separates a keyword list that ranks from one that gathers dust is intent grouping and a system for revisiting it. That is what this guide gives you.
Key takeaways
Search intent, not search volume, decides whether a keyword is worth targeting in 2026.
One primary keyword plus two to three secondary keywords per page prevents cannibalization.
Long-tail phrases convert at roughly 2.5x the rate of head terms and are far easier to rank.
A prioritization score (volume, difficulty, business value, intent match) beats gut feel.
Keyword research is a recurring calendar task, not a one-time project you finish and forget.
How to do keyword research: the five steps that matter
Google rewards pages that answer a specific need, so the workflow below is ordered by what moves rankings, not by what fills a checklist. Follow it in sequence the first time, then loop steps three through five every month.
Step 1: Build seed keywords from how customers actually talk
Seed keywords are the two- to four-word phrases that describe what you sell or write about, written in your customer's words rather than your own. Pull them from sales calls, support tickets, the questions people email you, and the language reviews use. A plumber's seed list starts with "blocked drain" and "hot water not working," not "hydro-jetting services."
This step is where industry insiders go wrong most often. We use jargon; searchers use symptoms and outcomes. Spend twenty minutes reading your own reviews and competitor reviews before you touch a tool. The phrases people repeat there are your highest-converting seeds.
Step 2: Expand the list with keyword research tools
Feed each seed into a keyword research tool to pull related terms, questions, and long-tail variants along with their monthly search volume and difficulty. Free options like Google Keyword Planner and Google Search Console show you real demand and the queries you already appear for. Paid tools such as Ahrefs, Semrush, and Mangools go deeper on difficulty scores and competitor gaps.
One underused source is your own Google Search Console. Its performance report lists queries you already earn impressions for but rank on page two, which are the fastest wins you will ever find. I start every audit here before spending a rupee or dollar on a paid tool.
Step 3: Group every keyword by search intent
Search intent is the reason behind a query, and it falls into four buckets: informational (learning), navigational (finding a specific site), commercial (comparing options), and transactional (ready to act). Sort your list into these four before writing anything. A page that matches intent ranks; a page that fights it never will, no matter how good the writing.
The practical test: open an incognito window, search the keyword, and look at what Google already ranks. If page one is all how-to articles and yours is a product page, the intent is informational and your product page will not break through. Match the format Google is already rewarding. This single habit has salvaged more of my clients' content than any technical fix.
Step 4: Judge difficulty honestly, then score for priority
Keyword difficulty estimates how hard it is to rank based on the strength of pages already ranking. Do not chase a keyword with a difficulty of 70 when your site is six months old. Instead, score every keyword on four factors and let the math decide what to write next.
Give each keyword a 1 to 5 rating on: monthly search volume, ranking difficulty (inverted, so easier scores higher), business value (how close the searcher is to buying), and intent match (can you genuinely satisfy it). Add the four numbers. Anything scoring 14 or above goes to the top of your calendar. This removes the emotion, and it stops you from writing the flashy high-volume term that will never rank over the boring one that pays the bills.
Here is how that plays out in practice. Say a new bookkeeping firm is weighing "accounting software" (volume 5, difficulty 1, business value 2, intent match 2, total 10) against "how to fix a VAT return error" (volume 2, difficulty 4, business value 5, intent match 5, total 16). The first term looks tempting because of its traffic, but the firm cannot outrank the software giants ranking for it, and the searcher wants a product, not a bookkeeper. The second term scores higher because the firm can actually rank for it, and anyone searching it has a problem the firm is paid to solve. The score points to the right page every time your instinct wants the vanity term.
Step 5: Map keywords to a publishing calendar
Assign one primary keyword and two or three secondary keywords to each planned page, then slot that page onto a dated calendar. Grouping keywords this way avoids cannibalization, which happens when two pages target the same term and split your ranking signals. A calendar also turns a 200-row spreadsheet into a manageable one-page-per-week plan.
This is the step every competitor guide skips, and it is the whole reason The SEO Calendar exists. A keyword list with no publishing dates is a wish, not a plan.
