Search Intent: Why It Matters More Than Volume

Picture this: you run a tea shop. Someone walks in and asks for "green tea." What do they want? Maybe to buy a pack. Maybe to learn how it differs from black tea. Maybe they're looking for a place to have a cup nearby. The same set of words — three completely different intentions.

In search it's the same. A query is words; intent is the intention behind those words. If a page answers a different intention than the one the person came with, it drops out of the top results, even if it's perfectly optimized for the words. Search engines can now recognize intent better than we can ourselves.

The four types of intent

Any query can be assigned to one of four classic types (or to a mixed one, see below).

1. Informational — "I want to learn"

The person is looking for information. Not to buy, not to go anywhere, but to figure something out.

Examples: "what is the flu," "how to brew matcha," "semantic core in plain words," "the difference between 4G and 5G," "causes of back pain."

What it's served by: a long article, an FAQ, a tutorial video, a Wikipedia article. No "buy right now."

2. Commercial (or transactional) — "I want to buy"

The person is choosing what to buy or who to buy from.

Examples: "buy a down jacket moscow," "iphone 15 pro 256gb," "pizza delivery nearby," "book a dentist appointment," "birthday cake."

What it's served by: a product/service page, a price list, an order form, reviews, a "buy" button.

3. Navigational — "I want to get there"

The person already knows the brand/site and is heading to it.

Examples: "vkontakte," "ozon account login," "semantic core semantic core" (branded), "mail.ru login," "avito."

What it's served by: the brand's own homepage. It's almost pointless for third-party sites to enter this niche — unless you're writing a review/guide about that brand.

4. Mixed — "all of it at once"

A single query can carry several intentions. "iPhone 15" — someone wants to buy, someone wants to read a review, someone wants to compare it with the 14. Search engines handle such queries through diversity in the results — the top will hold stores, reviews, and comparisons all at once.

If you land on a mixed intent, you have to decide which subtype you're going for and cover it more deeply than anyone else.

The main rule

A page must serve the intent the user came with. If the intent is off, the page won't win, even if it's perfectly optimized.

An everyday example. You're a barista; you have a coffee shop brand and a blog about roasting. A user googles "buy coffee moscow." That's commercial intent: the person wants to buy coffee now. If you give them a long post about the origins of Arabica, they'll close the tab in 5 seconds and head to Ozon.

The search engine sees this "5-second exit" (as a signal), concludes "the page didn't fit," and drops it down in the results.

How to determine intent when it's unclear

The universal method for any query is to open the results and look at what's currently in the top 5. The search engine has already chosen what best answers this query. If the top is all stores, the intent is commercial. If it's articles, it's informational. If it's maps, it's local intent, and so on.

This is the "gold standard" for determining intent: you don't ask yourself what should be in the results, you look at what's already there.

Signals in the words themselves

Sometimes you don't feel like checking the results, and you can estimate from the words.

This is a guideline, not a law. For tricky queries, it's better to check the results.

Why intent matters more than volume

Volume shows how many times a query is searched. Intent shows why. These are two different dimensions, and they multiply: if you pick a query with 10,000 impressions but miss the intent, you get 0 traffic. If you pick a query with 200 impressions and hit it, you get 50–100 visits with a 5–10% conversion rate.

I often see beginners grab "fat" queries with 50–100 thousand impressions and try to build pages for them. Without checking intent, that's pointless. Better to take 30 mid-volume queries (500–3,000 impressions each) that are definitely yours than one fat query with an unclear nature.

Check your queries right now

We have a free tool that analyzes intent for a list of queries in a minute. Paste up to 20 phrases into /free/intent — Claude will sort each one by type and explain why.

This is the basic step in any work with a core: analyze the intent first, then group into clusters, then design pages. Otherwise you'll build a page that's "everything for everyone," which lands with no one.

In the next article we'll figure out what search volume is and how to read it correctly in Wordstat.