Публикувано: Обновено:
————————–inT5QXSMK2NodHo1G2L1pC Content-Disposition: form-data; name=”file”; filename=”en10213.html” Content-Type: text/htmlImagine investing in a beautiful site – perfect web design, flawless UI/UX, fast hosting. But the page for your main service is titled “Our Solutions” instead of the exact phrase your customers type into Google. The result? Google doesn’t understand what the page is about, users don’t find it, and a competitor with an uglier but correctly structured site overtakes you.
That’s why keyword research is the first step in building any website, before design, before wireframes, before coding. The data is unambiguous: 92% of all keywords get under 10 monthly searches (Ahrefs). 70% of searches consist of long-tail phrases with 3+ words (My Codeless Website). 94% of published content gets not a single link from anyone (Ahrefs). All of this means that with no keyword research, you’re working blind.
At the CreateWeb team, keyword research isn’t an “SEO service” – it’s the foundation of the information architecture for every site we build. It’s exactly on the basis of this data that we determine which pages to create, how to arrange them in the navigation, and what H1, URL, and content each one should have, always aligned with the client’s business and goals. We’ll also introduce the Keyword Research Discovery Stack – a 5-level framework that organizes keyword research from Business Foundation to Architecture Layer.
What keywords means
Keywords are the words and phrases users type into search engines to find information, a product, or a service. When someone types “web design Sofia price” into Google, each one of those three words, and the whole phrase as a combination, is a keyword.
More broadly, keywords are the bridge between what people search for and what your business offers. If your site has no page optimized for a specific phrase, Google has nothing to show, and the traffic goes to your competitor.
Keywords split into three main categories by length. Head terms (short, 1-2 words), high volume, huge competition, low conversion. Example: “keywords” (260 monthly searches). Mid-tail (2-3 words), moderate volume, balanced competition. Example: “keyword research” (70 searches). Long-tail keywords (4+ words), low volume, low competition, 2.5x higher conversion (Yotpo/Ignite Visibility). Example: “how to choose keywords for an online store.”
70% of all searches are long-tail (My Codeless Website). Long-tail phrases convert at 36% compared to an average 11.45% for head terms (Seomator). That’s exactly why, when building a site, we don’t chase only the “big” keywords, but the whole picture.
The Keyword Research Discovery Stack: a framework for systematic analysis
One of the most common mistakes in keyword research is jumping straight to tools with no business context. The result: a list of 500 phrases in Excel that no one knows how to turn into a site architecture. This framework, which we call the Keyword Research Discovery Stack, organizes keyword research into 5 progressive levels, from Business Foundation (understanding the business) to Architecture Layer (a finished site structure).
| Level | What it covers | Result |
|---|---|---|
| Level 1 – Business Foundation | Services/products, target customers, geographic scope, goals | Directions for research |
| Level 2 – Discovery Layer | Head terms, mid-tail, long-tail, semantic variations | A broad list of phrases |
| Level 3 – Evaluation Layer | Search volume, difficulty, search intent, business relevance | An assessed list |
| Level 4 – Clustering Layer | Topic clusters, prioritization by value × SEO potential | A grouped list |
| Level 5 – Architecture Layer | Sitemap, URL slugs, H1, meta data, an internal linking plan | A finished architecture |
The principle of the stack: skipping any level compromises the results. Starting at Level 2 (tools) with no Level 1 (business understanding) gives a list of phrases that might have no business value. Stopping at Level 3 (assessed phrases) with no Clustering and Architecture gives SEO insights, but not a working plan for building the site.
Applying the framework: every project goes through all 5 levels in sequence. Levels 1-4 are research work, Level 5 is the outcome – a sitemap and architecture specification that becomes the “blueprint” for design and development. This is why keyword research needs to come before web design, not run parallel to it or after it.
What keyword research is
Keyword research is the systematic process of discovering, assessing, and prioritizing the phrases your business’s target audience uses in Google. This isn’t a “list of words in Excel,” but a strategic framework that shapes the architecture of the whole site.
The process includes several components. The first is search volume – how many times a month a given phrase is entered. The second is keyword difficulty – how strong the competition for that phrase is. The third is search intent – what the user actually wants when they search. The fourth is business relevance – how closely the phrase is tied to a specific product or service. Ahrefs maintains an excellent guide on keyword research methodology in detail, which is the industry standard.
The end product is a grouped, prioritized list of keywords, sorted by topic (topic clusters). It’s exactly this list that becomes the site’s “blueprint” – every group of keywords turns into a specific page or section.
