How to Choose a Domain Name for a Business That Actually Works

Choose a domain name that is short, easy to spell after hearing it once, and easy to say out loud. Prefer the .com if you can get it. Favor a brandable name over a keyword-stuffed one, because exact-match keyword domains no longer carry the search advantage people think they do. Screen for trademark conflicts and social handle availability before you buy, not after.

Most naming advice stops at "make it memorable," which is useless. What follows is the actual working process: the criteria that matter, the ones that do not, the legal check nobody does until it is too late, and how to decide when the perfect name is taken.

What makes a good business domain name?

A good domain name survives being spoken out loud. If you can say it once to a stranger in a noisy room and they can type it correctly without asking you to repeat it, you have a good name. Everything else (length, cleverness, keywords) is secondary to that one test. Short and spellable beats descriptive and clever almost every time.

What is the radio test?

The radio test asks whether someone could hear your domain name on the radio, remember it for two minutes, and type it correctly. Say the name out loud to three people who have never seen it written. If any of them ask "how do you spell that?" the name has failed. It is the single cheapest naming filter and almost nobody runs it.

Names fail the radio test in predictable ways. Deliberate misspellings ("Kwik", "Lyft") work only if you have marketing money to teach people the spelling. Homophone traps ("SiteWrite" vs "SightRight") force a coin flip. Doubled letters at a word boundary ("MetalLathe") make people hesitate. Any of these can be overcome with budget, and none of them should be a problem you volunteer for.

Is .com still better than other extensions?

Yes, for most US businesses. The .com is still the default people type, the default they assume, and the one they will land on if they misremember your extension. If you build on a .io and someone owns the .com, a meaningful share of your traffic will end up at their door. That is the real cost, and it does not go away.

That said, alternative extensions are genuinely fine in specific cases. If your audience lives entirely inside a niche where the extension is native and understood, the penalty is small. Developer tools on .dev, AI products on .ai, and design studios on .studio all read as intentional rather than as a consolation prize. Country codes make sense when you serve one country. What does not work is picking an odd extension purely because it was cheap and available, and hoping customers will remember.

If you choose a non-.com, choose it because it says something. If you chose it because the .com was taken, you did not choose it, you settled, and your customers will feel the difference every time they type it.

Should I use keywords in my domain name?

No, not as an SEO strategy. Exact-match keyword domains (bestdentistchicago.com) lost their ranking advantage many years ago, and search engines now treat them as neutral at best. A keyword domain will not rank you. It can actively hurt you by making a real brand impossible to build and by reading as spam to the exact customers you want.

What keywords still do is set expectations. A name that hints at the category can lower the cost of explaining what you do, which is useful for a local service business where clarity beats brand. But there is a difference between a name that suggests a category and a name that is a search query with a .com on the end. The first is fine. The second boxes you in the moment you expand past that one service in that one city.

Brandable vs descriptive

Brandable names (invented or repurposed words with no literal meaning) are harder to launch and much better to own. They are ownable, trademarkable, and they scale when your business changes. Descriptive names get you understood on day one and then become a cage. Most companies that matter picked brandable, spent the early money teaching people what the word meant, and never had to rename when the product shifted.

If you are stuck for candidates, a domain name generator is a decent way to break out of the loop of your own five favorite words. Do not expect it to hand you the answer. Expect it to produce forty ideas, of which two are worth putting on a shortlist.

How do I check a domain name for trademark problems?

Search the USPTO trademark database before you buy, and search it for your category specifically, not just for an exact string match. Trademark rights are scoped to classes of goods and services, so "Apex" can legally exist as a plumbing company and a software company at the same time. A conflict only exists if someone holds the mark in a space close to yours.

Run the free searches yourself first: USPTO's search tool, a plain Google search of the name plus your industry, and a check of state business registries. If anything close comes back, pay a trademark attorney for an hour of their time before you spend anything on the domain. Registering a domain gives you zero trademark rights. Someone with a valid mark can take the name from you through a UDRP proceeding, and the money you spent on the domain, the logo, and the packaging goes with it.

Do I need matching social media handles?

Check them, and weight them, but do not let them veto an otherwise great domain. Consistent handles across your site and your channels make you findable and make you look like a real company. Perfect availability everywhere is rare, and a slight variation on one platform is a survivable problem. A trademark conflict is not.

The practical order: domain first, trademark second, handles third. Check the handles at the shortlist stage so you know what you are getting into, and accept a prefixed or suffixed variant if you have to. What you want to avoid is a name where every platform requires a different weird workaround, because customers will never keep track.

Should I register a new domain or buy one on the aftermarket?

Register a fresh name if you can find one that passes the criteria above, because it costs a low yearly fee and you own it in minutes. Buy on the aftermarket when the name you actually need is already taken and the difference in quality is worth real money. Short, clean, one-word .com names are essentially all registered already, so if that is your requirement, the aftermarket is your only route.

The honest math: a premium name can save you years of explaining who you are, and it can also drain a runway you needed for the product. If you are pre-revenue, a good coined .com you can register today at standard cost is usually the smarter allocation. If you are funded, established, or in a category where credibility is the whole sale, an aftermarket name earns its price. When you do go that route, an appraisal estimate gives you a reference point before you negotiate, and names that are available, listed for sale, or live in auction all surface from the same search when you buy domains in one place.

A practical process for shortlisting and deciding

Naming stalls because people try to find the perfect name in one sitting. Split it into stages and it becomes tractable.

• Generate wide, judge later. Get to 40 or 50 candidates before you eliminate anything. Mix coined words, real words used sideways, and two-word compounds.
• Kill on hard criteria. Cut anything unspellable, unpronounceable, hyphenated, or over about 15 characters. This usually removes half.
• Check availability on the survivors. Search each one and see whether it is unregistered, listed for sale, or heading to auction. Some names you assumed were dead are quietly purchasable.
• Run the radio test on the top eight. Say them out loud to real people. Watch for hesitation, not opinions.
• Trademark screen the top three. USPTO search plus a category search. Anything ambiguous goes to a lawyer or gets dropped.
• Check handles, then commit. Buy the domain, secure the handles the same day, and stop deliberating. The name matters less than what you build on it.

That last point deserves emphasis. Founders spend weeks agonizing over a name and then months with nothing on the domain at all. The name gets you a fair hearing. After that you are judged on the site, the product, and whether anyone can find you, and getting found means actually publishing something worth ranking on a consistent basis rather than launching a homepage and going quiet.

Pick the shortest name that survives being said out loud, that no one else has a legal claim to, and that you will not be embarrassed to say in a meeting five years from now. Get the .com if it exists at a price you can defend. Then go build.