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

Most naming processes fail in the same place. A team spends three weeks falling in love with a word, then discovers the .com sold for five figures in 2014, a competitor already uses a near-identical mark, and nobody can spell it over the phone. The fix is not more creativity. It is doing the elimination work earlier, while it is still cheap to change your mind.

Here is a sequence that holds up when money and lawyers get involved.

Start with constraints, not adjectives

Before anyone brainstorms, write down what the name has to survive. Not values — constraints. Concrete, checkable things.

• Where it gets said. If your growth channel is word of mouth, sales calls or podcasts, the name has to be spellable from hearing it once. If it lives entirely in search results and app stores, spelling matters less than distinctiveness.
• How far the category stretches. A name that describes your first product becomes a tax on the second. If you might sell software and services, or move from one country to five, avoid a name that pins you.
• Who else is in the space. List the ten names buyers already know. You are looking for what everyone sounds like so you can avoid it. If four competitors end in -ly and two start with Get, those doors are closed.
• Budget for the name asset. Decide the ceiling now. Not later, when you are emotionally committed. 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. If your ceiling is under a thousand dollars, you should be generating invented words, not chasing dictionary terms.

Write these on one page. Every later argument gets settled by pointing at it.

Generate wide, in categories, without judging

Quantity first. You want 100 to 200 raw candidates before you evaluate anything, because the good ones tend to arrive after the obvious ones are out of your system.

Work through name types deliberately rather than waiting for inspiration:

• Descriptive: says what you do. Easy to understand, hard to own, weak in trademark terms. Fine for local service businesses, risky for anything that wants to scale.
• Suggestive: gestures at a benefit or a feeling without stating it. This is where most durable brands live.
• Coined: invented words, or real words bent into something new. Cheapest to acquire the .com, strongest legally, most expensive to explain in the first two years.
• Arbitrary: a real word with no connection to the business. Distinctive and defensible, but you are buying a common word, and common words cost.
• Metaphor and reference: mythology, geography, trades, materials. Rich, but heavily mined — most obvious picks are long gone.
• Founder or place: credible in professional services, limiting almost everywhere else.

Do a pass on each type. Say the results out loud as you go; a name that is awkward in the mouth will not improve later. If you want a machine to widen the coined and suggestive columns, the keyword domain search approach lets you start from words that actually matter to your business and see what is buildable around them.

Kill ruthlessly, in the cheapest order

Run your list through filters ordered by cost. Free checks first, lawyer last.

Pass one: the say-it test

Read each name down a phone line to someone who has not seen it written. If they cannot spell it back, cut it. This removes a third of most lists, usually the cleverest third.

Pass two: the search test

Search each survivor. You are looking for an existing company with the same name in a nearby category, an unfortunate meaning in another language, or a first page you could never displace. A name that is already the name of a mid-sized firm in an adjacent sector is a lifetime of confusion.

Pass three: the domain test

Now check what the .com costs. Not whether it is "available" at registration price — that answer is almost always no for anything good — but what it would take to own it. Some are on the market with a price. Some are held by owners who will listen. Some are attached to a live business and will never move.

This is where most shortlists collapse, which is why you do it before you have a logo. Browsing what is actually for sale is often faster than starting from a word you invented: a category browse shows you real names at real prices, and any name you find there is by definition acquirable.

Pass four: the trademark test

For the final three to five names, run knockout searches in the trademark registers where you will operate, in your goods and services classes. Free databases will catch the obvious collisions. For anything you intend to build a company on, pay a solicitor or trademark attorney for a clearance opinion. This costs a few hundred to a couple of thousand and is the single best-value spend in the whole process. A forced rebrand in year three costs vastly more.

Decide what the .com is worth to you

The honest position: you do not need the exact-match .com to build a company, and plenty of good businesses run on .io, .co or a modified .com. But you should know what you are trading away.

Without the .com, you will pay for the confusion permanently — in typo-ed traffic to a stranger's site, in email addresses that have to be spelled out, in the awkward moment when a customer says "I couldn't find you". You may also face a future acquisition negotiation from a weaker position, because whoever holds the .com knows you need it.

Rough shape of the market: hyphens, misspellings and long compounds are cheap. Two-word brandable .coms with clean pronunciation run into the low four figures and up. Short, common, single-word .coms are five and six figures. Anything under ten thousand for a name your whole company will carry is not an extravagance; it is a line item.

Two practical points on buying. First, use escrow — neither side should have to send money or push the domain first. On Names.com transfers run through escrow by default, sellers pay a 15% commission only when a sale completes, and buyers pay no fee. Second, if the name you want is held privately, an owner-direct listing means you are negotiating with the person who can actually transfer it, not a broker chain.

Test the shortlist against reality, then commit

Do not put the final names to a public vote. Crowds pick the safest option, and safe names are forgettable. Instead, run three specific tests:

• The sentence test. Write the line a customer would use to recommend you: "You should try ___, they do ___." If the name makes the sentence clumsy, it will make every sentence clumsy.
• The stretch test. Imagine the name on your third product and in your second market. Does it still fit, or does it argue with you?
• The rival test. Put it in a list beside your competitors. Does it stand out, blend in, or sound like a knock-off of the market leader?

Then decide, and decide quickly. Names improve with use. A merely good name that you commit to and repeat for two years beats a perfect name you are still debating. The company makes the name mean something, not the other way round.

Set a deadline before you start. Two weeks for a small team is plenty. Anything longer and you are no longer naming — you are avoiding launching.