Naming a business is the first decision where you have almost no information and a disproportionate amount of pressure. You have not sold anything yet, you may not have finished deciding what the business does, and yet you are being asked to commit to a word that will appear on every invoice, every email address, every piece of packaging, and every search result for as long as the company exists. It is genuinely difficult, and the difficulty is why so many people either stall for weeks or pick something in ten minutes and regret it.

The way out is to stop treating naming as a flash of inspiration and start treating it as a filtering problem. You do not need a brilliant idea. You need a large enough pool of decent options and a set of filters strict enough to eliminate the ones that will cause problems later. Do that properly and the right answer tends to emerge from the process rather than from a moment of genius, which is fortunate, because moments of genius do not arrive on schedule.
This guide walks through the entire process: what actually makes a name work, how to generate options that are not all variations of the same idea, how to check availability across the four places that matter, how to test a name before you commit, and how to set up the basic identity assets once you have decided.
What this guide covers
- What makes a business name work
- Writing your criteria before you brainstorm
- Techniques that generate real options
- Cutting a long list down to a shortlist
- Checking domains, trademarks, and handles
- Testing the name before you commit
- Setting up the identity once you decide
- Frequently asked questions
What Makes a Business Name Work
There is no formula for a good name, but there is a reliable list of properties that good names share and bad names lack. Most naming disasters are a failure on one of these dimensions rather than a lack of creativity.

It survives being said out loud
Most business is still passed along verbally. Someone recommends you at a dinner, mentions you on a call, or answers the phone with your name on it. If a name has to be spelled out every time it is spoken, you have added friction to every single one of those moments, permanently.
The test is simple and unforgiving. Say the name to someone who has never heard it, then ask them to write it down. If more than one person in five gets it wrong, the name has a spelling problem, and no amount of clever branding fixes it. Deliberately misspelled names, dropped vowels, and creative letter substitutions all fail this test spectacularly.
It is distinctive within its own field
Distinctiveness is relative to your competitors, not to the dictionary. In a market where everyone is called something plus Solutions, plus Group, or plus Partners, an ordinary word becomes distinctive simply by not following the pattern. In a market full of invented one word names, a plain descriptive name stands out.
Write down the names of ten competitors before you name anything. If your candidate would blend into that list, it is too similar, regardless of how nice it sounds in isolation. Blending in is worse than sounding slightly odd, because a name that blends in does no work at all.
It still fits in five years
The most common regret is a name that describes exactly what the business did at the start. A company named after one city that expands nationally, a name containing one product category that later becomes a small part of the revenue, or a name built around a technology that gets replaced, all end up with a name that quietly contradicts the business.
Ask what happens to the name if you add a second product line, move to a different country, or serve a completely different customer. If the name would have to change, it is too narrow. This is the argument for names that suggest rather than describe.
It is legally and practically available
A name is not yours because you thought of it. It is yours when you can register the company, secure the domain, hold the handles, and defend it if someone else claims it. Availability is not a formality to check at the end. It is a filter that kills roughly two thirds of candidates, which is exactly why you need a long list rather than a favourite.
It works visually
A name has to survive being set in a logo, cropped into a square app icon, printed at small sizes, and reduced to a single letter on a favicon. Very long names become unreadable at small sizes. Names starting with a common letter are harder to make distinctive as a monogram. Names with awkward letter combinations can be difficult to set attractively in most typefaces.
You do not need a designer to check this. Type the candidate in a plain bold typeface, shrink it to the size of a browser tab, and see whether it is still legible. That is the size most people will encounter it.
