Keyword research,
step by step.

Choose your audience, market and seed keywords.

Start by writing down what you sell, who buys it and where those customers are. Choose a useful goal, such as product sales, qualified inquiries or trial signups. This keeps a large keyword list focused.

For a US campaign, select United States in your research tool. Use the intended language and keep the same country for competitor checks and tracking. For a local business, check the actual city or service area too.

Next, list the main products, services, problems and use cases. These are your seed keywords. Use a manageable starting set, such as five to ten topics. You can add more after the first round.

Find seeds in your product navigation, sales calls, support emails, customer reviews and competitor categories. Include common names and synonyms. Customers may search for “sit stand desk” even if your catalog calls it a “height-adjustable workstation.”

If the site already gets search traffic, check its existing queries too. You may have a page close to ranking well that deserves attention before a new article. The Search Console workflow below shows how to find these opportunities.

Save: your audience, goal, country, language and starting topics.

Find keyword ideas with Ahrefs or OpenSEO.

A research tool expands each seed into more searches. At this stage, collect useful possibilities. You will check the numbers and remove weak ideas in the next steps.

How to do keyword research in Ahrefs

1. Open Keywords Explorer, enter your seed keywords and select United States.
2. Open Matching terms to find searches containing those topics.
3. Use the question filter to find searches beginning with words such as “how,” “what” or “which.”
4. Use include and exclude filters. For desks, you might explore “small,” “tall,” “electric,” “best” and “versus,” while removing unrelated jobs or software searches.
5. Check Related terms for ideas that may use different wording.
6. Save relevant keywords to a list or export them. Keep country, volume, difficulty and available traffic metrics with each term.

Do not reject every small-volume keyword yet. First check whether it belongs to a useful group. The report names above follow Ahrefs’ keyword research guide; availability depends on your plan.

How to use DataForSEO with OpenSEO

DataForSEO supplies SEO data. OpenSEO gives you an interface for working with it. You can use the hosted OpenSEO app or self-host it. The self-hosted version uses your own DataForSEO credentials. Open-source software still has data-usage costs. See the official OpenSEO project for setup options.

1. Open Keyword Research and enter a seed such as “standing desk.”
2. Set the target market to the United States and check the language setting available in your version.
3. Review the keyword ideas and their volume, difficulty, CPC and intent data.
4. Open the search results for promising terms to see the pages already ranking.
5. Save useful keywords. Add tags for the topic, planned page or priority.
6. Repeat with other seeds. Use the saved list to build your page plan and choose terms to track.

These steps use OpenSEO’s documented keyword research and saved-keyword features. Menu details can change between versions.

Using DataForSEO directly

If you prefer an API, spreadsheet connection or your own research script, DataForSEO offers separate data requests for different tasks. An API lets software request data automatically. You can skip the technical setup by using OpenSEO.

Use the same location and language for comparable requests. Start with small batches, save the results, and avoid requesting the same data repeatedly. Keep the data date and provider in your worksheet.

Can you do keyword research for free?

Yes. Free sources can build a useful starting list, although they may show less data or take more manual work.

• Google Keyword Planner: open Discover new keywords, enter seeds or a relevant website, and set your location. It provides keyword ideas and search estimates; the detail shown depends on account access. Follow Google’s Keyword Planner instructions.
• Google Trends: compare topics in the target country. Check seasonality and rising related searches. Its 0–100 scale shows relative interest, not monthly search counts.
• Search Console: find queries already bringing your own site impressions and clicks.
• Autocomplete, People Also Ask and related searches: collect questions and variations from real results. Their appearance does not prove a particular search volume.
• Reviews and communities: find customer wording and problems worth testing in a keyword tool.

Save: one working list of relevant ideas, with the source of each idea.

Check which keywords competitors rank for.

Your search competitors are the sites appearing for your target keywords. They may include stores, publishers, forums and directories. Pick a few that cover the same audience or products as you.

Look at both their keywords and their best-performing pages. A page often ranks for many related searches, so the page view helps you spot whole topics rather than isolated phrases.

In Ahrefs

Enter a competitor in Site Explorer. Review Organic keywords and Top pages, using US data. Then use Content Gap to find searches competitors rank for that your site does not. Start with terms shared by several relevant competitors and remove brands or products you do not serve. See Ahrefs’ content gap walkthrough.

In OpenSEO or DataForSEO

OpenSEO’s Domain Overview shows ranking keywords and top organic pages. Compare relevant topics with your own list. With DataForSEO directly, Ranked Keywords supplies competitor terms and Domain Intersection helps compare two sites.

A competitor ranking for a term is a reason to investigate it. Check that the topic fits your business and that you can provide a useful answer. Copying every competing page creates a large plan without a clear purpose.