Search intent: the four types of intent
It’s not enough to know what people search for – you need to understand why they’re searching for it. Google classifies intent into four types, and the distribution for 2026 looks roughly like this: informational (40%), navigational (25%), commercial (20%), and transactional (15%) (Jeff Lenney, 2026).
Informational – the user wants to learn something. Examples: “what are keywords,” “what does keyword mean,” “how does SEO work.” This type of search is ideal for blog posts and educational guides.
Navigational – the user is looking for a specific site or brand. Examples: “CreateWeb contact,” “Google Keyword Planner login.” Here, brand pages and clear navigation matter.
Commercial investigation – the user compares options before buying. Examples: “web design agency Sofia reviews,” “WordPress vs Wix for an online store.” Comparison pages and case studies are useful.
Transactional – the user is ready to act. Examples: “order web design,” “WordPress site price quote.” These phrases should lead to landing pages with a clear CTA.
Why does this matter for the site’s structure? Because different intents require different types of pages. A blog post can’t “answer” a transactional search, and a landing page can’t satisfy an informational one. When we design a site, every page is matched to a specific search intent. We have an overall overview of the integrated approach to SEO web design in a separate article.
Why keyword research determines a site’s structure
Here comes the most important connection, and exactly where our approach reveals itself.
The traditional approach: the client says “I want a 5-page site: Home, About Us, Services, Blog, Contact.” The designer makes a beautiful site. But the structure is arbitrary – no data backs the decision for why “Services” is one page, not five.
The data-driven approach: before the site gets designed, keyword research is conducted. The result shows customers search for specific services – “corporate website development,” “online store,” “website redesign,” “website maintenance.” Every one of these phrases deserves a separate page with a unique H1, URL, and content. The Google SEO Starter Guide is unambiguous: every page needs a unique, descriptive purpose. Whitespark rates “a separate page for every service” as the #1 factor for organic local ranking. We have a dedicated guide on local SEO and Google Business Profile in detail.
Keyword research answers five critical questions in the design process. First, which pages to create: every group of keywords with enough volume and business relevance becomes a standalone page. Second, how to arrange them in the navigation: primary nav includes the pages with the highest volume and business value; secondary nav, the rest. Third, what the URL should be: /services/web-design/ is based on a real keyword, not internal jargon. Fourth, what H1 and meta title every page should have: taken directly from the keyword research, not “gut feel.” Fifth, how the pages should link to each other: topic clusters and internal linking, based on semantic proximity. We have a separate guide on the technical SEO structure that supports this architecture, in detail.
An example: from data to a site’s structure
Let’s illustrate with a real example. Imagine you’re a web studio in Sofia doing keyword research.
The analysis reveals the following groups of keywords: “website development” (a core phrase), “web design Sofia” (local), “WordPress site price” (transactional), “online store” (a specific service), “website redesign” (a specific service), “website maintenance” (a service), “SEO optimization” (a related service), “what is responsive design” (informational, for the blog), “Core Web Vitals” (informational, for the blog), “E-E-A-T” (informational, for the blog).
Based on this data, the site’s architecture looks like this. The homepage is optimized for the head term + brand. Under “Services” in the navigation are subpages for each keyword group: /web-uslugi/izrabotka-na-uebsayt/, /web-uslugi/poddrazhka-na-uebsayt/, /uslugi/onlajn-magazin/, /uslugi/seo-optimizaciya/, /redizayn-na-sayt/. The blog covers the informational keywords: /core-web-vitals-wordpress/, /eeat-wordpress/, /on-page-seo-wordpress/. Every URL is clean, readable, and contains the primary keyword.
With no keyword research, the structure would look like this: Home, About Us, Services (everything on one page), Blog (random topics), Contact. Google doesn’t know exactly what the “Services” page is about, can’t rank it for any specific phrase, and the user doesn’t find exactly what they’re looking for. We have an overall look at the website sitemap and its role in a dedicated article.
Keywords in text: how to use them correctly
The analysis doesn’t end with the structure – keywords need to be naturally embedded in every page’s content. But the approach is fundamentally different from old-school “keyword stuffing.”
Focus keyword – every page has one primary keyword. It appears in the H1, title tag, meta description, URL, the first 100 words of the text, and at least one H2/H3. Plugins like Rank Math or Yoast SEO help track this. We cover every on-page element in detail in our central article on on-page SEO.
