Coined names versus real words: choosing between them

Real-word domains buy you instant comprehension and cost more money. Coined names cost less to acquire and more to explain. Which trade you make depends on one number: how much you can spend on being understood over the next three years, through advertising, sales calls, packaging or search.

That's the decision. Everything after this is detail on making it well.

What you get in each direction

A real-word domain uses existing vocabulary. Monzo isn't one; Mint is. It might be a single dictionary word, a descriptive phrase, or a compound of two real words such as Salesforce or Mailchimp: half-coined but readable.

The benefits:

• No spelling friction. Someone hears it on a podcast and types it correctly. Founders underrate this constantly, because most brand discovery still happens through ears and memory rather than clicks.
• Borrowed meaning. A word like Anchor or Compass turns up with associations already attached. You didn't pay to build them.
• Category signal. If the word touches your sector, a prospect knows what you do before you open your mouth.

A coined name is invented. Zapier, Vercel, Klarna, Xero. Sometimes a Latin or Greek root reshaped, sometimes pure phonetics with no ancestry at all.

The benefits:

• Ownership. Nobody else is using it. The trademark route is cleaner and the search results page belongs to you.
• Availability. The .com is far more likely to be registerable, or cheap.
• Room to move. An invented name doesn't lock you into a category you may want to leave.

The price gap is large and there is no point pretending otherwise

Be honest about the numbers before you fall in love with anything. A clean, short, single-word .com in a commercially active category — finance, health, logistics, software — typically sits in the five- to six-figure range on the aftermarket, and the best of them sit far above that. A two-word real phrase that reads naturally might land anywhere from the low four figures to the low five. A pronounceable invented .com of five to eight letters can often be registered outright, or bought for a few hundred to a few thousand pounds if somebody got there first.

So the question was never which type is better. It's whether the comprehension a real word buys is worth the difference, given how you actually acquire customers.

Two situations where it plainly is:

• Consumer sales with a short consideration window. Impulse and low-consideration purchases live or die on immediate understanding. A self-explaining name removes a step from the funnel.
• Word of mouth and offline channels. Radio, podcasts, van livery, event sponsorship, shelves in shops. If the name has to survive being said aloud once, spelling is everything.

Two where it plainly isn't:

• A sales team and a long funnel. When a human spends forty minutes explaining your product, the name isn't doing the explaining. Spend the money on the product.
• You expect to broaden. A name welded to your first product becomes an obstacle at your second.

Where invented names quietly win: lawyers and search results

This is the part nobody weighs properly at the naming stage, and it arrives later as legal invoices.

Trademark law protects distinctive marks. Descriptive terms are hard or impossible to register. Generic terms are impossible, full stop. Call a coffee subscription FreshBeans and you will struggle to stop a competitor launching something adjacent. Call it Kelva and you have a strong mark from day one, plus a genuine ability to defend it.

Search works on similar logic. Put a common word into a search engine and you're up against dictionaries, Wikipedia, unrelated businesses and twenty years of accumulated content. Put an invented name in and you own page one within weeks of launching. Branded search means people typing your name specifically. It's the highest-intent traffic that exists, and coined names collect all of it.

The counterweight is real. Common words attract non-branded search intent that invented names never will. If someone searches your category and your domain contains those words, you pick up a small relevance signal and a click-through advantage in the listings. Modest, not decisive, nothing like the edge it was fifteen years ago. But it exists. If that's the angle you want, start with a keyword-led search rather than browsing at random.

Most good names are hybrids anyway

The binary is convenient and slightly dishonest. Strong brand names usually sit between the poles.

• Real word, wrong category. Apple for computers. Oracle for databases. You get the spelling ease of real vocabulary and the distinctiveness of a coinage, because nobody expects that word in that context. Often the best deal available, and cheaper than in-category words because fewer buyers are chasing them.
• Two real words joined. Facebook, Snowflake, Stripe-adjacent constructions. Spellable, distinctive enough to trademark, and vastly more available than single words.
• Modified real word. Lyft, Flickr, Tumblr. Cheap, available, and taxed forever: every time you say it aloud you also say "spelled with a Y". Price that in.
• Root-derived coinage. Verily, Vercel, Auralis. Invented, but it feels like a word and carries a faint suggestion of meaning through the Latin.

Early and budget-bound? The second and fourth options hand you most of the upside for a fraction of the outlay. A curated set aimed at early-stage founders will show you what that price band actually holds, which beats theorising about it from a spreadsheet.

Six tests worth an afternoon

Run every shortlisted name through these. An afternoon here prevents expensive mistakes later.

• The phone test. Say it once to someone unfamiliar. Ask them to write it down. If they get it wrong, you will pay for that error thousands of times over.
• The search test. Search the name today. What's competing for it? A common word with a strong incumbent on page one is a long, costly climb.
• The trademark test. Check the relevant registry for your classes and territories before you spend anything. A cheap domain attached to a blocked mark is not cheap.
• The five-year test. Write down what you might be selling in five years. Does the name still fit? Category words fail here more than anything else.
• The exact-match test. Does the .com exist, and can you get it? Building on a name whose .com belongs to someone else means leaking traffic permanently and negotiating from weakness permanently.
• The other-language test. Selling abroad? Check the name isn't unfortunate elsewhere. Coined names fail this more often than founders expect, because invented syllables keep landing in real vocabularies.

Deciding, in one pass

Choose a real word if customers must understand you in under three seconds, you sell through physical or audio channels, you have a category you intend to stay in, and four to six figures won't hurt.

Choose coined if a sales motion or content strategy does the explaining, you want a defensible trademark, you might broaden the product, or the price difference would eat into runway that matters.

Choose a compound or an off-category real word if you want most of both and can't fund a pure single word. That describes most businesses assessing themselves honestly.

A process note for the coined route: verify availability against the live registry instead of trusting a generated list. A brilliant invented name that's already taken is a wasted afternoon. The Name Studio checks everything it invents against the registry, so what appears on screen can actually be registered. And in either direction, take the .com if it's within reach. People relitigate the naming argument for years. Nobody reopens the extension.