Service · Semantic SEO · Koray Framework

Rank topics,
not
keywords.

Chasing keywords one page at a time is why your traffic plateaued. Google’s NLP reads topics and entities now — not strings. Using Koray Tuğberk GÜBÜR’s Topical Authority framework (Topical Maps, Entity-Attribute-Value coverage, Cost-of-Retrieval, 41 authorship rules) I make you own the entire concept space — so you compound for years without buying a single backlink.

1+N
Topical map size
41
Authorship rules
3mo
Build cycle
EAV
Coverage model
Tanzid · Semantic LeadKoray-trained →
NLP-aligned for
Google NLU
BERT
MUM
ChatGPT
Perplexity
Gemini

K

Framework attribution · Holistic SEO & Digital

This service follows the Semantic SEO framework pioneered by Koray Tuğberk GÜBÜR.

Topical Authority, the Entity-Attribute-Value model, Cost of Retrieval, and the 41 authorship rules referenced here originate from Koray's published work and the Topical Authority Academy. I implement his framework — credit where credit is due.

↗ topicalauthority.digital
What is Semantic SEO?

Search engines
read meaning,
not keywords.

Semantic SEO is the practice of structuring a website so that search engines (and the LLMs trained on their data) understand the meaning, relationships, and authority of your content — not just match a query string against a page.

The keyword era is over. Google's NLU stack — BERT, MUM, and the LLM-augmented ranking layer — reads concepts, entities, and topical relationships. A page that "ranks for one keyword" loses to a site that owns the whole topic graph around it.

Koray's framework gives this a precise shape: a measurable formula, a hierarchical content structure, and 41 writing rules that compound into Topical Authority.

The formula · Koray, 18 May 2022

One equation,
two multipliers.

Topical Authority isn't a metaphor. Koray formalized it as a literal formula — coverage plus history. Both terms compound; neither works alone.
Topical Authority
=
Topical Coverage
+
Historical Data
Term 1 · Topical Coverage
How completely you map the subject.

Every entity, attribute, sub-topic, predicate, and user question within your niche — captured by a Topical Map and shipped as interconnected pages. No content gap, no semantic blind spot.

Term 2 · Historical Data
How long you've been consistent.

Publishing rhythm, engagement signals, dwell time, return visits, content updates. Google's confidence in you compounds the longer you keep the coverage healthy. Newness is a tax; consistency is the offset.

The Topical Map · Koray's pillar

Map the
whole topic,
not the keyword.

A hierarchical mini-Wikipedia of your subject. One macro context at the center, 4–8 meso topics around it, dozens of micro questions at the leaves. Built before a single word is written.
Entity-Attribute-Value model

Teach search
the facts,
not the prose.

Every statement on a page should reduce to an Entity, an Attribute, and a Value. LLMs and Google's NLU layer index these triples — not your adjectives.

The triple is the unit.

An E-A-V triple is the atomic unit of structured knowledge. Example: Tanzid Al Tuhin → occupation → SEO consultant. A page covers an entity by stating its attributes; the more attributes covered with values, the higher its coverage score.

Koray's framework requires complete EAV coverage per entity — not 80%, not "the important ones." When Google's NLU model parses the page, every triple it expects to find should be there, declared in clean prose + schema.

This is also what AI search engines retrieve. ChatGPT pulls triples to assemble an answer. If your page doesn't surface the triple, the citation goes elsewhere.

Entity
Attribute
Value
AI-First SEO
definition
Optimizing for LLM retrieval
AI-First SEO
duration
4-week sprint format
AI-First SEO
cost
discussed on call
AI-First SEO
targets
ChatGPT, Perplexity, AIO, Bing
AI-First SEO
practitioner
Tanzid Al Tuhin
Topical Map
inventor
Koray Tuğberk GÜBÜR
Topical Map
levels
Macro · Meso · Micro
Topical Map
output
Hierarchy of pages + internal links
Generative-Engine Content
format
Pillar + 9 clusters + FAQ hub
Generative-Engine Content
word_count
25,000+ words, hand-written
Cost of Retrieval · Koray's invention

Make it cheap
for Google to
read your site.

Search engines have a finite budget per URL — crawl time, parse time, NLP cycles. Koray's "Cost of Retrieval" frames the goal as minimizing that budget. Lower cost = higher rank likelihood.
High cost · most sites

Google works hard
to understand you.

