
A step-by-step framework for naming your business: brainstorming methods, domain and trademark checks, and how to choose a name that's memorable and ownable.
Why Most Business Names Fail Before the Business Does
A business name has one job that almost nobody writes down: it has to survive being repeated. A customer hears it once, at a barbecue or in a podcast ad or from a contractor’s mouth through a closed truck window, and then has to carry it home, type it into a search bar, and land on you instead of someone else. Every naming failure is a failure somewhere along that chain — the name couldn’t be remembered, couldn’t be spelled, couldn’t be found, or turned out to belong to somebody else.
Founders tend to optimize for the wrong moment. They judge candidate names by how they feel in a brainstorm — clever, meaningful, personally resonant — when the moment that matters is six weeks later, when a half-interested stranger tries to reconstruct the name from memory. “Clever in the meeting” and “recoverable from memory” are different properties, and the second one pays the bills.
The other quiet killer is sequencing. Most people fall in love with a name first and check whether they can own it second. By the time they discover the .com is parked, a trademark is registered in their category, or the provincial registry rejects the name as too similar to an existing corporation, they’re emotionally committed and start rationalizing workarounds — a hyphenated domain, a slightly different spelling, a “we’ll deal with it later.” Those workarounds become permanent taxes on the brand.
This guide runs the process in the right order: understand the naming conventions available to you, generate broadly, kill ruthlessly, and only then run the ownership checks — domain, trademark, and registration — before anything gets printed on a truck.
The Four Naming Conventions and When Each One Works
Almost every business name falls into one of four buckets, and each trades memorability, meaning, and ownability differently.
Descriptive names say what the business does: Toronto Drain Cleaning, North Shore Bookkeeping. They’re instantly clear, which is genuinely valuable for local service businesses where the customer is searching for the service, not the brand. The costs are real, though: descriptive names are nearly impossible to trademark because no one can own plain language for the thing itself, they blur into every competitor using the same words, and they box you in — Toronto Drain Cleaning can’t gracefully expand into Mississauga or into full plumbing.
Founder and family names — Hamilton & Sons, Drier Group — carry built-in accountability and age well in trades, law, and anything reputation-driven. They’re reasonably ownable and feel trustworthy, but they say nothing about what you do, so they usually need a descriptive tagline doing the explaining for the first several years.
Evocative names borrow a real word or image and attach it to the business by association: Patagonia, Slack, Forge, Harbour. This is the sweet spot for most modern brands — real words are easy to remember and spell, while the metaphor gives the brand somewhere to go. The risk is collision: real words have existing search results, existing trademarks in adjacent categories, and existing meanings you don’t control.
Invented names — coined words like Kodak or blends like Netflix — are the most ownable of all. The domain is usually available, the trademark path is cleanest, and search results start empty. The price is that an invented name arrives meaning nothing; every ounce of association has to be bought with marketing. That trade is fine for a funded startup and expensive for a two-person firm that needs the name to do some selling on day one.
There’s no universally correct bucket. A solo electrician and a software company should not name themselves the same way. Pick the convention that matches how your customers will actually encounter and search for you, then generate within it.
The Creative Process: Generating Names That Stick
Good names rarely arrive as lightning bolts. They come out of volume and constraint — a wide, slightly embarrassing longlist narrowed by cold criteria. The mistake is brainstorming five names, debating them for a week, and picking the least objectionable one. Aim for fifty to a hundred candidates before you judge any of them.
Start with raw material, not names. List the words of your category: the tools, the outcomes, the feelings customers have before and after, the geography, the materials, the verbs. A roofing company’s sheet might include ridge, crest, summit, shield, dry, overhead, slate, apex, shelter. Then work the list: translate words into other languages you have a real connection to, chase synonyms two steps out, pair unexpected words together, clip syllables, blend fragments. Most of what this produces is garbage. That’s the point — the garbage is the path to the three good ones.
As the longlist forms, apply the structural tests that predict memorability. Shorter beats longer; one to three syllables is the comfortable zone for a name people will say aloud. Stress patterns matter — names that start with a hard consonant and land on a strong first syllable tend to stick. And run the radio test early: say the candidate in a sentence out loud to someone who hasn’t seen it written, then ask them to spell it. If “Kwikflo” comes back as Quickflow, Kwickflo, and Quikflo, you’ve found a name that will leak customers to typos forever.
Finally, sleep on the shortlist. A name that still sounds right after two weeks of casual use — said on fake phone answers, typed in fake email signatures — has passed a test no brainstorm can simulate. Names wear in or wear out, and you want to know which before the signage is ordered.
