Kazanan
How to Do Keyword Research
Keyword research connects the language customers use with a business goal. Its useful output is a prioritized page plan explaining which need each page answers, rather than a large spreadsheet of phrases. This guide combines intent, data sources, tool selection and measurement limits.
1. Start with the business goal and seed terms
Content marketing planning begins with a sale, demo, enquiry or learning objective. Record the target country, language, customer group and existing pages. A consumer in Turkey and an English speaking procurement team may use different vocabulary and decision criteria.
Collect seed terms from sales questions, support tickets, internal site searches and customers’ descriptions of their problems. Expand “CRM” through industry, integration, feature and use case. Separate branded from unbranded searches; a brand mention does not automatically imply purchase intent.
Short tail describes a broader topic and long tail a more specific need; phrase length alone does not establish the difference. Related concepts help explain the subject. Do not present them as a separate Google system called “LSI keywords” or assign an unsupported percentage of all searches to long tail.
2. Match intent to the page format
A B2B SEO plan needs different answers for learning, comparison and purchasing queries. These are starting categories, and results can combine several intentions. Check a tool’s suggested label against the actual current results before choosing a page type.
| Intent | Example | Possible page |
|---|---|---|
| Informational | What is CRM? | A guide with a definition, examples and practical use cases. |
| Navigational | Product name login | The official login or help destination the user needs. |
| Commercial investigation | CRM tools comparison | A comparison explaining criteria, differences and limitations. |
| Transactional | Request a CRM demo | A product or service page with clear scope, form and next step. |
3. Read Google data in context
Add country, dates and scope to your content measurement checklist. Different numbers for the same term do not necessarily indicate an error: tools use different samples, groupings and time periods.
Keyword Planner
Search Console
Trends and search results
4. Choose paid tools for the work you need
Connect the existing pages and indexing findings from a technical SEO audit to your research. Test the same country, terms and competitors across candidates. Check export, user and update limits before buying. No difficulty score guarantees success.
| Tool | Useful workflow | Important limit |
|---|---|---|
| Ahrefs Keywords Explorer | Google research, Parent Topic, Traffic Potential and SERP inspection. | KD reflects links to leading results; traffic potential is not a forecast for your website. |
| Semrush Keyword Magic Tool | Country/language, questions, intent and topic grouping filters. | KD and personal difficulty are vendor models; review intent labels manually. |
| Ubersuggest | Keyword ideas, competitor gaps and results analysis. | Free access is limited; volume and difficulty are estimates. |
| Moz Keyword Explorer | A candidate to evaluate for keyword discovery. | Verify its current interface and metrics; do not assume an old Priority formula remains valid. |
| KWFinder | Keyword and competitor domain research. | Check location, language, SERP date and plan limits. |
5. Turn competitors and metrics into priorities
For ecommerce SEO research, commercial competitors may differ from websites appearing in search. Inspect each query in the target country and language. Are categories, products, guides, video or local results prominent? Identify information you can provide instead of simply copying competitors.
Record query, intent, country/language, data date, volume, trend, difficulty, existing URL, business relevance and production effort. Average volume can conceal seasonality. CPC reflects advertising demand, not your profit. Difficulty scores use different methods; avoid treating a single universal threshold as a publishing rule.
Prioritize needs your product serves and your team can answer convincingly. A competitor gap might be a useful opportunity or an irrelevant market. Writing more words or selecting a low score does not secure rankings. Check whether you have the product knowledge, evidence and useful next step to meet the need.
6. Build topic clusters and a page map
In local SEO, match services and areas to the business’s actual scope. Before creating similar pages for every city or phrase variant, decide whether a distinct user need justifies another URL. Connect a central topic page to supporting explanations where the connection helps readers.
Ecommerce
B2B and SaaS
Local services
Editorial websites
7. Improve the page with clear language and evidence
During content production, make the main need clear in the title, introduction and relevant headings. Write a useful description, readable URL and alternative text describing an image’s function. Avoid changing an established URL merely to add a keyword; plan redirects when a change is necessary.
Do not target a fixed keyword density. Artificial repetition makes a page harder to read. Check which job each similar page performs: two URLs appearing for one query do not automatically establish a penalty. Where genuine overlap creates confusion, consider consolidation, clearer scope or better linking.
AI Overviews and AI Mode do not require a special keyword file or mandatory new schema. Conversational questions, useful tables and descriptive product images can help readers. They do not guarantee voice search, featured answers or image visibility.
8. Measure search visibility and business outcomes separately
Use the measurement plan to compare Search Console query and page data by country, device, brand and comparable dates. GA4 reports landing page engagement and key events separately. Linking the systems does not produce a complete direct revenue attribution for every search query.
Keep location, device and sampling frequency consistent in rank tracking tools. If impressions rise while clicks fall, examine intent, result appearance and the page title; this alone does not demonstrate lost sales. Log changes and reassess when sufficient data accumulates. Avoid promising a fixed success date or a first position.
Frequently asked questions
Anything else? Write to us.
Can I start without a paid tool?
Use Search Console, Trends and manual result inspection first. Check Planner access conditions; creating accounts or setting up billing is not an automatic research step.
How many keywords should a page target?
Define one main need and the related questions necessary to explain it. There is no fixed keyword count, density or mandatory LSI list; scope depends on intent.
Is the highest volume always the best choice?
No. Consider product fit, result format, competition, seasonality and decision stage together. A smaller relevant need can be more useful than broad demand you cannot serve.
Is Planner competition the same as SEO difficulty?
No. Planner measures advertiser competition. SEO tools use different proprietary models, which cannot fully describe your website’s circumstances or predict the final ranking.
How often should research be updated?
Review when products, target markets, demand or search results change. Set a regular review schedule and compare seasonality before reacting to a single declining month.
When should results appear?
Indexing, competition, content and technical conditions affect timing. There is no fixed month guarantee. Check publishing and collection first, then visibility and business outcomes.
Turn research into an actionable content plan
Review your target market, existing pages and the business outcome you need to measure.
