How to Do Keyword Research: A Step-by-Step 2026 Guide
How do you do keyword research? A step-by-step process with free tools: seed keywords, search intent, volume and difficulty, clustering and page mapping.
The short answer to how to do keyword research is this: you start with a few seed keywords that describe your business, expand them with real search data, read the search intent of each phrase from the results on the first page, and connect keywords with the same intent to a single page. Volume and difficulty are inputs to that decision, not the decision itself. The result is a content plan where every row maps to a page and a business goal.
In this post I describe the process I use on both the SEO and the Google Ads side. I go step by step through working with free tools, reading the metrics and turning the list into pages.
What is keyword research and why does it matter?
Keyword research is the work of finding the words people use to search for the problem you solve. It matters because it bases content and ad decisions on real search behavior instead of assumptions. The language a business uses about itself and the language customers type into the search box are often not the same.
Google highlights this gap in its SEO Starter Guide. It notes that users who know a lot about a topic may search with different keywords than someone new to it, and that anticipating these differences can have a positive effect on search performance.[4] Keyword research lets you see those differences in data instead of guessing them.
Research has three concrete outputs:
- Page plan: Which pages need to exist and which question each one will answer.
- Priority order: Which page gets written first, based on demand, intent and closeness to the business.
- Shared language: Content, advertising and product teams talking with the same words.
How to do keyword research: the step-by-step process
Short answer
How do you do keyword research?
Keyword research starts with seed keywords that describe the business. These keywords are expanded with real search data, and the intent of each phrase is read from the first-page results. Next, priority is set by demand, difficulty and business value. In the final step, keywords with the same intent are grouped and each group is matched with a single page.
- Define 5-10 seed keywords that describe your business and your customer's problem.
- Expand the list with Search Console, autocomplete and Keyword Planner.
- Look at the first-page results for each phrase and mark its search intent.
- Prioritize by volume, difficulty and business value.
- Group keywords with the same intent and connect each group to a single page.
- After publishing, update the list with Search Console data.
I unpack these steps in order below. The first two steps make the list longer, and the later steps narrow it down and make it meaningful.
1. Define your seed keywords
A seed keyword is a short phrase that describes your business or topic in its most general form. Five to ten seeds are enough to begin. The best way to find them is to listen to the customer’s voice: questions that reach the sales team, form messages, live chat logs and phrases used in quote calls.
Do not limit seed keywords to product names. Add phrases that describe the problem, the solution and comparisons too. For accounting software, for example, “how to issue an e-invoice” is as much a seed as “bookkeeping software”.
2. Expand the list with real search data
In the expansion stage, the goal is to collect the real queries that branch out from your seeds. These are the free sources I use here:
- Google Search Console: If your site is live, this is the first stop. The Performance report lists the queries your site appears for on Google, along with clicks, impressions, click-through rate (CTR) and average position.[2]
- Google Ads Keyword Planner: You enter words or a website to get new keyword ideas, and you see monthly search estimates and ad cost for each keyword.[1]
- The search results page: Autocomplete, the “People also ask” box and the related searches at the bottom of the page show directly how people phrase the topic.
- Google Trends: It shows how interest in a topic changes over time and how it is distributed by region.
Do not filter the list at this stage. You have to collect before you can eliminate; the number will shrink in the next step anyway.
3. Mark the search intent of each phrase
Search intent is the result a user wants to reach when typing a query. Guessing intent from the keyword alone is misleading. The most reliable method is to search the query and look at the type of results on the first page. Whatever type of page Google puts forward is the best signal of what users expect for that query.
| Intent type | User’s goal | Usually on the first page | Example query |
|---|---|---|---|
| Informational | To learn | Guide, definition post | “what is a long tail keyword” |
| Navigational | To reach a specific site | Brand site, login page | “search console login” |
| Commercial investigation | To compare options | List, comparison, review | “best seo tool” |
| Transactional | To buy or apply | Product, category, service page | “seo consulting price” |
If different types appear mixed on the first page, the intent is mixed. In that case, pick the dominant type or meet both needs on your page. I covered reading intent from the writing side as well in the how to write SEO-friendly content post.
4. Prioritize
Prioritization means weighing every phrase on the list by demand, difficulty and business value together. None of the three decides on its own. I explain how I read these metrics separately in the next section.