The Kill Tests: How to Eliminate Names Before You Commit
Once you have a shortlist of five to ten survivors, the job flips from creative to adversarial. You are now trying to kill each name, and any name you can kill cheaply at this stage would have cost far more to kill after launch.
First, the say-it-out-loud battery. Answer a phone with it. Use it in “I work at ___.” Say the web address: “that’s ___ dot com.” Names that need a spelling assist every time — silent letters, dropped vowels, soundalike words — fail here. So do names that are awkward to say twice in a row, which is exactly what happens in referrals.
Second, the meaning sweep. Search the name plus a few unflattering words. Check what it means or sounds like in the other languages of your market — in Canada, French is not optional thinking if you ever plan to sell nationally, and in any major city your customer base speaks dozens of languages. Check slang dictionaries. Plenty of companies have discovered post-launch that their name is a punchline somewhere that matters.
Third, the collision scan. Search the exact name in quotes. Search it with your city and your category. Look at who already lives in those results: a defunct company with the same name and bad reviews, a business two provinces over in a similar trade, a product in an unrelated category that dominates page one. None of these is automatically fatal, but each one is a cost you should price in deliberately rather than discover later.
Fourth, the growth test. Will the name still fit when you add a second service, a second city, or a second decade? Names with a city, a single service, or a trend baked in (anything that sounds unmistakably of this year) carry an expiry date. You don’t have to optimize for an empire you may never build, but you should know which names have ceilings.
What survives all four tests goes on to the ownership checks — and those checks, not taste, make the final cut.
Domain Availability: Checking and Securing the Name Online
The domain check is the cheapest, fastest ownership test, so run it first. Type each surviving candidate into a registrar’s search and see what’s actually available — not just whether the .com is taken, but who has it and what they’re doing with it.
The hierarchy is straightforward. The exact-match .com is still the default people type and the safest long-term asset. For a Canadian business, the matching .ca is close behind — it signals local, performs perfectly well in Canadian search, and is often available when the .com isn’t. If the .com is parked by a squatter asking four figures, owning the .ca plus the exact-match name everywhere else is usually a better deal than a renamed business. If the .com is an active company — especially in a related industry — treat that as a kill signal, not an inconvenience: you’d be sending them your customers and their reputation would be bleeding onto you.
When the exact match is gone, modifiers can work, but choose them carefully. A natural extension of the name — adding the category, like northshorebooks becoming northshorebookkeeping — survives word of mouth. Hyphens, doubled letters, and creative misspellings mostly don’t; every one is a fork in the road where some fraction of typed traffic goes to someone else.
While you’re checking domains, check the handles. Search the name on the social platforms that matter for your category, and check whether a sensible email pattern is possible. A name is a bundle of addresses, and you want the whole bundle.
Two practical habits: keep a shortlist private until you’re ready to buy, and when you decide, register immediately — domains cost less than lunch, so registering two or three finalists while you complete the trademark and registry checks is cheap insurance. Buy the .com and .ca together if you can, set them to auto-renew, and register them in an account the business owner controls, not a web designer’s personal account. Recovering a domain from a departed contractor is a genuinely miserable process.
Trademark Basics: Search Before You Fall in Love
A quick framing first: none of this is legal advice — trademark clearance is a job for a registered trademark agent or lawyer, and the few hundred dollars a professional opinion costs is small next to a forced rebrand. But you can and should do preliminary screening yourself, because it kills doomed names for free before you spend money on the survivors.
The core concept is that trademarks protect names within categories of goods and services. Two companies can share a name in unrelated industries; they generally can’t in the same one. So your search isn’t “does this name exist anywhere” — it’s “does this name, or something confusingly similar, exist in or near my category, in the markets where I’ll operate.”
Start with the Canadian Trademarks Database on the Canadian Intellectual Property Office’s site. Search the exact name, then obvious variants — phonetic spellings, singulars and plurals, the name with and without its descriptive word. Look at what each hit is registered for, not just whether it exists. If you’ll sell into the United States, repeat the exercise on the USPTO’s trademark search. Then remember that registration isn’t the whole picture: in Canada, businesses can acquire common-law rights just by using a name in the market, so a plain web search for the name in your industry matters too.
Two more concepts worth knowing before you talk to a professional. First, distinctiveness: invented and evocative names are strong candidates for registration, while descriptive names (“Toronto Drain Cleaning”) are weak or unregistrable, because the law won’t let one company own the plain description of a trade. Second, registering your corporate name or a domain gives you essentially no trademark protection by itself — they are different systems answering different questions. Clearing the name is about avoiding a fight; registering the trademark, if your budget allows, is about being able to win one.
