Keyword research: building a semantic core for Uzbek websites
The semantic core defines your site structure. A mistake here means six months of content targeting queries that bring no customers.
What a semantic core is
It is the complete list of queries you should be found for, grouped so each group maps to exactly one page.
It is a structural map rather than a word list: it tells you directly how many pages you need and what each covers.
Where to source queries
Google Keyword Planner, search autocomplete, related searches at the bottom of results, and Google Search Console if the site already gets impressions.
Also: questions customers ask your sales team, and queries competitors rank for.
Query types and intent
Commercial queries with 'order', 'buy', 'price' map to service pages. Informational queries with 'how', 'why', 'what is' map to blog articles.
Do not mix them: a blog post will not win a commercial query, and a service page will not win an informational one.
Uzbek specifics: bilingual search
People search in Russian, Uzbek Latin, sometimes Cyrillic, and often mix languages in one query. Tools cover Uzbek poorly, so that semantics is gathered manually via autocomplete.
The Uzbek page must be optimised for Uzbek queries rather than translated Russian keywords — they are different phrase sets.
Clustering and cannibalisation
A cluster is a group of queries Google treats as one topic, identifiable because the same pages rank across them. One cluster, one page.
Cannibalisation happens when two of your pages compete for one cluster: Google cannot decide which to show and both lose ground.