The types of names, and what each costs you
| Type | Example pattern | Strength | Weakness |
|---|---|---|---|
| Descriptive | City Dental Care | Instantly understood, helps early search | Generic, hard to trademark, limits growth |
| Suggestive | Evergreen, Northwind | Evokes a feeling, room to grow | Needs some explanation at first |
| Invented | Zolvex, Numera | Fully ownable, domains available | Means nothing until you build meaning |
| Founder name | Harrow and Sons | Personal, credible in trades and services | Hard to sell, tied to one person |
| Compound | Fieldnote, Rooftop | Familiar words, distinctive combination | Popular approach, many are taken |
| Metaphor | Anchor, Beacon, Summit | Memorable, visual, easy to say | Heavily used, trademark conflicts likely |
None of these is correct in general. Descriptive names are genuinely better for a local service business that lives on search. Invented names are genuinely better for a product company that intends to expand. Choose the type that matches your situation before you start generating, because trying to generate all six types at once produces a scattered list you cannot compare.
Writing Your Criteria Before You Brainstorm
Almost everyone skips this step, and it is the reason most naming sessions produce a hundred options and no decision. Without criteria, every name is arguable and nothing can be eliminated, so the list grows and the decision never gets made.
Spend twenty minutes writing answers to these questions before you generate a single name.
- Who has to understand this immediately? A name aimed at procurement managers has a different job than one aimed at teenagers.
- Where will it be seen most? A search result, a van, a shelf, an app icon, and an email signature all reward different things.
- What must it never rule out? Future products, future countries, future customer types.
- What feeling should it carry? Pick three adjectives. Serious, warm, and precise is a different brief than playful, bold, and cheap.
- What are the hard constraints? Maximum length, must be one word, must be pronounceable in a second language, must not contain a specific competitor's word.
- What is the domain policy? Decide in advance whether an exact match domain is required or a reasonable variant is acceptable. Deciding this later, under pressure, is how people talk themselves into bad compromises.
Write these down where you can see them during the brainstorm. Their entire purpose is to let you reject options quickly and without argument later.
Techniques That Generate Real Options
Unstructured brainstorming produces roughly fifteen names, twelve of which are variations on the same idea. Structured generation produces sixty or more that are genuinely different, which is what a filtering process needs.

Build the raw material first
Before naming anything, write four word lists of about twenty entries each. What the business does, in verbs. What the customer gets, in outcomes. What the field is made of, in physical objects and materials. What the feeling is, in adjectives. These lists are the raw material for everything that follows, and building them separately stops you from circling the same three words.
Combine across lists
Take a word from one list and pair it with a word from another. Most combinations are terrible and that is expected. The point is volume, and the good ones only appear once you have generated the bad ones. Compound names like Fieldnote or Brightwater come directly out of this exercise.
Go sideways into other fields
Some of the strongest names come from a completely unrelated domain that shares a structure with yours. Sailing, cartography, geology, printing, botany, astronomy, and old trades are all rich sources of concrete words that sound distinctive in a modern business context precisely because they came from somewhere else. A logistics company borrowing from navigation, or a research tool borrowing from surveying, gets a name that feels considered rather than generic.
Use a generator to break your own patterns
Human brainstorming has a strong bias toward the words you were already thinking about. A generator has no such bias, which is exactly its value. You are not looking for it to hand you the final answer. You are looking for the one candidate in fifty that contains a sound or a fragment you would never have reached on your own, which you then develop yourself.
The Business Name Generator is useful at this stage for filling out a long list quickly. Run it several times with different input words from your four lists, keep anything that makes you pause even slightly, and do not evaluate as you go. Judging during generation is what keeps lists short.
Fill out a long list fast, then do the filtering yourself. Volume first, judgement second.
Try the Business Name GeneratorModify what you have
Once you have a list, run modifications across it. Change a word to its older or more specific form. Try the name in another language, checking carefully what it means there. Add or remove a prefix or suffix. Drop one syllable. Try the plural. These small transformations frequently turn a decent candidate into a strong one, and they cost almost nothing.
Set a target and stop judging
Aim for sixty names before any evaluation. Sixty sounds absurd until you remember that availability checks will eliminate most of them. If you start with eight, you will finish with one or two, and you will end up choosing a name because it was available rather than because it was good.