Save: the relevant keyword, competitor URL, missing topic or answer, and any existing page you could improve.

Understand search volume, difficulty and traffic potential.

Keyword metrics help you compare ideas. They are estimates, and each provider calculates them differently. Keep the country, source and date next to the numbers.

Search volume: how often is the term searched?

Search volume is an estimate of searches per month. It is not the number of unique people or visits your site will receive. An annual average can hide a seasonal peak or sudden change.

Check monthly history before choosing a topic. A Christmas-related keyword and an everyday product search need different publishing schedules. Investigate unusual spikes. Do not add all close variants together and assume the total is a separate audience.

Keyword difficulty: how hard might it be to rank?

Keyword Difficulty, or KD, is a tool’s estimate of ranking competition. Ahrefs KD is based on the referring domains of the top ten organic pages. Referring domains are the different websites linking to a page.

DataForSEO has its own difficulty model. A score of 20 from one provider does not necessarily mean the same thing as 20 from another. Neither score can decide whether your page matches the search or has the experience needed to answer it.

Traffic potential: how much could the whole page attract?

A useful page can rank for many related terms. Ahrefs’ Traffic Potential estimates the total organic traffic of the top-ranking page across its keywords. It can reveal a topic larger than the exact phrase you entered. It remains a competitor-based estimate, not a forecast for your site. See Ahrefs’ keyword strategy guide.

CPC and ad competition: what do advertisers pay?

CPC means cost per click in paid search. It can suggest commercial interest, but a high CPC does not prove that a keyword will be profitable for you. Google Ads competition describes advertisers, not organic ranking difficulty. Google’s metric definitions explain the distinction.

Clicks, trends and business value

Domain Rating and similar authority scores describe a provider’s view of a link profile. They are not Google scores or guaranteed ranking requirements. Our SEO metrics guide explains the terms used when comparing sites and placements.

Save: the useful metrics and a short note about what the numbers miss.

Check search intent and the pages already ranking.

Search intent means what a person wants to do when they search. A SERP is simply a search engine results page. Review it for your target country before deciding what to create.

These labels are starting points. Also note the page type, such as a guide, product category, service page or tool; the format, such as a comparison or tutorial; and the angle, such as small rooms or tall users.

Check several leading pages. What do they answer? What do they show? Is the advice recent? Can you offer better tests, clearer examples or a simpler way to complete the task? If results are mixed, record that rather than forcing one interpretation.

Check the whole results page

Look for ads, shopping results, Maps, videos, forums, featured snippets and AI answers. These can change how visible an ordinary website result is. For a quick fact, the answer may already be on Google. For a purchase or detailed comparison, there may be a stronger reason to click.

For example, guest posts versus niche edits helps someone compare methods. A niche edits service page helps someone evaluate and request a service. Related words do not always mean the same page should rank.

Save: intent, preferred page type, important questions, competing URLs and the date checked.

Choose keywords that fit your business and your site.

Clean your list before planning pages. Remove duplicate entries, wrong meanings, places you do not serve and products you do not offer. Keep competitor brand terms separate so they do not make your possible audience look larger than it is.

For each remaining idea, ask:

1. Is it relevant? Would the right customer make this search?
2. Can we help? Can we create the page or tool the results call for?
3. Is there useful demand? Do searches, customer questions or existing impressions support it?
4. Can we compete? What would it take to match the best results and add something useful?
5. Is it worth the work? Could the visits lead to sales, inquiries or another agreed goal?

Use volume and KD filters to reduce a very large list, then review the shortlist yourself. There is no universal “good volume” or KD cutoff. A new site with little relevant content faces a different challenge from an established specialist.

Mark each idea improve now, create next, research further or skip. Add one sentence explaining why. You can compare these choices again after the keywords are grouped into pages.

Save: a shortlist with a clear reason for each priority.

Build keyword clusters and cover the topic properly.

A keyword cluster is a group of searches one page can answer well. For example, “standing desk for tall people” and “best desk for a tall person” may belong together. “How to fix a standing desk motor” is a different task.

Compare the search results for each phrase. When many of the same URLs rank and the reader wants the same thing, one page is often a sensible choice. This is called SERP overlap. Similar wording alone is weaker evidence.

Ahrefs can group a list using Clusters by Parent Topic. Treat those groups as a first draft and inspect important ones. The Parent Topic is based on the leading page’s search traffic, so it may need adjustment for your page plan. Ahrefs’ clustering guide explains the method and comparing two result sets.

In OpenSEO, use saved-keyword tags to keep page groups organized. In a spreadsheet, add a “cluster” column. If AI groups the list, check whether each suggested group actually shares intent and ranking pages.

What is a topical map or topical graph?

