750 million domain names × what people search

How crowded is your business idea?

Every domain name is a small bet someone paid for. We read about 750 million of them and set them against what people actually search for. Type an idea and see where sellers crowd, where buyers are left waiting, and which niches next door you have not thought of yet.

Try mobile dog grooming meal planning app home brewing candle making

Where sellers and buyers meet

A domain name is a founder’s bet. A search is a customer’s need.

Put millions of each side by side and you can see both minds at once, and the places where they collide.

Why it matters

Keyword tools show you what people type; the map also shows who already bet on it. In a few minutes a vague idea turns into three or four sharp hypotheses, and a list of traps to skip.

How it works

We take each registered name, split it into words and slide a two-word window along them. Linguists call these pairs bigrams. Simply counting how often each bigram occurs in a text is enough to read it statistically: which ideas recur, which travel together, which are rare.

mobiledoggrooming.com→mobile dogdog grooming→searches · ad price

Google counted bigrams across millions of books for its Ngram Viewer, to watch words rise and fall over centuries. We count them across domain names, read as one text written by millions of founders: what they hoped to sell, and which pains they bet customers would pay to solve. Then we look up each pair as a search phrase: how often people type it into Google, and what advertisers pay to show up for it.

Those searches are not for the websites. They are people looking for the thing itself: a groomer for their dog. So without opening a single site, you can tell which field a name was built to play in, and how much demand that field has.

How to read the counts

Each count is a small paid bet

The number of domain names, out of about 750 million registered since 2006, that contain a pair of words. Somebody looked at that corner of the market and decided it was worth a name. Other spellings and word orders ("pressure wash", "wash pressure") are folded into one row and listed under "also".

10,000+ names

Crowd

Many bets already, plenty with serious budgets. A small product arriving alone is easy to miss.

300 – 10,000 names

Middle

Steady, paid interest without a stampede. A narrow job here is often not owned by anyone yet.

Under 300 names

Tail

Small and specific: cheap ground for a first test, and sometimes just a pair nobody needs.

Searches, “per name” and advertiser cost — what the labels mean

The counts show supply: how many people put a name on a field. The search labels show the other side: how often people in the US search Google for the pair itself as a phrase, in five levels from very low to very high; we recalculate them several times a year. "Per name" puts the two together: room to spare means many searches for each registered name, crowded for its demand means many names chasing few searches. "Advertisers pay" (little, average, a lot) shows what advertisers pay per click on the pair: a hint of its commercial value.

Customer searches are not founder searches: people looking for a coffee shop want coffee, not a business plan. Pairs without a label have no estimate, which is not the same as no demand.

How to use it

From a vague idea to a cheap test in five steps

"Get prompt" copies the results with explanations, so your own AI chat can help with steps 2 to 4.

  1. 01

    Start from a field you know a little

    A hobby, your trade, a problem you have had yourself. If you type a problem as a verb, like reduce anxiety, there may be no exact pair: names are usually built from nouns. Look at the nearby pairs instead, such as "anxiety relief" or "anxiety treatment".

  2. 02

    Look for the gaps

    Sort by searches per name and look for tail or lower-middle pairs with high or very high searches and "room to spare". Switch to Broader to see the neighbours outside your words.

  3. 03

    Note the pairs to avoid

    Crowded heads, and pairs marked "crowded for its demand", where many names chase few searches.

  4. 04

    Check the search results by hand

    Search a few narrow phrases around each pair in a signed-out browser. A pair with room to spare can still be owned by one strong site, a marketplace or a video platform.

  5. 05

    Test cheaply

    Build one small free tool or page for the best pair, and keep it only if strangers find it and use it.

Limits

What the map can’t tell you

Odd cuts and spellings

Names are split into words by a program, so some pairs are cut in odd places ("block chain"), and one business can hide behind a dozen spellings.

Pure brands leave no trace

A company called Oracle leaves no trace under "database". Most small businesses name themselves the hybrid way: in canopusdentalclinic.com the map keeps "dental clinic" (20,018 names), while the rare "canopus dental" never appears: a pair needs at least 100 names. Big brands show up as the trades around them: "iphone repair" (4,497), "google ads" (2,184).

Bets, not winners

Counts mix today’s activity with old fashions, and many names have since expired. A count shows where people bet, not who won.

No rankings, no feasibility

It cannot tell you who ranks on the first page today, or whether you can build the product and sell it.

Read it the way you would read a crowd from a window: look at clusters of related pairs, ask who registered those names and what they were trying to sell, and compare it with what you know. The direction for your business is still your call.