🔍 SEO / GEO

Keyword Cluster / Semantic Core

semantic core · keyword cluster · keyword clustering · keyword research · query map · keyword map

Definition

A semantic core is a structured database of search queries grouped by thematic cluster and by intent, used as the basis for site architecture and content planning in SEO. It is the Russian-market name for what Western SEO calls keyword research plus clustering.

What a semantic core is

A semantic core is a map of the queries a site should appear for in the search results. Not simply a keyword list, but a structured database in which every query is tied to a specific page.

The term belongs to the SEO methodology that grew up around Yandex, the dominant search engine in Russia and several CIS markets, where a complete, formally documented semantic core is a standard project deliverable. The Western equivalent is keyword research plus clustering plus a keyword-to-page map — the same work, usually delivered as a looser set of artefacts. The practical difference is emphasis: the semantic core tradition treats exhaustive coverage of the query space and a strict one-cluster-to-one-page rule as the starting point of site architecture, not as a later optimisation step.

In e-commerce a semantic core spans every level of the purchase funnel:
– Top of funnel (TOFU): “how to choose a fridge”, “which vacuum cleaner is best” — blog articles
– Middle (MOFU): “coffee machine ranking 2025”, “smartphone comparison” — round-ups and comparisons
– Bottom (BOFU): “buy iPhone 15 256gb”, “order furniture delivery” — categories and product pages

How to collect the semantics

Data sources

Tool	What it gives you
Google Keyword Planner	Volumes for Google; the baseline source in most markets
Google Search Console	The real queries a site already gets impressions for
Yandex Wordstat	Query frequency in Russia and CIS markets; its word-picker expands the semantics
Search suggestions (Google and Yandex)	Long-tail queries and variations
Ahrefs / Semrush / Serpstat	Competitor semantics; the queries other sites rank for

Clustering

Once collected, all queries are grouped on the principle of one intent, one page. For example:

Cluster: "nike mens trainers"
  - nike mens trainers               (19,000/mo)
  - nike trainers for men            (12,000/mo)
  - buy nike mens trainers           (8,500/mo)
  -> Target page: /catalog/nike-mens-trainers/

Clusters with informational intent go to the blog; commercial ones go to categories and product pages.

Semantic cores for e-commerce: the specifics

In online retail the semantic core is unusually large: thousands of SKUs generate thousands of product-level queries. A few principles:

Product pages are promoted on exact queries (model plus SKU plus “buy”). Writing content for every product page by hand is unrealistic — use templates with variables.
Categories are promoted on broad commercial queries (“laptops”, “Samsung smartphones”). The category content — description, H1, meta tags — should carry the cluster queries.
Informational content in the blog is built on TOFU and MOFU queries that commercial pages cannot serve.

Important: do not try to answer informational queries with category pages or the other way round. The search engine sees the intent mismatch and the page does not rank, however many keywords it contains.

Frequently asked questions

Where do you start when building a semantic core for an e-commerce site?

The standard order: 1) collect the head terms, the broad queries, for each product category; 2) expand them with Yandex Wordstat, Google Keyword Planner and search suggestions; 3) add competitor semantics, meaning what competitors rank for, through Ahrefs, Semrush or Serpstat; 4) group everything by intent and by site page. For a large catalogue, automated clustering pays for itself.

How often should a semantic core be refreshed?

Every 6 to 12 months for a baseline review. Triggers for an unscheduled refresh: launching a new category, a seasonal product going live, a change in the search engine's algorithms. Position monitoring against the current core should be monthly.

What is keyword clustering and why is it needed?

Clustering is the splitting of the whole query list into groups, each of which will be promoted through a single page. Queries with the same search intent are merged into one cluster. It prevents cannibalization, where several pages compete for the same query and each eats into the other's position.