You open Yandex Wordstat, type "buy a cake," and it shows 142,000. That means that 142,000 times a month people on Yandex search for something related to buying a cake.
It sounds as if, once you make a page about cakes, 142,000 people will come to you. But that's not how it works. If everyone who searches "buy a cake" went to a single store, every confectioner would be a millionaire.
Let's figure out what this number actually means.
What search volume is
A query's search volume is the number of impressions for a query in search over a given period (usually a month). Not people, but impressions specifically. If one person searched "buy a cake" three times in a month, they show up in the statistics three times.
Wordstat counts every variant with those words: "buy a cake," "buy a birthday cake," "where to buy a cake moscow," "buy a cake for a child," and so on — all of it is one "fat" number.
When an SEO looks at 142,000, they're looking at a family of queries, not a single exact one.
The three types of volume in Wordstat
This is the most confusing spot for beginners, but you can't work with a core without it.
Broad ("no quotes")
You just type buy a cake into Wordstat. You get the widest number — it includes all
word forms and all variations with those words in any order.
buy a cake in broad form yields: "buy a cake moscow," "buy a little cake as a gift," "buy a big cake,"
"i'll buy cakes for a wedding," "cake buy cheap," and hundreds of others. The sum of impressions for all these phrases.
This number is the largest and the most useless for estimating real demand. It's good as a "first rough take" — to find out whether the topic exists at all.
Exact ("in quotes")
You type "buy a cake" (in double quotes). Wordstat shows only the queries where
exactly these words appear, in any order and word form. So "i'll buy a cake," "cake bought,"
"buy cakes" all count. But "buy a cake moscow" does not (it has an extra word).
This number is already much smaller than the broad one. For "buy a cake" in quotes, it's usually 8,000–15,000 a month, instead of 142,000 broad.
Precise ("!buy !cake")
Quotes plus exclamation marks before each word. This is the strictest — only the exact
word form in the given order. So "i'll buy a cake," "buy a little cake" no longer count.
Really only buy a cake and buy a cake. You'll get a small number — 500–2,000 impressions.
This number is the most honest. It's "how many people search for exactly this phrase."
Which to use
| Task | Type of volume |
|---|---|
| Estimate whether the topic exists at all | broad (no quotes) |
| Roughly how many people search in this branch | exact (quotes) |
| Precise estimate of demand for a specific phrase | precise (quotes + ! ) |
| Picking keywords for the core | look at everything |
| Deciding "build a page or not" | precise |
From impressions to real traffic
Let's say we took the precise volume of the phrase "custom cake moscow" — 2,000 impressions a month. How many of them will come to our site if we're in the top?
The chain looks like this:
Impression → Click in top 10 → Click on our position specifically → Landed on the site
CTR (click-through rate) — the share of clicks out of impressions:
| Position in top 10 | Share of clicks out of impressions (CTR) |
|---|---|
| Position 1 | ~28% |
| Position 2 | ~14% |
| Position 3 | ~10% |
| 4–5 | ~5–7% |
| 6–10 | ~2–3% each |
| 11+ (second page) | under 1% |
Half of impressions go to "the-organic-that-gets-no-clicks" — people refine the query, close the tab, click an ad. Take this table with a grain of salt — real CTR depends on what's in the results (product cards, video, maps, images), on the niche, and on how appealing your heading is.
Take our example with "custom cake moscow" (2,000 precise impressions). If you're at position 3: 2,000 × 10% = 200 visits a month. If at position 7: 2,000 × 3% = 60 visits.
200 visits from a single phrase is already decent. Conversion into an inquiry in the cake niche is usually 3–7%, which means 6–14 leads a month from one query in the top. From there, factor in the average order value.
Why queries with 50,000+ impressions are often not what they seem
The fattest queries look tempting — there's so much traffic there. But there are three traps.
Trap 1: fat queries are brands you have no shot at. "Buy a laptop" — 250,000 impressions. Top 10 — DNS, Citilink, M.Video, Svyaznoy. Competing with them is impossible: they pay more for SEO than your business is worth.
Trap 2: fat queries have mixed intent. "iphone 15" — 90,000 impressions. Someone wants to buy, someone to read a review, someone to check the specs. If you build a page for purchasing, you'll only appear in the results for a third of these people. See more: search intent.
Trap 3: fat queries have high volume from bots and manipulation. Especially in B2B niches and technical topics. When you see a strangely "fat" number for a narrow topic, check the precise volume — it's usually 10–20 times smaller.
The right approach: mid-volume queries
Real money is made on queries with 500–5,000 impressions a month (by exact volume). Their characteristics:
- Less competition — big sites don't spend SEO resources on every such small item.
- More specific intent — the phrasing is longer, the person knows exactly what they want.
- Higher conversion — a target customer, not "a person off the street."
- Easier to reach the top 10 — a good page structure and a couple of quality links are enough.
The "50 mid-volume queries in the top" strategy is far more effective than "1 fat query at position 8–10."
Seasonality: the same number in different months means different things
Wordstat shows the monthly average, but real traffic can swing by 10–20 times depending on the season. "New Year gifts" — 200 impressions in July and 80,000 in December. The average of 25,000 is misleading.
Wordstat can show query history — the "Query history" button under the search field. It's a chart of monthly peaks. Always check seasonality for any seasonal product (clothing, sports, holidays, education).
We'll cover seasonality in more detail in a separate article, but for now the rule is: a flat chart — you work year-round; a wavy chart — plan content 2 months before the peak.
Frequently asked questions
Which matters most: broad, exact, or precise volume?
Why does a query with 50,000 impressions give my site 100 clicks a month?
Should I chase the fattest queries?
Next article: the long tail — why 200 queries at 20 impressions each are stronger than 1 at 4,000.