A topical map shows a subject, its main subtopics and the pages that cover them. A topical graph also shows how those topics connect. It helps you spot missing explanations and useful internal links.

For this guide, the main topic is keyword research. Its branches include:

• Basics: meaning, purpose, keyword types and search intent.
• Discovery: seeds, tools, competitors, customer questions and Reddit.
• Analysis: volume, difficulty, trends, clicks and business value.
• Planning: clusters, page mapping, content gaps and priorities.
• Publishing: briefs, on-page SEO, internal links and useful examples.
• Measurement: rankings, Search Console, conversions and updates.
• Special cases: local search, ecommerce, B2B, paid search, video and AI.

The connections matter. Search intent affects the page type. Clusters affect how many pages you need. Page mapping helps prevent overlap. Tracking shows when the map needs to change.

How to find missing subtopics

For each important topic, list the definitions, options, comparisons, steps, costs, examples and problems a reader needs to understand. Check your keyword list, customer questions and leading pages for gaps. Ask whether the reader can complete the task after using your content.

You may also hear about entities. These are specific things or concepts in a topic: Google, Ahrefs, search volume and Search Console are examples here. Explain their relevant relationships naturally. A list of names inserted into an article does not make it complete.

Save: keyword groups, missing subtopics and connections between planned pages.

Assign each keyword group to a page.

Keyword mapping means choosing the URL that should cover each group. Give the page a primary keyword, supporting phrases, a clear purpose and a planned next action.

Improve an existing page or create a new one?

Improve an existing page when it already covers the same need and can support the right format. Its rankings, links and history may make it the best starting point. Create a new page when the search needs a meaningfully different answer or action.

What is keyword cannibalization?

Cannibalization is a problem when similar pages compete for the same intent and hurt the site’s performance. Simply seeing two URLs for one keyword does not prove it. Check which page ranks, whether the wrong one appears, and whether positions or traffic keep shifting between them.

You can clarify each page’s role, improve internal links, or combine overlapping content. If a page is retired, assess a redirect to the closest relevant replacement. Do not merge useful pages just because they share a phrase.

Plan internal links at the same time

A pillar page is a main guide that introduces a broad topic. Supporting pages explain narrower topics. Link them where the next page helps the reader. A comparison can link to a service, while the service can link back to details about the method or price.

For example, a buyer can move from the link building services overview to the guest posting service, then read the guest post pricing guide. Each page answers a different question.

Save: the target URL, page type, primary keyword, supporting questions and useful internal links.

Write a brief, then create the right content.

A content brief tells the writer who the page is for, what it must answer and what evidence is available. Use the keyword group and live search results to build a short outline.

• Main answer: what should the reader understand or do first?
• Important sections: which steps, comparisons, costs or problems need explaining?
• Proof: what tests, screenshots, examples, data or experience can you show?
• Page format: does the task need an article, service page, category, video or working tool?
• Next step: which product, service or related guide helps the reader continue?

For the tall-user desk comparison, show usable height ranges, stability tests, measurements and product limitations. Explain how each desk was assessed. These details help a shopper choose and give your page a reason to exist.

Where should keywords appear?

Use the main topic naturally in the title tag, main heading and opening explanation. Use descriptive subheadings and related wording where it helps. For a new page, choose a short, clear URL. Write a meta description that accurately describes the benefit of visiting.

Use descriptive internal-link text and image alt text that explains the image. Avoid repeating the exact phrase in every heading or forcing it into unrelated alt text. There is no useful keyword-density target to hit.

Check that the published page loads well on phones, has one clear main heading and is available for crawling and indexing. Keep a useful existing URL unless there is a real reason to change it.

Where promotion and link building fit

Once the page is useful, decide how the right audience will find it. Link from relevant pages you already own. Share it through channels that suit the subject. Original research or a strong expert story may fit digital PR placements; a service or resource may suit a relevant editorial placement.

Paid links should be qualified appropriately, such as with rel="sponsored", in line with Google’s outbound-link guidance. Placements do not guarantee rankings.

Save: the brief, publication or update date, owner and next review date.

Track rankings, search traffic and useful actions.

Use a rank tracker to monitor a chosen keyword list in a set location and device. Use Search Console to learn which queries actually show your site. Use analytics and sales data to see whether those visits help the business.

How to use Google Search Console for keyword research

1. Open the Search results Performance report.
2. Filter the country to United States for a US campaign.
3. Choose a useful period and compare it with an earlier one. Use the same season last year where relevant.
4. Open a page and inspect its queries, impressions, clicks, CTR and average position.
5. Find relevant questions the page could answer better, or searches reaching the wrong page.
6. Update the keyword map and decide whether to improve the page or create a distinct one.

