In SEO and natural language processing (NLP), we often work with four terms: phrase, keyword, entity, and query. Although each has a distinct meaning, in practice, they often overlap.
What you will find in this article:
Phrase
Any sequence of words. It may or may not be typed into a search engine.
e.g. “winter running shoes”
A phrase can include entities and keywords, but doesn’t necessarily express search intent.
Keyword
A phrase with search potential, important for SEO.
e.g. “running shoes”
This is usually a specific topic we optimize content for, measuring its popularity and competition.
Entity
A concept semantically recognizable by Google, regardless of its wording.
e.g. “Nike” as a brand, “Warsaw” as a city
Entities are fundamental to semantic search and Google’s Knowledge Graph.
Query
A specific phrase entered by a user in the search engine.
e.g. “where to buy Nike running shoes in Warsaw”
Each query may include keywords, entities, and phrases, but its form is variable and unpredictable.
These sets are not always distinct. A good example is “Air Jordan” – it can function as:
- a phrase – it’s just a sequence of characters/words,
- a keyword – if we optimize content for that brand,
- an entity – as Google understands it refers to a specific brand,
- a query – if someone types exactly “Air Jordan” into the search engine.
How do these concepts overlap?
| Example | Phrase | Keyword | Entity | Query |
|---|---|---|---|---|
| Nike | ✅ | ✅ | ✅ | ✅ |
| running shoes | ✅ | ✅ | ❌ | ✅ |
| Adidas UltraBoost | ✅ | ✅ | ✅ | ✅ |
| how to choose running shoes | ✅ | ✅ | ❌ | ✅ |
Understanding these differences and relationships allows for better content planning, alignment with user intent, and the creation of effective SEO and content marketing strategies, especially with regard to AI summaries and search (ChatGPT, AI Overviews and AI mode in Google, Perplexity, Copilot), where semantics and LLM optimization are key.
Contact me if you need more clarification 🙂