Canadian Incorporation and Business Name Registration Rules
Separate from trademarks, there’s the question of what governments will let you register, and Canada has its own machinery here.
If you incorporate federally under the Canada Business Corporations Act, your proposed name is actually examined. It needs a distinctive element (the part that identifies you), usually a descriptive element (the part that says what you do), and a legal element — Inc., Ltd., Corp., Limited, Incorporated, or a French equivalent. Federal applications are screened against a NUANS report, a search that compares your proposed name against existing corporate names and trademarks across databases, and names that are confusingly similar to existing ones, misleading about what the business is, or that suggest government connection can be refused. Most provinces have parallel requirements for provincial incorporation, with details that vary — some require a name search report, some run their own screening — so check the rules for your jurisdiction rather than assuming they match the federal ones.
There’s also the escape hatch many founders use: incorporate as a numbered company (the registry assigns something like 12345678 Canada Inc.) and register your brand as a trade name — an “operating as” name — separately. That decouples the legal shell from the brand, which is convenient, but note carefully what it doesn’t do: registering a trade name in a province is mostly a disclosure requirement, not a grant of exclusive rights. In most cases it won’t stop someone else from using a similar name, and it won’t protect you if you step on someone’s trademark.
If you’ll operate under any name other than your own legal name — which includes sole proprietors using a brand — most provinces require you to register that business name, and you’ll want the name settled before you open bank accounts, sign leases, and file for tax accounts, because changing it later means changing all of them. The practical sequence: clear the name against trademarks and registries first, then incorporate or register, then lock down the domains the same week.
SEO and AI Findability: Can Anyone Actually Search for You?
A name now has to perform in two retrieval systems: search engines and AI assistants. Neither changes what makes a good name, but both raise the cost of certain bad ones.
The first findability test is simple: when someone searches your exact name, can you plausibly own the results page? Names that are common dictionary words on their own — a company called simply “Bloom” or “Apex” — are fighting every other use of that word on earth, and a small business rarely wins that fight. The standard fix is a two-word name or a word pair that becomes unique in combination: “Apex” is unfindable, “Apex Roofing North Vancouver” as a consistent brand string is not. Invented names start with an empty results page, which is the easiest page to own.
The second test is collision with bigger entities. If your candidate name is shared by a celebrity, a popular product, a TV show, or a large company anywhere, search engines and AI assistants will resolve the ambiguity in their favour, because the systems answer with the most likely referent. You don’t need a unique name in the universe — you need to be the obvious answer for your name within your category and geography.
For AI findability specifically, consistency is the lever. Assistants and answer engines learn what your business is from the consistent co-occurrence of your name with your category, your city, and your details across your website, your Google Business Profile, directories, and reviews. A name helps when it’s distinctive enough to function as a stable entity and spelled one way everywhere; it hurts when it has variant spellings, an ambiguous word boundary (is it SunRise or Sun Rise?), or punctuation that gets dropped in databases. Skip special characters, skip deliberate misspellings, and decide the exact canonical string — capitalization, spacing, the works — before you create the first account.
None of this means choosing a keyword-stuffed name; a name doesn’t need your service in it to rank for your service. It means choosing a name that machines can reliably tell apart from everything else called something similar.
Making the Final Call: A Decision Checklist
By this point the process has done most of the deciding for you. You generated wide, killed names that fail the spoken tests, and eliminated candidates with domain, trademark, or registry conflicts. What’s left is usually two or three names, all viable, and the final pick is the only genuinely subjective step.
Run each finalist through one last pass: it can be said and spelled after being heard once; the .com or .ca (ideally both) is in hand; the preliminary trademark screen came back clean in your category and a professional has sanity-checked the one you’ll commit to; the name clears or has a plan for incorporation and business name registration in your jurisdiction; the exact-name search results are winnable; the social handles exist in some consistent form; it won’t embarrass you in the other languages of your market; and it leaves room for the business you might be in five years.
Then pick the one you can say with a straight face, and stop. Naming has a long tail of second-guessing, and the marginal difference between two clean, ownable, pronounceable names is far smaller than founders believe. Brands earn most of their meaning after the choice — a name that seems flat today becomes inevitable after two years of good work attached to it, and the cleverest name in the world can’t rescue a business that under-delivers.
The last step is unglamorous and essential: write the canonical name down — exact spelling, spacing, capitalization, legal name versus operating name — and use that string everywhere, from the domain registration to the bank account to the Google Business Profile. The checks you ran in this process were all about making sure the name could be owned. Consistency from day one is how you actually own it.
Want help implementing this?
Get a free proposal for your branding setup. We’ll show you exactly where the opportunities are.
Get Free ProposalRelated Articles