Where you place the secondary keywords matters as much as picking them. Put the primary keyword in the title, the opening answer, one H2, and the closing lines. Work the secondary terms into subheadings and body copy where they read naturally, near the primary term but never stacked in the same sentence. If you find yourself repeating an exact phrase to hit a density target, stop; Google reads topics and synonyms, and forced repetition reads worse to humans and to AI engines alike.
The best keyword research tools, and when each earns its place
You do not need every tool. You need the right one for your stage.
Google Keyword Planner and Google Search Console cost nothing and are enough to launch a small site or local business. Google Trends adds seasonality, which matters if your demand spikes around holidays or events. When you outgrow the free tier and need competitor gap analysis, Ahrefs and Semrush are the industry standards, with Semrush's related-keyword and intent data being particularly strong. Mangools (KWFinder) sits in the middle on price and suits solo bloggers and small agencies who find the big platforms overwhelming.
My honest take after testing all of them on client budgets: start free, prove the process works, then pay for a tool only once the manual gap analysis is costing you more time than the subscription. Tools speed up research; they do not replace the intent judgment in step three.
The mistake that quietly kills most keyword research
The costly error is treating keyword research as a one-time project. Teams run it once at a site launch, build a content plan, and never look again. Six months later the plan is stale, competitors have moved, and half the keywords have shifted intent because Google updated what it shows.
Keyword research is a maintenance habit, not a founding event. Search demand moves with seasons, news, product launches, and algorithm updates. The sites that win treat it as a recurring calendar entry: a light monthly pass to check Search Console for new page-two opportunities, and a deeper quarterly review to reassess difficulty and refresh the priority scores. Building this rhythm is exactly what a monthly SEO checklist is for. The compounding effect is real. A site that revisits its keywords every month for a year ends up with a map its competitors cannot copy, because it reflects live demand rather than a snapshot from last January.
Keyword research for AI search: AEO and GEO
AI engines have changed what a keyword is. Queries typed into ChatGPT, Perplexity, and Google's AI Overviews average around 23 words, compared with roughly four words on classic Google search, because people describe their whole situation instead of typing fragments. Nearly a third of the US population is forecast to use generative AI search in 2026, so this is no longer a fringe channel.
For your research, this means three shifts. First, mine full questions, not fragments; the "People Also Ask" box and AI chat logs are gold. Second, write self-contained answers of 40 to 60 words directly under question-style headings, because AI systems extract passages, not whole pages. Third, track branded and comparison queries, since AI answers cite sources and higher-intent AI referral traffic converts far better than a cold organic click.
None of this replaces the five steps above. Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) sit on top of solid intent-first research. The foundation is the same; the surface area is larger. Google's own guidance still points in the same direction: create helpful, reliable, people-first content and let the SEO serve the reader, not the algorithm.
How many keywords should one page target?
Target one primary keyword and two to three closely related secondary keywords per page. The secondary terms should be variations or subtopics of the primary, not separate ideas. This keeps the page focused enough to rank while capturing the natural phrasing people use, and it prevents two of your own pages from competing for the same term.
Frequently asked questions
How long does keyword research take?
A first pass for a small site takes three to five focused hours: an hour on seeds, an hour on tools, and the rest on intent grouping and scoring. After that, budget thirty minutes monthly for maintenance. The initial time is front-loaded; the ongoing habit is light once the system is in place.
Is keyword research still worth it with AI search?
Yes, and arguably more than before. AI engines still pull answers from ranked, well-structured content, and they favor pages that match intent precisely. The work shifts from chasing volume toward mapping real questions, which is where higher-converting traffic now lives.
Can I do keyword research for free?
You can run the entire five-step process using Google Keyword Planner, Google Search Console, Google Trends, and the "People Also Ask" results. Paid tools save time on competitor gap analysis and difficulty scoring, but they are an accelerator, not a requirement for getting started.
Where to go from here
Pick five seed keywords today, run them through the five steps, and score the results. Put the top three on a dated calendar for next month. If you want the recurring structure that keeps this from becoming another abandoned spreadsheet, the monthly checklists and marketing calendars at The SEO Calendar are built for exactly this rhythm.
