We Tested 200 Domains. Not a Single Real Word Was Free.

What exactly did we test?

We wanted a fair, boring, reproducible experiment — not a cherry-picked scare story. So we built two lists of 50 words each and treated them identically.

Bucket A was 50 real, single English words: everyday nouns and adjectives with genuine brand appeal — otter, meadow, lantern, ember, quartz, zephyr, willow, cinder, and so on. These are exactly the words people wish they could own.

Bucket B was 50 coined words: invented or blended strings that don't appear in a dictionary but read like modern startup names — zenlio, brixon, glyntic, ploomex, quivex, fintavo, and the like.

Every word was checked on two extensions, .com and .io, giving us 200 total availability checks. We ran them live through the same DNS-plus-registry pipeline our search tool uses, so a 'free' result means the domain genuinely resolved to nothing at the moment we looked. No estimates, no scraping old lists — a live snapshot taken in one sitting.

How many real words were still available?

Zero. Out of 100 real-word domains — 50 on .com and 50 on .io — exactly none were free.

That is not a rounding artifact. It's a clean 0%. Every single common English word we tried was already registered, on both extensions. The .io results are the part that surprised even us: people treat .io as the 'escape hatch' when the .com is gone, but for plain dictionary words that escape hatch is already bolted shut. If the word exists in English, someone has parked, developed, or is squatting on both the .com and the .io.

The practical takeaway is blunt: if your naming strategy is 'find a nice real word that's somehow still open,' your success rate in our sample was literally 0 in 100. You are not unlucky. The inventory simply doesn't exist anymore.

Did made-up names do any better?

Dramatically better. Of the 100 coined-word domains we tested, 39 were available — a 39% hit rate versus 0% for real words. But the average hides the most useful part, which is how the two extensions split:

On .com, 5 of the 50 coined names were free — a 10% availability rate. On .io, 34 of the 50 were free — a 68% availability rate. Same invented words, same session, and .io came back available almost seven times as often as .com.

To put the whole experiment in one table: real words scored 0% on .com and 0% on .io; coined words scored 10% on .com and 68% on .io. The single biggest lever on whether a domain is open isn't the extension and it isn't luck — it's whether the word already exists in English.

For the curious: five coined .coms were unregistered at check time, including glyntic.com and ploomex.com. We're not telling you to buy them — domain availability changes by the minute and these may already be gone — but they show the pattern. A slightly odd, invented string is the difference between a locked door and an open one.

Why is .io the secret weapon for coined names?

The gap between 10% (.com) and 68% (.io) on the exact same words comes down to age and volume. The .com registry has been accepting registrations since 1985 and holds over 150 million active domains; four decades of founders, investors, and squatters have swept through it. Even weird invented strings on .com have often been grabbed speculatively.

.io is younger, smaller, and — until recently — mostly of interest to a technical crowd. That means the coined-name shelf is far less picked-over. Our data is a live demonstration of it: the same made-up word that's taken on .com is, more often than not, sitting open on .io.

This isn't a recommendation to reflexively grab .io for everything — a consumer brand chasing mainstream trust may still want to fight for a .com. But if you're building a developer tool, an API, or an infrastructure company, the numbers say .io gives you a genuinely bigger pool of open, brandable names to choose from.

Why is every good real word already gone?

There are only so many real words. Common, brandable English vocabulary — the punchy, one-word kind — numbers in the low tens of thousands, not the millions. That finite pool met an effectively infinite demand decades ago.

Three forces finished the job. First, time: .com has had 40 years of continuous registration, and the best words went early. Second, investment: domain investors hold huge portfolios of dictionary words precisely because they're scarce and resell for five and six figures. Third, defensive registration: established companies buy the .io, .co, and .net of their .com to protect the brand, quietly removing those words from circulation on every extension at once.

So when your dream one-word .com is taken, it usually isn't in active use — it's parked, listed for sale, or held in a portfolio. That's cold comfort, but it reframes the problem correctly: you're not competing for a word that's about to free up. You're competing against a market that decided long ago these words are assets.

What should you actually do with this?

Stop spending your first week of naming hunting for an unclaimed real word. Our data says that search returns nothing. Spend that energy on the strategy that actually has open inventory: coin a name.

The best modern brand names are invented — Stripe, Vercel, Twilio, Figma, Notion (a real word that was an expensive exception). Invented names are available more often, easier to trademark, and easier to rank for in search because you're not fighting a generic dictionary term. Our 39% coined-name availability rate versus 0% for real words is the availability side of that same argument.

A practical loop that works: generate a batch of pronounceable coined names, check them live across .com and .io, and keep the ones that are open on the extension you want. That's exactly what ZeroTaken is built to do — describe your idea, get brandable invented names, and see real-time availability instead of clicking through a registrar one guess at a time.

How honest is this data — and what are the caveats?

This is a 200-domain snapshot, not a census of the entire DNS. A different set of 50 real words would still, in our experience, come back overwhelmingly taken — but your exact percentages will vary with the words you pick and the day you check.

Availability is also a moving target. Every number here reflects the moment we ran the check; a name that was free this morning can be registered by lunch, which is exactly why a live check beats any static 'available names' list you find lying around the internet.

And 'available' is not the same as 'good.' A coined name has to be pronounceable, spellable when said aloud, and free of unfortunate meanings in other languages before it's worth buying. Availability just gets you to the starting line. The honest headline stands, though: in a fair test, real words were 0-for-100, and invented names were the only ones with an open door.