86/100
  • Unstructured prose, no semantic HTML
  • Zero JSON-LD schema
  • No internal cluster wiring
  • Buried answers, no inverted pyramid
  • Slow load, layout shift, INP > 200ms
  • Inconsistent entity references
  • Low cost · Koray framework

    Google reads you
    in one pass.

    22/100
  • Semantic HTML5 tags throughout
  • Article + Person + FAQPage JSON-LD
  • Hub-spoke topical internal links
  • Answer-first paragraphs (≤60 words)
  • Core Web Vitals green; INP < 100ms
  • Consistent entity + sameAs wiring
  • 9 of Koray's 41 authorship rules

    How a sentence
    gets written.

    The full 41-rule set governs every comma in the content. These nine are the most consequential — the ones that bend an LLM retrieval result in your favor.
    Rule 04

    One macro per page

    Each URL covers a single macro context. No “ultimate guides” that try to cover three topics.
    Example
    Title: "How LLMs decide which sources to cite" — not "AI SEO + Schema + Content Pack"
    Rule 07

    H2 as a user question

    Every H2 phrased the way a real user would type it. Direct question form.
    Example
    "What signals does ChatGPT use to pick a citation?" — not "ChatGPT citation factors"
    Rule 11

    40-word extractive answer

    First paragraph of each section answers the question in ≤ 40 words. LLMs retrieve from here.
    Example
    "ChatGPT cites sources with structured schema, answer-first formatting, and entity ties to authoritative profiles..."
    Rule 14

    Full EAV coverage

    Every entity declared with its full attribute set. Definition, attributes, examples, related entities.
    Example
    Entity = "Topical Map" → Inventor + Levels + Output + Use case — all stated
    Rule 17

    Macro-to-micro flow

    Page opens at the highest level of abstraction, descends into specifics. Inverted pyramid.
    Example
    TL;DR → definition → mechanism → examples → edge cases → references
    Rule 19

    Internal linking with intent

    Anchor text describes the linked page’s subject precisely. No “click here”, no “learn more”.
    Example
    "Read the topical-map methodology guide" — anchor = "topical-map methodology"
    Rule 22

    Source context match

    The page author’s expertise (declared via schema + bio) must match the topic.
    Example
    Article about Semantic SEO authored by Person with sameAs LinkedIn = SEO consultant
    Rule 26

    Sentence-level clarity

    One claim per sentence. Subject + verb + object + (optional) clause. No three-clause sentences.
    Example
    "Topical Authority compounds." (good) — not "Topical Authority, which Koray formalized..., compounds."
    Rule 31

    Predicate consistency

    Use the same verb when stating the same relationship across pages. “is_a”, “has_attribute”, “used_by”.
    Example
    Always: "Topical Map is a hierarchy" — not sometimes "represents" or "describes"
    → All 41 rules applied per page during the engagement. Full rule set is Koray’s, used here under his published framework.
    Historical Data · Term 2 of the formula

    Coverage starts
    the clock. Time
    compounds the trust.

    Building Topical Authority isn't instant. Google needs to verify your coverage is real — by watching you behave like an authoritative source over time.

    What history actually measures.

    Coverage gets you noticed. Historical Data is what makes Google trust the coverage isn’t a one-time burst.
  • Consistent publishing cadence — weekly or bi-weekly
  • Engagement signals — dwell time, scroll depth, returns
  • Updates on old content — proves you tend the garden
  • Brand search volume — people typing your name
  • Mention velocity in trusted sources over time
  • M1M3M6M9M12Coverage shippedTrust thresholdCompounding
    How search engines actually think

    Six stages, one
    decision.

    Every query Google or an LLM receives passes through these six stages of NLU pipeline. Semantic SEO ships content that wins at stages 04–06 — the ones most agencies don't even know exist.
    01
    Tokenization

    Query split into tokens. "best AI SEO consultant" → [best, AI, SEO, consultant].

    → tokens
    02
    Entity recognition

    NLU identifies entities in the query: "AI SEO" (concept), "consultant" (role).

    → entity set
    03
    Intent classification

    Macro intent decided — informational, navigational, transactional, or commercial.

    → intent type
    04
    Context expansion

    Query expanded with co-occurring entities, synonyms, related concepts. Hidden in the model.

    → expanded query
    05
    Document retrieval

    Index probed for pages with matching entity coverage, low cost-of-retrieval, strong source context.