5. Cluster and map to pages
In the final step you group keywords that answer the same intent and match each group with a single page. This is where the list turns into a plan. I give the details in the clustering section below.
How do you read search volume, difficulty and business value?
Search volume, difficulty and business value are three separate metrics that determine a keyword’s priority, and each one answers a different question. Volume answers “how many people search for it”, difficulty answers “how much effort it takes”, and business value answers “what this visitor brings me”.
Search volume
Search volume is an estimate of how many times a query is searched per month on average. Keyword Planner provides this value as a monthly search estimate.[1] I read this number not as a promise of traffic, but as an indicator of scale for comparing queries. The volumes in paid SEO tools are also based on their own estimation models and vary from tool to tool.
Google Trends measures something different. According to Google, each Trends data point is normalized by dividing it by the total searches of that geography and time range, and then scaled on a range of 0 to 100.[3] So 100 in Trends does not mean “100 searches”; it means the highest relative interest in the selected period. Use Trends to see seasonality and the preference between two phrases, not for volume.
Keyword difficulty
Keyword difficulty (KD) is a score that estimates how hard it is to reach the first page for a query. Each tool calculates this score with its own method, so a 40 in one tool is not the same as a 40 in another. Instead of using the score alone, I check the first page myself: Is it full of large, established sites? Are there outdated pages that only skim the topic? Are forum or Q&A pages ranking? The second and third cases signal that there is room for a better page.
Business value
Business value is how close a visitor arriving through a keyword is to your business. A practical way to measure it is to look at the advertising side. Keyword Planner also shows the average cost for your ad to show on searches for a keyword.[1] Queries where advertisers are willing to pay a high cost per click are usually queries with high commercial value.
Choosing by volume only
- Generic, mixed-intent keywords rise to the top
- You start where competition is highest
- Visitors may be far from the business
- Results stay unclear even months later
Choosing with intent and value
- Long-tail queries with clear intent rise to the top
- You start with winnable queries
- Visitors are closer to the next step
- Each page has a clear goal and measurement
Why are long-tail keywords valuable?
A long-tail keyword is a search phrase, usually made up of three or more words, that has low volume on its own but clear intent. Its value comes from that clarity: it is unclear what someone typing “seo” wants, while someone typing “category page seo for an ecommerce site” is clear about what they need.
Long-tail queries are high on my priority list for three reasons:
- Easier rankings: Competition is usually lower for these queries, and there is room for a page that covers the topic in depth.
- More qualified visitors: The more specific the question, the more likely a visitor is to take the next step when your page answers it.
- Fit for AI answers: Users type longer, question-shaped phrases into AI assistants and AI Overviews. Passages that answer such questions clearly are better suited to being quoted. I covered this in detail in the what is GEO post.
In Turkish, the long tail also multiplies through inflectional suffixes. Phrases meaning “keyword research”, “how to do keyword research” and “keyword analysis” usually share the same intent. Instead of opening a separate page for each of these variations, you should use them naturally on the same page.
How do you group keywords and map them to pages?
Keywords are grouped by bringing together queries that show the same results on the first page. If you search two queries and see that most of the top 10 results overlap, Google treats those two queries as the same intent. Those two queries should be served by a single page.
For each group I define the following:
- Primary query: The most representative phrase in the group. The page’s title and URL are built around it.
- Supporting queries: Variations and questions with the same intent. They are covered in subheadings and the FAQ.
- Page type: Blog post, service page, category page or product page. The column in the intent table is useful here.
- Existing page: Is there already a page on the site that answers this intent? If so, strengthen that page instead of opening a new one.
The last item is the most frequently skipped step. When two pages on the same site target the same intent, they compete with each other in rankings; this is called keyword cannibalization. Google’s spam policies also define pages created to rank for similar queries, which lead users to intermediate pages instead of the real destination, as doorway abuse.[5] Pages duplicated from a template for every city or every keyword variation come close to this definition.
How do you build a shared keyword list for SEO and Google Ads?
Because SEO and Google Ads draw from the same search box, working from a single keyword list gives both sides more data. The list is the same, but the way it is used differs: on the SEO side, keyword groups become a page and content plan; on the ads side, they become ad groups and negative keywords.