Impressions count appearances in search. CTR, or click-through rate, is clicks divided by impressions. Average position combines many searches; it is not one fixed ranking. Search Console omits some query data and cannot show topics for which your site has no visibility.

Ahrefs Rank Tracker and OpenSEO Rank Tracking can monitor selected keywords over time. Keep the location and device settings consistent. Record qualified inquiries, purchases or trial signups by landing page too.

Google explains how Search Console and Analytics work together: one describes visibility before the click, the other behavior after arrival. Their totals use different definitions and will not match exactly.

A monthly check is a useful starting point. Revisit the broader plan quarterly and sooner when products, demand or results change. Record the date of each content change so you can compare before and after without assuming every movement came from your edit.

Save: the result, what you learned and the next change to test.

Extra ways to find useful ideas

Use Reddit, customer questions and AI alongside the tools.

Find the language customers actually use

Read sales questions, support tickets, product reviews and relevant Reddit threads. Look for repeated problems, failed attempts, comparisons and details people care about. A desk shopper may talk about wobble or knee space before using the phrase “ergonomic furniture.”

1. Find communities where your customers discuss the problem.
2. Search a topic directly or use a query such as site:reddit.com "standing desk" "tall".
3. Save the thread URL, date, question and useful wording.
4. Look for the same concern across different threads and customer conversations.
5. Test the resulting keyword ideas in Ahrefs or OpenSEO and inspect the search results.

Practitioner discussions about Reddit as a keyword source and keywords with reported zero volume describe this approach. Use them as ideas to test. Reddit votes do not measure search volume or represent every customer.

Use AI to speed up the organizing work

AI can suggest seeds, clean an export, label questions and propose keyword groups. Give it your actual products, audience and data. Then check the groups against search results and decide which ones matter.

If the AI tool has live SEO-data access, check which source it used and when. Without that access, treat its keyword and metric suggestions as unverified. Do not upload private customer details just to generate seed ideas.

How do AI answers change keyword research?

Add one question to your SERP review: why would someone still visit this page? A working calculator, detailed comparison, original test or useful template can give them a reason. A basic definition may be answered directly in the results.

Longer AI prompts can reveal follow-up questions, but they do not need a separate page for every wording. Google describes related searches behind AI answers as query fan-out. Its AI-search guidance keeps established SEO principles central and advises against creating pages for every query variation.

Track observable AI referrals and relevant visibility data alongside Search traffic and conversions. A prompt-monitoring sample is not a count of everything people ask. Where available, Google’s generative AI performance reports provide another view of your pages’ visibility.

Apply the process to your business

Local, ecommerce, B2B and other keyword research.

Local SEO and international markets

Combine services with cities, neighborhoods and service areas. Also check searches without a place name that trigger local results. Review Maps and business listings as well as normal websites. A national volume estimate may hide a small but valuable local need.

Create location pages for places you actually serve and provide useful local details. For a wider local campaign, compare the website and profile work with the Local Stack service. When targeting another country, research its wording, competitors and results separately. Translating the keyword list is only a starting point.

Ecommerce

Research categories, product types, brands, models, features, sizes and use cases. Broad buying searches may need a category page. A model search may need a product page. Comparisons and setup questions usually need supporting content. Plan filters carefully so you do not create many thin pages for every combination.

B2B and SaaS

Ask sales teams about use cases, integrations, alternatives, pricing and objections. Different people in the same company may search during the purchase. A small-volume term can be useful if it brings the right buyer. Measure lead quality and trials or sales, not just visitor counts.

Blogs and new websites

Choose a focused set of questions you can answer from experience. Build useful coverage around one relevant area, then expand. Avoid broad topics solely because they have large volume. Use early Search Console data to improve the first pages and find the next questions.

YouTube and marketplaces

Use the platform’s own searches and data where available. YouTube viewers may want a demonstration; marketplace visitors may want a specific product. Google volume does not measure demand on either platform. YouTube’s Trends tab can help explore searches and content gaps, subject to availability.

SEO versus paid-search keyword research

Both start with the audience, intent and offer. SEO also asks which page can earn organic visibility. Paid search adds bids, budgets, match types and negative keywords that exclude unsuitable ad searches. A phrase-match setting in Google Ads is not an instruction to repeat an exact phrase throughout an SEO page.

Put it all in one place

A keyword research worksheet and worked example.

Keep the full export in one sheet and the page plan in another. The export holds individual terms. The page plan holds one row per keyword group, so close variants do not turn into duplicate content assignments.

Common mistakes and questions

What should you avoid in keyword research?

The common mistakes are choosing by volume alone, confusing ad competition with SEO difficulty, ignoring the search results, publishing a page for every variant, copying competitors without adding value, and measuring success only by rankings.