    → candidate set
    06
    Ranking + citation

    Final ordering. For AI Overviews / LLMs, top candidates feed the answer generator — citation chosen here.

    → result + cite
    Source Context · Koray's E-A-V triangle

    Who says it
    matters as much
    as what is said.

    Three vertices form a Source Context: the entity making the claim, the topic the claim is about, and the predicate connecting them. Misalign any vertex and the claim doesn't index as authoritative.
    Entitywho
    Topicwhat
    Predicatehow
    Source
    Context

    The three vertices.

    Koray frames Source Context as a triangle. All three vertices must align for Google’s NLU model to treat your page as authoritative on the topic.
  • EntityWho — the author + brand, declared via Person/Organization schema, sameAs to Wikidata, LinkedIn, GitHub.
  • TopicWhat — the macro context of the page, declared via H1 + first paragraph + topical map position.
  • PredicateHow — the verb connecting entity to topic. "Founded by", "expert in", "consultant for" — kept consistent site-wide.
  • The full workflow · end to end

    Nine stages,
    one pipeline.

    From raw intent to compounding traffic — the entire path. Stages 1–8 run in a 12-week build; stage 9 runs forever. Each stage produces an artifact you keep.
    01

    Search Intent Research

    Week 1
    Mine real user queries, classify intent (informational / commercial / navigational / transactional), and surface the Central Search Intent — the macro purpose every page in the build will serve.
    GSC mining
    AlsoAsked
    AnswerThePublic
    SERP scrape
    ↳ Central Search Intent doc
    02

    Topical Mapping

    Week 2
    Build the full hierarchy. One macro context, 4–8 meso subtopics, 30–60 micro question nodes. Sign-off happens here — no word gets written until the map is approved.
    Manual mappingKoray's Topical Map agentSERP cluster mining↳ Approved topical map
    03

    Entity Research

    Week 3
    Extract every entity in the topic graph. Attach attributes, declare predicates, define values. Build the E-A-V coverage matrix that drives every page.
    Google NLP API
    Wikidata
    Knowledge Graph audit
    Schema.org
    ↳ EAV coverage matrix
    04

    Website Structure Planning

    Week 4
    Map the Topical Map onto your URL hierarchy. Decide pillar slugs, cluster routes, breadcrumb model, and the internal link graph before any page exists.
    Sitemap diagramsURL conventionsIA review↳ Information architecture
    05

    Semantic Content Creation

    Weeks 5–10
    Hand-write each page following Koray’s 41 authorship rules. One macro per page, H2s as user questions, 40-word extractive answers, full EAV coverage, consistent predicates.
    41 authorship rulesEAV matrixNLP scoring↳ 40–60 drafted pages
    06

    On-Page Optimization

    Week 10
    JSON-LD wrap per entity, semantic HTML5 tags, meta descriptions phrased as extractive answers, image alt with entity references, Core Web Vitals tuning.
    JSON-LD
    Semantic HTML
    Schema validator
    Lighthouse
    ↳ Schema + technical on-page
    07

    Internal Linking

    Week 11
    Build the semantic content network. Anchor text describes destination intent precisely. Hub-spoke + lateral cluster links + breadcrumb closure — every node reachable in ≤ 3 hops.
    Link graph auditAnchor dictionaryOrphan check↳ Internal link graph
    08

    Authority Building

    Weeks 11–12
    sameAs wiring across Wikidata, LinkedIn, Crunchbase, GitHub, Medium. Author bylines + Person schema. Optional Digital PR layer to seed external entity references.
    Wikidata
    LinkedIn
    sameAs schema
    PR layer
    ↳ Authority profile network
    09

    Continuous Optimization

    Week 12 + ongoing
    Koray’s rule: audit every page every 6 months. Re-tune for new query patterns, refresh EAV gaps, update anchor text as NLP models evolve. Historical Data starts compounding here.
    Citation tracker6-month audit cadenceTopic Consolidation↳ Configuration calendar
    9
    Stages, all owned by me
    12w
    Build cycle
    Continuous loop
    9
    Artifacts you keep
    Deliverables · what you keep

    Every artifact,
    in your CMS.

    No vendor lock-in. The map, the schema, the rules dictionary, the configuration calendar — all yours.

