How to choose a domain name: the seven checks to run before you buy

How to choose a domain name is really a question about order. Everybody knows the criteria: short, spellable, memorable, ideally a .com. What separates a decision that takes an afternoon from one that takes three weeks is which check you run first, because each one has a different cost and a different kill rate.

Below are seven checks in the order that fails candidates fastest and cheapest. Run them as a funnel, not a scorecard. The point is to stop thinking about a name the moment it fails, rather than to grade every candidate on all seven.

Check one: can it be said once and typed correctly

Free, instant, and it removes more names than anything else on this list. Say the candidate out loud to somebody who has not seen it written down, then ask them to type it. Anything that comes back with a different spelling is out.

The usual casualties are creative vowel swaps, doubled letters, words with two accepted spellings, and compounds whose seam is ambiguous. Every one of those is a small permanent cost on every conversation about the business, paid forever, in exchange for looking slightly more available today.

Check two: is the .com actually free right now

This has to come second, and running it late is the single most expensive mistake in naming. Not third, not last, second: immediately after the candidate proves it can be spoken.

The reason is emotional rather than technical. A name you never seriously considered costs nothing to lose. A name you have lived with for two weeks, said out loud in three meetings and started picturing on a sign is a name you will talk yourself into buying badly, or into taking on a worse extension, or into fighting a holder for at four figures. Checking early means you never form the attachment.

Do it against the registry rather than against a search box that also wants to sell you something. Registrar search is fast and accurate on availability, but when a name is gone it is designed to move you sideways onto whatever it has in stock. The verdict you want is a plain yes or no from the record itself. That is what the domain availability checker returns, and what the console prints beside every candidate it generates.

Check three: does it survive a trademark search

Cheap to run, catastrophic to skip. Search the national trademark register for your country, and the register for any market you plan to sell into, for the word itself and for anything phonetically close in your class of goods.

This is a screening step and not legal advice. What it reliably catches is the obvious collision: an existing registered mark in the same category with the same or nearly the same name. That is the case that ends in a rebrand after you have printed things, and it is worth twenty minutes to rule out. Anything ambiguous is a question for a lawyer, and the fee is small compared to the alternative.

Check four: what does it mean somewhere else

Two minutes of searching the candidate as a plain word, in the languages of any market you might reasonably enter. You are looking for an existing meaning, an unfortunate one, or an established company that already owns the association.

Invented words are not exempt. A coined name assembled from Latin roots very often turns out to be an ordinary noun in a language you do not speak, and the ones that go wrong tend to go wrong memorably.

Check five: does it still fit in five years

Ask what the name commits you to. A name that names your first product, your first city or your first channel is the clearest possible name on day one and a liability the moment you add the second of anything.

This is the trade the descriptive names lose. They buy immediate clarity with future room, and the bill arrives exactly when things are going well. Slightly abstract names cost a little explaining at the start and stop costing anything after that.

Check six: the handles and the shelf around it

Check the social handles you actually use, and check whether the obvious near misses are taken by somebody with a reputation. You do not need an exact match on every platform, and chasing one will cost you better names. What you do need is to know in advance whether the account you want requires a prefix, and whether a similarly named business already occupies the search results you were expecting to own.

Take this as information rather than a veto. It is the one check on this list that is allowed to lose to a genuinely better name.

Check seven: write the sentence you will say a thousand times

Put the name in the sentence you will use to introduce the business, out loud, in full. "We're building X, it's a Y for Z." The candidate that makes that sentence shorter and easier to say is usually the right answer, and by this point you rarely have more than three left, all of which have passed everything above.

If the three are still tied, look at what each one is worth. The inputs that make a name valuable, length, pronounceability and extension, are the same inputs that make it easy to remember, so a candidate that estimates into a different class is telling you something real. Domain name value covers how to read that figure, and what it cannot tell you.

Register the same hour you decide

The window between deciding and registering is the only period in which somebody else can take the name, and it is a window you control. Registration is a small yearly fee at any registrar and it is the cheapest insurance in the entire process. Announce afterwards.

Keep the losing candidates and one line on why each one lost. Somebody will reopen the question in six months, and a written answer is faster than having the argument again. If you want the criteria discussed at more length rather than as a funnel, how to choose a domain name for a business takes them one at a time, and business name ideas covers the patterns that still have free .com names behind them.