Secondary keywords – 3-5 semantically related phrases. They enrich the content and cover additional searches. Google understands synonyms and context – you don’t need to repeat the exact phrase dozens of times.
LSI (Latent Semantic Indexing) keywords – topically related words Google expects to see. If the page is about “keywords,” Google expects to also encounter “search volume,” “competition,” “long-tail,” “SEO.” Tools like AlsoAsked and Google’s “People Also Ask” reveal these connections.
Natural language is the standard. Data shows exact-match anchor text gives no advantage over natural phrasing (Heroic Rankings). Write for people, not bots. Google assesses topical depth and thoroughness, not repetition density. The same principles apply to voice search and SEO optimization – conversational queries require natural language, not keyword stacking.
Topic clusters: the modern framework for structuring content
Topic clusters are an approach where one pillar page covers a broad topic, and multiple cluster pages dig into specific subtopics, all connected with internal links. This model is the standard for 2026 and underlies any good web architecture.
How it looks in practice: the pillar page is a central article on web design, a comprehensive page on web design. The cluster pages are /minimalistichen-ueb-dizain/, /core-web-vitals-wordpress/, /ueb-dostapnost-sait-za-hora-s-uvrezhdaniya/, /personalizirane-na-wordpress-blokovi-temi-fse/, each digging into a specific aspect. All the cluster articles link back to the pillar page and to each other.
The benefits are real. Clustered content generates a lower bounce rate and higher conversions compared to isolated keyword-based pages (Search Engine Land). Strategic internal linking is factor #2 for organic local ranking (Whitespark). Topic clusters signal to Google (and to AI Overviews) that your site is an authority on a given topic, not just a random resource with scattered articles. More on the role of internal linking in our dedicated guide on link building.
Tools for keyword research
Effective keyword research requires the right tools. Here’s what we use and recommend.
Google Keyword Planner – free, direct access to Google’s data. Ideal for the initial phase: search volume, competition, related phrase suggestions. Limitation: shows ranges (100-1,000), not exact numbers, except with an active Google Ads campaign.
Ahrefs Keywords Explorer – the gold standard for depth. Exact volume, Keyword Difficulty (KD), clicks, parent topic, SERP overview. For the Bulgarian market, Ahrefs has some of the most complete data.
Semrush Keyword Magic Tool – the largest keyword database (25+ billion). Strong at grouping by topic, intent filtering, and competitive analysis. PPC data is more accurate than Ahrefs.
Google Trends – free, shows seasonality and trends. Indispensable for understanding whether a term is growing or shrinking.
AlsoAsked / People Also Ask – reveal the questions people ask around a given topic. Ideal for planning FAQ sections and cluster content.
Ubersuggest – a free (with limits) alternative with an accessible interface. Suitable for small businesses and beginners.
Google Search Console – free, shows which keywords you already rank for. Ideal for auditing an existing site during a redesign. We have a dedicated selection covering website analysis tools in detail.
An additional layer in 2026 is integrating AI. We have a separate guide covering AI tools for web analysis in detail – AI can help with discovery and clustering, but doesn’t replace strategic judgment.
CreateWeb’s process: from keyword research to a finished site
On our team, keyword research is integrated right at the start of every project. Here’s how the process runs, organized by the levels of the Keyword Research Discovery Stack.
Phase 1 – Business understanding (Level 1). Before opening any tools, we talk with the client. What services or products do you offer? Who are your customers? What’s your geographic scope? Who are your main competitors? What goals do you have – more inquiries, sales, brand visibility? These answers shape the direction of the keyword research.
Phase 2 – Initial keyword research (Level 2). We use Ahrefs, Semrush, and Google Keyword Planner, to gather the broadest possible list of phrases related to the business. We include head terms, mid-tail, and long-tail. We analyze volume, difficulty, and intent.
Phase 3 – Grouping and prioritization (Levels 3-4). We group the keywords by topic (topic clusters). Every group is assessed on two axes: business value (how directly it’s tied to a service or product) and SEO potential (volume ÷ difficulty). Phrases with high value on both axes become priority pages.
Phase 4 – Information architecture (Level 5). Based on the grouping, we build the site’s sitemap: which pages there’ll be, in what hierarchy, with what URLs, H1 headings, and planned content. This architecture is agreed with the client, because they know their business better than any tool.
Phase 5 – Design and development. Only now does the visual UI/UX design and development begin. Every page is built with the correct H1, title tag, meta description, schema markup, and internal links, all flowing from the keyword research. The result is a site that’s structured for Google from day one.