Cutting a Long List Down to a Shortlist
Now the criteria earn their keep. Work through the long list in passes, each one cheap and fast, so the expensive checks only run on names that deserve them.
Pass one: say every name out loud
Read each candidate aloud as though answering the phone with it. This one action eliminates a surprising number of names that looked fine written down, and it takes about five minutes for sixty names. Anything that makes you hesitate, that runs two similar sounds together awkwardly, or that sounds like something else entirely when spoken quickly, comes off the list.
Pass two: the spelling test
Send your remaining names to three or four people as an audio message, and ask them to reply with the spelling. Do not send the written version. The names that come back consistently misspelled are telling you exactly what will happen every time a customer tries to find you after hearing about you.
Pass three: check for unintended meaning
Search each remaining name on its own and then alongside a word from your industry. You are looking for existing companies, obviously, but also for slang meanings, unfortunate abbreviations, and any strong association you would rather not inherit. Check what the initials spell. Check what happens when the name is written as a single lowercase string in a domain, because word boundaries disappear there and occasionally produce something embarrassing.
Pass four: the tab test
Set each survivor in a plain bold typeface, shrink it to browser tab size, and look at it. Then look at just the first letter, which is what a favicon usually becomes. Names that turn into a grey smudge or share an unremarkable initial with every competitor are worth demoting now rather than after you have paid for a logo.
You should be down to about eight names. That is the right number to take into availability checks, because you will lose most of them there.
Checking Domains, Trademarks, and Handles
This is where naming projects actually get decided, and where doing the checks in the wrong order wastes the most time. Check the cheapest and most fatal things first.

Trademarks come first, not last
A trademark conflict is the only failure on this list that can force you to rename after launch, so check it before you get attached. Search your national trademark database, and the regional one if you trade across borders. You are looking for marks that are the same or confusingly similar in the same class of goods or services.
Two points that surprise people. First, similar counts, not just identical, so a name that merely sounds like an existing mark in your field can still be a problem. Second, class matters, so an identical name in a completely unrelated industry is often fine. If anything looks close and the business matters to you, this is the one stage worth paying a professional to review.
The company register
Every country has a register of company names, and most will simply refuse a registration that is identical or nearly identical to an existing one. This check is usually free and takes seconds. Note that registering a company name is not the same as owning a trademark, and neither one grants the other.
Domains
Check the exact match first, then sensible variants. Have your policy from the criteria stage ready, because this is where discipline collapses.
A few practical rules. Avoid hyphens, because they get lost when the name is spoken. Avoid numbers for the same reason, since nobody knows whether to type the digit or the word. Watch for accidental word breaks in the concatenated form. And if the exact domain is held by an active business in a related field, treat that as a red flag about the name rather than as a domain problem, because your customers will confuse you with them.
When you settle on the domain form, that same string is usually the cleanest basis for your page URLs too. Running your name and your main service phrases through the Slug Generator shows you exactly what they look like as lowercase hyphenated URLs, which is worth seeing before you commit, because an awkward slug is a small permanent tax on every page you ever publish.
Social handles
Check the handle on the two or three platforms you will genuinely use, not all of them. Chasing a perfect match across eight networks eliminates good names for no benefit. A consistent handle is nice. An identical handle everywhere is a luxury, and adding a short prefix or suffix on one platform costs you nothing real.
Record the results properly
Build a simple table with your eight names down the side and four columns across: trademark, register, domain, handles. Fill it in completely before making any decision. The pattern is almost always that two names survive cleanly, three have one solvable problem, and three are dead. Seeing that laid out makes the decision far easier than holding it all in your head.
Testing the Name Before You Commit
You now have one or two survivors. Before registering anything, put them through a short reality check. This takes a day and prevents the specific kind of regret that only appears after the signage is printed.

The phone test
Answer a real phone call with the name for a day. Say the full greeting you would actually use. Names that are a pleasure to write can be genuinely awkward to say forty times, and you only discover that by saying them forty times.