These are the data each side brings to the other:
- From ads to SEO: The search terms report in your ad account shows the real queries that triggered your ads. Queries that do not appear in Search Console, because you are not yet visible organically, can show up here.
- From SEO to ads: Queries with high impressions but low click-through rate in Search Console are good candidates for ad copy and landing page tests.[2]
- Shared negative list: Intents unrelated to your business, such as “free”, “job listing” or informational “what is” queries, can waste ad spend. On the SEO side, the same queries may be content opportunities.
I describe how I set up ad management from this angle on the performance marketing page.
What are the most common keyword research mistakes?
The most common mistake in keyword research is treating the list as the result. The list is an input; the real output is the plan showing which page will serve each keyword group. Other mistakes I see often:
- Choosing by volume only: The generic keyword at the top is usually the hardest and the most mixed in intent.
- Treating Trends as volume: The 0 to 100 scale is relative interest, not a search count.[3]
- Opening a page for every variation: Queries with the same intent should be grouped on a single page.
- Pasting the keyword into the text: Google states that repeating the same words unnaturally often goes against its spam policies.[5]
- Skipping existing data: Research done for a live site without opening Search Console ignores the most reliable data you already have.
- Doing it once and stopping: Search behavior, products and competitors change; the list should change too.
Can you do keyword research with AI?
AI tools speed up the idea generation and grouping stages of keyword research, but they do not replace search data. A language model can suggest plausible variations of how users might ask about a topic. Whether those variations are actually searched can only be confirmed by Search Console, Keyword Planner or an SEO tool.
I use AI for these tasks:
- Seed expansion: Listing the sub-problems, synonyms and different phrasings of a topic.
- Grouping draft: Pre-grouping a long keyword list by intent. I then verify the groups by checking the SERP.
- Question extraction: Separating the questions in a group as FAQ and subheading candidates.
I do not take numbers such as volume, difficulty or cost from a language model. Those numbers must come from measured data.
Where should you start?
The shortest way to start keyword research is to work from the data you already have. If your site is live, export the last three months of queries from Search Console, group them by intent and connect each group to an existing page.[2] Groups without a page are the first pieces of content you should write.
If your site is new, start with 5-10 seed keywords, expand the list with Keyword Planner and the search results page, and then narrow it with the checklist above. You can read how I turn this list into a strategy together with technical SEO and measurement on the SEO services page, and about my own background on the about page.
Key terms
- Seed keyword
- Seed keyword is a short phrase that describes a business or topic in its most general form and serves as the starting point of keyword research. You enter seed keywords into tools to list the long-tail queries and questions that branch out from them.
- Keyword difficulty
- Keyword difficulty is a score that SEO tools calculate with their own methods to estimate how hard it is to reach the first page for a query; it is often shortened to KD. Scores are not comparable across tools and should be checked against the actual SERP.
- Search volume
- Search volume is an estimated value showing how many times a query is searched per month on average in a given region. Google Ads Keyword Planner and SEO tools estimate this value from their own data. It is not a promise of traffic, but an indicator of scale for comparing queries.
Frequently asked questions
Can keyword research be done for free?
Yes. Google Search Console, Google Ads Keyword Planner, Google Trends and the autocomplete and related searches on the results page provide enough data for most small and mid-sized sites. Paid tools save time on competitor analysis and difficulty scores, but they are not required to get started.
Should I pick the keyword with the highest search volume?
Not always. High-volume generic keywords often carry mixed intent and face strong competition. Long-tail queries with lower volume but clear intent that connects directly to your business usually rank faster and bring more qualified visitors.
Should every keyword get its own page?
No. Keywords that answer the same search intent should be grouped on a single page. Opening a separate page for every small variation makes your pages compete with each other and can come close to the doorway definition in Google's spam policies.
How often should keyword research be refreshed?
Every time you plan new content or a campaign, and every few months for existing pages using Search Console data, is a good rhythm. For seasonal topics, periodic checks in Google Trends show you in advance when demand starts to rise.
Can the same keyword list be used for Google Ads and SEO?
Yes, and it is often useful. The same list produces a page and content plan on the SEO side, and ad groups and negative keywords on the ads side. The search terms report in your ad account also reveals real queries you have not yet seen on the organic side.
Sources
The full topic: Content SEO guide ↗ · Semantic SEO guide ↗