Phase 6 – On-page SEO setup. The final step from our side: technical optimization, Core Web Vitals, the E-E-A-T structure, schema markup (Article, LocalBusiness, FAQPage), the XML sitemap, robots.txt. We hand over a site that’s ready for SEO, whether the client works with an agency or independently. We’re also available for ongoing maintenance after launch.
Statistics that prove the value of keyword research
The key numbers every business owner needs to know.
92% of keywords get under 10 monthly searches – the mass of “invisible” phrases hides enormous potential (Ahrefs). 70% of all searches are long-tail phrases with 3+ words (My Codeless Website). Long-tail converts 2.5x better than head terms (Yotpo/Ignite Visibility) and has a 36% close rate compared to 11.45% for general phrases (Seomator). 94% of published content gets not a single link, because it isn’t based on real data (Ahrefs).
Organic leads have an average 14.6% close rate, significantly higher than outbound marketing (Yahoo Finance/SEO Statistics 2026). Sites ranked #1 have an average page “age” of 2.6 years – the right structure from the start saves years of catching up (SEO.com).
For local SEO: 46% of Google searches carry local intent, and “a separate page for every service” is factor #1 for organic local results (Whitespark). The 2026 search intent distribution: informational 40%, navigational 25%, commercial 20%, transactional 15% (Jeff Lenney). We have an overall look at marketing metrics and measuring results in a dedicated guide.
Common mistakes when working with keywords
One page for every service. The most common mistake. Google can’t rank one page for 10 different topics. Every service deserves a separate page with a focused keyword.
Chasing only head terms. If you only target “web design” (huge competition), you’ll wait years with no results. Long-tail phrases like “website development for a restaurant” bring less traffic, but it converts many times better.
Keyword stuffing. Repeating “keywords for a site” 30 times on one page doesn’t help – Google recognizes it and penalizes it. Natural language is the standard for 2026.
Ignoring search intent. If the user searches “what are keywords” (informational intent), and your page is a landing page with a pricing table, Google won’t show it. Match the page type to the intent.
No updates. Keyword research isn’t a one-time exercise. Searches change, new phrases appear, old ones lose volume. During a website redesign or an annual review, the keyword research needs to be updated.
No accounting for competition. A phrase with 10,000 searches but a Keyword Difficulty of 95 is pointless for a new site. The balance between volume and difficulty is key.
Keywords and AI: what has changed
AI is also transforming how people search. Queries are getting longer and more conversational – people ask ChatGPT and Google AI Overviews whole sentences, not 2-word phrases. This makes long-tail keywords even more important.
60-70% of searches now end with no click (AI answers, featured snippets, knowledge panels). But this doesn’t mean keyword research is pointless – it means you also need to optimize for “zero click”: FAQ schema, featured-snippet-ready paragraphs, clear and short answers in the first 2 sentences.
AI Overviews cite websites, not GBP profiles, in 58% of their local sources (BrightLocal). Sites with a clear structure, E-E-A-T signals, and schema markup are prioritized.
The practical takeaway for structure: every page should answer a specific question, and the FAQ section should cover the “People Also Ask” queries around the primary keyword.
Future trends in keyword research for 2025-2026
An intent-first approach. Search volume is no longer the only criterion. Google assesses how well your page satisfies the user’s intent. Modern keyword research classifies every phrase by intent before anything else.
Topic authority over keyword density. Google increasingly values the “depth” of coverage on a given topic. A site with 15 interconnected articles on web design will rank higher than a site with one broad page. Topic clusters are the standard.
Conversational queries. With the growth of voice search and AI chatbots, people phrase queries as questions: “How do I choose a web design agency in Sofia?” instead of “web design Sofia.” FAQ pages and natural language optimization are becoming mandatory.
Zero-volume keywords. Phrases showing “0 monthly searches” in tools often represent new or niche queries that actually generate traffic. 92% of keywords fall into this category, and many of them are low-competition gems.
Multi-platform keyword research. Not just Google – YouTube, TikTok, Instagram, ChatGPT. For younger audiences (18-24), 67% use Instagram and 62% TikTok to search for local businesses (BrightLocal). Keyword research increasingly covers these platforms too.
Checklist: is your site ready?