The email test
Write out the email address, then read it aloud as if giving it to someone over the phone. Long names produce long addresses. Names with ambiguous spelling produce addresses you have to spell out every time. This is a small friction that repeats for the entire life of the business.
The recall test
Mention the name once to five people, in passing, without emphasis. Ask them two days later what it was. This is a harder test than asking them to choose a favourite, and it is far more informative, because it measures the only thing that matters: whether the name sticks without effort.
The context test
Write the name into the sentences it will actually appear in. A search result title. An invoice header. A sentence introducing yourself at an event. A one line description in a directory. Some names that seem strong in isolation deflate the moment they sit inside a normal sentence, and you would rather find that out now.
Ignore the wrong feedback
When you ask people about a name, most will tell you whether they like it, which is close to useless. Preference is not the question. Ask them what the business does, ask them to spell it, and ask them what it reminds them of. Those three answers tell you whether the name works. Whether a friend likes it tells you about your friend.
Also be careful about consensus. A name everyone finds mildly agreeable is usually a name nobody will remember. Some resistance is normal for distinctive names, and it fades quickly once the name is attached to something real.
Setting Up the Identity Once You Decide
When you commit, move fast and register everything on the same day. A name that is available on Monday can be gone by Friday, and partial ownership, where you have the domain but somebody else grabbed the handle, is a genuinely annoying position to be in.
The same day checklist
- Register the domain, plus one or two obvious defensive variants and common misspellings.
- Register the company with your national authority.
- Claim the social handles you will use, even if the accounts sit empty for months.
- File the trademark application if the business justifies it.
- Set up the primary email address on the domain.
- Write the name down exactly as it should always be written, including capitalization, and use that form everywhere from day one.
The first visual assets
You do not need a full brand identity to launch, and paying for one before you have customers is usually premature. What you do need is a consistent wordmark, a colour, and an icon small enough to work everywhere.
The browser tab icon is the piece most people forget, and its absence makes a new site look unfinished in a way visitors notice without being able to say why. A single letter or a simple shape in your brand colour is enough, and the Favicon Generator will produce the required sizes from one image so the tab, the bookmark, and the mobile home screen icon all look intentional.
Connecting the name to the physical world
If the business has any physical presence, packaging, a vehicle, a shopfront, a market stall, a printed card, you need a way for people to get from seeing the name to reaching you without typing. A QR code on packaging or a card removes the step where someone tries to remember your name later and fails.
Point it at a real destination rather than your homepage where possible, such as a booking page or a product page, and print it large enough to scan from a normal distance. The QR Code Generator creates one you can download and drop into a print file, and it is worth doing before your first print run rather than after.
Write down how the name is used
One page is enough. How the name is capitalized, whether it takes a legal suffix in formal contexts, the exact tagline, the colour values, and the file names of the logo versions. Without this, three months of inconsistent usage accumulates across documents and profiles, and cleaning it up later is tedious. Writing it down on the first day costs fifteen minutes.
A One Week Naming Schedule
Naming expands to fill whatever time you give it, so give it a week and treat the deadline as real. Spreading the work across days matters more than the total hours, because the gaps let bad candidates lose their shine and good ones prove they are memorable.
| Day | Work | Output |
|---|---|---|
| Monday | Write criteria, list ten competitors, build the four word lists | A one page brief |
| Tuesday | Generate without judging, using every technique | Sixty or more candidates |
| Wednesday | Leave it alone entirely | Distance |
| Thursday | Say aloud, spelling test, meaning check, tab test | Roughly eight survivors |
| Friday | Trademark, register, domain, and handle checks | A completed availability table |
| Weekend | Phone, email, recall, and context tests | One clear winner |
| Monday | Register everything, write the usage page | A name you own |
The Wednesday gap is not padding. Names that still feel right after two days away are the ones worth checking, and the ones that suddenly look obvious and dull were never going to survive anyway. Deciding on the same day you generate is how people talk themselves into their fourth best option.