    Map + drafting half

  • Central Search Intent + customer language report
  • Complete Topical Map — macro / meso / micro hierarchy
  • Entity-Attribute-Value coverage matrix (per page)
  • Source Context audit + author byline schema
  • 40–60 page drafts per Koray's 41 rules
  • Predicate dictionary — consistent verb usage
  • Schema + ship half

  • JSON-LD schema wrap on every page
  • Hub-and-spoke internal linking graph
  • Cost-of-Retrieval profile + Core Web Vitals fix
  • sameAs entity wiring to Wikidata + LinkedIn
  • 6-month content configuration calendar
  • Topical Authority tracker (rankings + AI citations)
  • Who is this for?

    Is the framework
    for you?

    Six profiles where Semantic SEO returns the most. If none fit, the call ends with me saying so — no soft sell.
    01

    Sites that backlinks can't fix

    You've spent on links, gone nowhere. Coverage + history is the lever you haven't pulled.

    ★ Best fit
    02

    Founder-led SaaS in a focused niche

    Narrow vertical, defined audience, defensible expertise. The topic map maps cleanly.

    ★ Strong fit
    03

    Affiliate / publisher building a niche

    Coverage-driven verticals (recipes, tools, software, reviews) — Koray's case studies live here.

    ★ Strong fit
    04

    Hit by a Helpful Content Update

    HCU penalizes thin coverage. The framework rebuilds the comprehensiveness Google now measures.

    ★ Urgent fit
    05

    B2B with long buyer research cycles

    Decisions made over weeks; buyer reads 10+ pages. Topical Authority owns that whole journey.

    ★ Strong fit
    06

    Site planning to leverage programmatic SEO

    Programmatic at scale only works on a healthy semantic foundation. Map first, multiply second.

    ★ Smart fit
    FAQ · 8 questions

    The honest answers.

    If yours isn't here, email hello@tanzidaltuhin.pro — answered within 24 hours, by me, not a bot.
    01Why credit Koray Tuğberk GÜBÜR so prominently?

    Because honesty matters. Topical Authority, the EAV model, Cost of Retrieval, and the 41 authorship rules are his framework — published, taught, and battle-tested in his case studies. I learned this from him; I implement it for clients; I credit him for it. Anyone selling "semantic SEO" without acknowledging Koray either invented something different or is borrowing without attribution.

    02Did you study under Koray directly?+

    I went through the Topical Authority Academy and applied his framework on multiple client projects. I'm not on his team, I don't represent Holistic SEO & Digital, and I don't claim certification. I'm a practitioner of his methodology — the way a developer might be a practitioner of Test-Driven Development without working at Pivotal.

    03How long until I see results from Semantic SEO?+

    The Topical Coverage lift is visible 60–120 days after the map ships. The Historical Data multiplier is what compounds — typically you see meaningful compounding from month 6 onwards. Koray's published case studies show 0 → 128,000 organic sessions in 123 days, but that's the high end on a perfectly mapped niche.

    04Will this work for an existing site with a lot of content?+

    Often the right move is consolidation, not addition. We map your current pages onto the Topical Map, identify gaps, identify cannibalization, and sometimes delete or merge 20–30% of pages. Koray calls this "Topic Consolidation" — and it usually lifts traffic by itself.

    05Does this replace classical SEO?+

    No — it absorbs it. Technical SEO, Core Web Vitals, schema, and on-page basics are part of Cost of Retrieval. Backlinks still help. The framework just changes the priority: coverage first, history second, links third.

    06Can you do this for Bangla content?+

    Yes. The framework is language-agnostic. EAV, topical maps, predicates work the same in Bangla. Many Bangladeshi affiliate and B2B sites are wide-open semantic opportunities because no one's mapped them properly yet.

    07Why doesn't every SEO agency do this?+

    Because it's slow, demanding, and doesn't resell. A topical map takes weeks. 41 authorship rules take training to internalize. Most agencies sell packages: 4 blog posts, 10 backlinks, monthly. Koray's framework requires the agency to be a bottleneck on quality, not a volume play.

    08What happens after the 12-week build?+

    Content Configuration starts — Koray's rule is "audit every page every 6 months." We move to a quarterly cadence: add micro-coverage where gaps emerge, retune anchor text, refresh EAV. Most clients stay on retainer for this; some take the calendar and run it internally.

    Free topical-coverage audit · Q3 2026

    See the gaps
    holding you flat.

    Send me your site and I’ll map your current topical coverage against your niche — free — and sketch the macro context you’d need to own to start compounding. The sketch is yours to keep whether or not we work together.

    Scroll to Top