How to name a company: a process that survives contact with reality

Naming projects tend to die in the same spot. Three weeks of falling in love with a word, and then the discovery: the .com went for five figures in 2014, a competitor in an adjacent sector already uses something almost identical, and nobody can spell it when you say it down the phone. More creativity won't save that. Doing the elimination earlier will, because changing your mind is cheap right up until it isn't.

What follows holds together once money and lawyers arrive.

Write the constraints before anyone says a word out loud

Not values. Constraints — the checkable kind, on one page, before the brainstorm.

• Where it gets said. If growth comes from word of mouth, sales calls or podcasts, the name has to be spellable from a single hearing. If it lives in search results and app stores, spelling matters less than being distinct.
• How far the category stretches. A name that describes your first product becomes a tax on the second. If software might turn into services, or one country into five, don't pin yourself.
• Who else is in the space. List the ten names your buyers already know. You are mapping what everyone sounds like in order to avoid it. Four competitors ending in -ly and two starting with Get means those doors are shut.
• Budget for the name asset. Set the ceiling now, while you are still capable of walking away. A serious range for a good one-word or two-word .com is four to six figures; brandable coined .coms often sit in the low-to-mid four figures. Under a thousand dollars, you are inventing words, not shopping for dictionary terms.

Every argument in week two gets settled by pointing at that page.

Quantity comes first, and it isn't inspiration

Get to 100 or 200 raw candidates before you evaluate a single one. The decent names show up after the obvious ones have left your system.

Work the types deliberately instead of waiting to be struck:

• Descriptive: says what you do. Instantly understood, almost impossible to own, weak as a trademark. Fine for a local plumber, risky for anything with ambition.
• Suggestive: gestures at a benefit without spelling it out. Most durable brands live here.
• Coined: invented, or a real word bent out of shape. Cheapest .com, strongest legal position, and two years of explaining yourself.
• Arbitrary: a real word with nothing to do with the business. Distinctive and defensible. Also a common word, and common words cost.
• Metaphor and reference: mythology, geography, trades, materials. Rich seam, heavily mined. The obvious picks went long ago.
• Founder or place: credible in professional services. Limiting nearly everywhere else.

Do a pass on each. Say every result aloud as you go, because a name that's awkward in the mouth stays awkward. To widen the coined and suggestive columns by machine, the keyword domain search approach starts from words that already matter to your business and shows what can be built around them.

Cut in order of what the check costs you

Free filters first. The lawyer last.

The say-it test

Read each name down a phone line to someone who has never seen it written. If they can't spell it back, it's gone. This usually takes out a third of the list, and it's almost always the cleverest third.

The search test

Search every survivor. You are hunting for an existing company of the same name in a nearby category, an unfortunate meaning in another language, or a first page of results you could never shift. Sharing a name with a mid-sized firm in an adjacent sector buys you a lifetime of confusion.

The domain test

Now find out what the .com actually costs. Not whether it's "available" at registration price, because for anything good the answer is no. What would it take to own it? Some names carry a public asking price. Some sit with owners who'll take a call. Some are attached to a live business and will never move.

Shortlists collapse here, which is precisely why you do it before the logo exists. Working backwards from what's genuinely for sale is often faster than working forwards from an invented word: a category browse puts real names at real prices in front of you, and anything you find there is acquirable by definition.

The trademark test

With three to five names left, run knockout searches in the registers covering the markets you'll operate in, in your goods and services classes. Free databases catch the obvious collisions. Then pay a solicitor or trademark attorney for a clearance opinion on the one you intend to build on. A few hundred to a couple of thousand, and the best-value money in the whole exercise. A forced rebrand in year three costs a different order of magnitude.

What you're really trading away without the .com

Plainly: you don't need the exact-match .com to build a company. Businesses run perfectly well on .io, .co and modified .coms. But know the bill.

Without it, you pay for the confusion forever. Typo-ed traffic landing on a stranger's site. Email addresses spelled out letter by letter. The customer who says "I couldn't find you". And if you ever try to buy the .com later, you're negotiating from the weak side, because the holder knows you need it.

The rough shape of the market: hyphens, misspellings and long compounds are cheap. Two-word brandable .coms that pronounce cleanly start in the low four figures and climb. Short, common, single-word .coms are five and six figures. Anything under ten thousand for the name your entire company will carry isn't an extravagance. It's a line item.

Two things about buying. Use escrow, so neither party has to send money or push the domain on trust; on Names.com transfers run through escrow by default, sellers pay a 15% commission only on a completed sale, and buyers pay nothing. And if the name is held privately, an owner-direct listing means you're talking to the person who can actually sign it over, rather than a chain of brokers.

Three tests, then stop deliberating

Don't put the finalists to a public vote. Crowds choose the safest option and safe names evaporate on contact. Run these instead:

• The sentence test. Write the line a customer would use: "You should try ___, they do ___." A name that makes that sentence clumsy will make every sentence clumsy.
• The stretch test. Put it on your third product, in your second market. Does it still fit, or does it start arguing with you?
• The rival test. Set it in a list beside your competitors. Standing out, blending in, or sounding like a copy of the market leader?

Then choose, fast. Names improve through use. A merely good name repeated for two years beats a perfect one still under discussion, because the company makes the name mean something and never the reverse.

Set the deadline before you begin. Two weeks is ample for a small team. Longer than that and you've stopped naming; you're avoiding launch.