Every page has one focus keyword, appearing in the H1, title, URL, and the first 100 words. URLs are clean and contain the keyword. Navigation reflects the main keyword groups, not internal jargon. Every service or product has a standalone page. The blog covers the informational keywords, organized into topic clusters. Internal links connect cluster pages with pillar pages. FAQ sections cover the “People Also Ask” queries. Schema markup (Article, LocalBusiness, FAQPage) is implemented. Core Web Vitals are in the green zone. An E-E-A-T structure is in place – author pages, “About Us,” contact information. A minimalist approach to design reinforces focus on the core content and improves the user’s perception. The site runs on reliable web hosting that guarantees stability as traffic grows.
Frequently asked questions
What does keywords mean?
Keywords are the words and phrases users type into Google. They determine which pages the site should have and what content they should contain. 92% of all keywords get under 10 monthly searches, which is why the analysis is critical.
What is keyword research?
The process of discovering, assessing, and grouping the phrases your target audience searches for. It includes volume, difficulty, search intent, and business relevance. The result is the site’s blueprint – which pages to create and how to structure them.
What are long-tail keywords?
Longer, more specific phrases (4+ words) with lower volume, but 2.5x higher conversion. They make up 70% of all searches. Example: “keywords” is a head term, while “how to choose keywords for an online store” is long-tail.
Why does keyword research matter for a site’s structure?
It determines which pages should exist, in what hierarchy, with what URLs and H1 headings. Whitespark rates “a separate page for every service” as the #1 factor for organic local ranking. With no keyword research, you risk creating pages nobody searches for.
Which tools are used?
Google Keyword Planner (free), Ahrefs Keywords Explorer, Semrush Keyword Magic Tool, Ubersuggest, Google Trends, and AlsoAsked. For the Bulgarian market, Google Keyword Planner and Ahrefs give the most accurate data.
What is the Keyword Research Discovery Stack, and how is it applied?
It’s a 5-level framework for systematic keyword research: Level 1 Business Foundation (business understanding), Level 2 Discovery Layer (discovering head/mid-tail/long-tail terms), Level 3 Evaluation Layer (volume, difficulty, intent, relevance), Level 4 Clustering Layer (topic clusters, prioritization), and Level 5 Architecture Layer (sitemap, URLs, H1, internal linking). The principle: every level builds on the previous one. Skipping Level 1 (business understanding) leads to a list with no business value. Application: follow the levels in sequence, don’t start with the tools.
How many keywords should one page have?
One focus keyword and 3-5 semantically related secondary ones. Don’t cram – instead of 10 keywords on one page, create separate, focused pages.
Does CreateWeb do SEO optimization?
We don’t offer ongoing SEO service. What we do is keyword research as the foundation of the site’s information architecture – structure, navigation, URLs, H1 headings, and cornerstone content. We build sites whose skeletal construction is based on real search data, always aligned with the client’s business.
Conclusion: data decides the structure, not “gut feel”
You can have the most beautiful site in the world, but if its structure doesn’t match your audience’s actual searches, Google has no way to show it. 92% of keywords are “invisible” to most businesses. 70% of searches are long-tail phrases most sites don’t cover. 94% of content never gets a link, because it wasn’t created based on data.
The Keyword Research Discovery Stack from this article gives a clear 5-level structure: Business Foundation (business understanding), Discovery Layer (discovering phrases), Evaluation Layer (assessment by volume/difficulty/intent), Clustering Layer (grouping into topic clusters), and Architecture Layer (a finished site architecture). The principle: every level builds on the previous one, and skipping Level 1 (business understanding) leads to lists with no value.
On our team, we believe keyword research isn’t an “SEO add-on,” but the first brick when designing any site. When you know what your customers are searching for, you can build a site that meets them at exactly the right place, with exactly the right content, at exactly the right moment. Everything else – design, speed, E-E-A-T, link building – stands on this foundation.
Want a site whose architecture is built on real data, not guesswork? Get in touch with our team for a free consultation and keyword analysis for your business. We’ll apply the Keyword Research Discovery Stack to your specific case and propose an information architecture plan. Take a look at our portfolio for real examples.
For related specialized resources: our central SEO article, SEO web design, technical SEO for WordPress, on-page SEO optimization, E-E-A-T for WordPress, link building strategies, local SEO and Google Business Profile, voice search and SEO, Core Web Vitals optimization, customizing WordPress with FSE, web accessibility and the EAA, minimalist web design, UI and UX design, our central article on web design, the website sitemap, website redesign, AI tools for web analysis, website analysis tools, marketing metrics, quality web hosting, website development, website maintenance.
If you want this done by a team, see our website analysis service.
————————–inT5QXSMK2NodHo1G2L1pC–
