Naming a brand when every good name feels taken

The complaint arrives in every naming project now, usually in week one: everything is taken. The trademark office receives hundreds of thousands of applications a year, Verisign's quarterly count of registered dot com names has sat far past one hundred and fifty million for years, and the short dictionary words went decades ago. Founders respond with the era's workarounds: a dropped vowel, a doubled letter, a dot ai suffix doing the work the name gave up on.

Look closer at what is actually exhausted. The crowded shelf holds descriptive names: the clear, sensible compounds that say what the business does. Swift Logistics, Bright Analytics, TrueNorth Advisory, in every spelling arithmetic allows. Scarcity lives precisely where every founder shops.

The shelf of distinctive names is close to infinite. English alone offers tens of thousands of usable words with no category attachment, and coined words multiply that without limit. Nobody had claimed Kodak until someone invented it. The wall founders hit is a preference, and the preference is the actual problem this essay works on.

A brand name has one commercial job: retrieval. When the buying situation arrives, some name surfaces in the customer's memory, and the business carrying that name gets the enquiry. Everything a strategy hopes for funnels through that half second.

Description feels safer because it seems to shorten the explanation. It rarely does. A descriptive name explains the category, and the category is the one thing the buyer already knew. What the buyer lacks is a handle for you specifically, and a name built from category vocabulary hands them a handle shaped exactly like your competitors' handles.

The division of labour that works: the name is the hook, the sentence after the name does the describing. Apple carries zero information about computers, which left it free to hold forty years of accumulated meaning. The description job belongs to positioning lines, category entries, and the written record, all of which can change. The name is the one word that must never need to.

Psycholinguistics has measured for decades what poets and novelists always practised: sounds carry meaning on their own. In the classic demonstration, people across languages match the invented word kiki to a spiky shape and bouba to a rounded one, agreement running near total. Klink's brand name experiments made it commercial: front vowels, the ee and ih sounds, read smaller, faster, sharper, and more feminine; back vowels, the oh and oo sounds, read larger, heavier, and slower. Plosive consonants snap; fricatives glide.

Lowrey and Shrum then showed the effect steers preference: buyers like a name more when its sound fits the product's promise, a crisp front vowel for the nimble app, a deep back vowel for the heavyweight machine. Literature got there first. Dickens built character in sound before biography: Gradgrind grinds, Scrooge scrapes, and a reader trusts the ear's verdict on page one.

Fluency completes the picture. Alter and Oppenheimer found that companies with easily pronounced tickers outperformed hard ones in early trading, purely on processing ease. The practical screen for a shortlist: say every candidate aloud, score what the sound promises against what the strategy promises, and keep the candidates where ear and strategy agree. A name whose sound argues with its positioning fights that argument in every single exposure.

The 2026 twist on the shortage is partly manufactured. Ask a language model for name ideas and it predicts the most probable answer, which means the statistical centre of every startup name it has read: soft two syllable coinages, familiar suffix endings, the vaguely Latin glow. Thousands of founders prompt the same tools with similar briefs, then meet each other at the trademark office holding near identical inventions.

This is the modal content trap wearing a name tag. Generators are genuinely useful as pressure tests and shortlist expanders, and genuinely dangerous as deciders, because their default output is the category's average, and average is the one quality a name exists to escape.

The countermove is the same one that governs AI content everywhere: feed the machine your specifics and reserve the judgement. Give it your positioning, your sound direction, your banned suffixes, and let it generate against those constraints. Then run the audition yourself. A rival's tool could produce the candidate list; only your strategy can produce the choice.

Trademark law quietly agrees with memory science. Marks are graded on a spectrum of strength, and descriptive names sit at the weak end: hard to register, expensive to defend, forever sharing the register with lookalikes. Coined and arbitrary names sit at the strong end, where protection is broad and conflicts are rare. The name that is easiest to defend is the same name that is easiest to remember, and both for the same reason: nobody else was standing there.

The exact match dot com deserves demotion from requirement to preference. Buyers arrive through search, social profiles, and increasingly through assistants that resolve a spoken name to an entity, and every one of those routes runs on recall rather than typing. A distinctive name with a modified domain outperforms a compromised name with a perfect one, because the name is spoken, remembered, and asked for far more often than it is typed.

The order of operations that saves money: audition first, then screen the survivors through a trademark search in your classes and markets, then accept whatever sensible domain exists. Founders who run it backwards let a parking page veto their strongest candidate, which hands naming authority to whoever squatted fastest.

Every name is born empty. Google was a misspelt maths term, Kodak was invented for the sound of the shutter, and both acquired their meaning the only way meaning is acquired: deposits. Repetition, consistency, and years. Choosing a name is opening an account; the balance arrives later, and only if the deposits keep landing in one place.

This is where naming projects actually fail. The distinctive candidate feels exposed on day one precisely because it is empty, the descriptive candidate feels safe because it borrows the category's existing meaning, and committees walk toward the borrowed comfort. My own practice carries the choice in its name: tatva is the Sanskrit word for essence, the element beneath a thing, and it explained zero services on day one. The name held still, the work filled it, and the filling is the point.

So the last screen in the audition is a commitment, written down: one spelling, one pronunciation, one visual setting, repeated everywhere the business speaks, for years. A distinctive name with that discipline compounds. Without it, even a brilliant name stays an empty account with a lovely sound.