If you genuinely cannot spare a week, compress it to three days but keep the shape: generate, gap, filter, check, decide. What you must not do is collapse it into one sitting, because a single session produces a short list of similar names and no time for any of them to prove they are memorable.
When to stop and just pick
There is a point where more searching stops helping. If you have two clean, available, sayable finalists and you have been going back and forth for more than a day, the difference between them is smaller than the cost of the delay. Pick the one that was easier to say on the phone, register it, and move on to the part of the business that actually determines whether any of this matters.
Naming Mistakes That Cost Real Money
Almost every expensive naming problem is one of a short list of mistakes, and every one of them is avoidable at the point of decision rather than after launch.
Falling in love before checking availability
The single most common failure. Someone settles on a name emotionally, then discovers a trademark conflict, and instead of walking away, starts negotiating with reality. They add a word. They swap a letter. They take a worse domain. Each compromise weakens the name, and the final result is measurably worse than the second choice they never seriously considered. Check availability while you are still indifferent between candidates.
Naming for today's business
A name that pins you to one city, one product, or one technology is a name with an expiry date. Businesses that outgrow their names spend years explaining that they do more than the name suggests, or pay for a rename that resets brand recognition to zero. If your five year plan contradicts your name, change the name now, while it is free.
Choosing a name that only works written down
Names built on visual cleverness, unusual capitalization, a symbol in the middle, a deliberate misspelling, collapse the moment they enter speech. They cannot be dictated over a phone, they get typed wrong into search boxes, and they are impossible to say on radio or in a podcast. If the joke only lands visually, it is not a name, it is a logo idea.
Picking by committee
Group naming decisions converge on the option nobody objects to, which is reliably the blandest one on the list. Distinctive names always attract some resistance, because unfamiliarity feels like wrongness. Gather input widely, but let one person decide. Ask the group diagnostic questions rather than preference questions, and weigh what they remember over what they like.
Confusing a company registration with ownership
Registering a company name gives you the right to trade under it. It does not stop a trademark holder in your field from requiring you to stop. People discover this in the worst possible way, through a letter, after investing in signage and packaging. The register and the trademark database are two separate checks and you need both.
Optimizing entirely for the domain
Choosing a name because a good domain was free produces names that no human would have arrived at independently. The domain matters, but it is one criterion among six. A slightly imperfect domain attached to a strong, memorable, sayable name beats a perfect domain attached to a name nobody can recall.
Never writing anything down
Six weeks after the decision, nobody remembers why the second choice was rejected, and the debate reopens every time a new person joins. Keep the shortlist, the availability table, and one paragraph explaining the decision. It ends the argument permanently and it takes ten minutes.
Naming Across Languages and Borders
If there is any chance the business will operate outside one language, a few extra checks now prevent a difficult conversation later.
Check what it means elsewhere
Run your finalists through a translation check in the major languages of any market you might plausibly enter, and specifically check slang rather than only formal meaning. The famous cautionary tales in this area are famous precisely because the check takes ten minutes and nobody did it.
Check whether it can be pronounced
Sounds that are natural in one language can be genuinely difficult in another. Consonant clusters that English speakers produce without thinking are hard in Japanese and Spanish. Some sounds do not exist at all in Mandarin. If a name cannot be said comfortably by your future customers, they will shorten it into something you did not choose, and that shortened form becomes your real name whether you like it or not.
Check the character set
Accented characters, ligatures, and non Latin scripts create real friction in domains, email addresses, and search. If your name contains one, decide early what the plain ASCII form is, register that too, and use it consistently in technical contexts. Leaving it undecided means five different spellings end up in circulation.
Check that it does not collide regionally
A name that is clear in one country may already belong to a well known company in another. This rarely creates a legal problem across unrelated markets, but it does create a search problem, because you will spend years ranking below them for your own name. Search the name alongside your industry term in each target market before committing.
The short version
- Write your criteria before generating names, or you will never be able to eliminate anything.
- Generate sixty candidates, not eight. Availability checks will destroy most of them.
- Say every name out loud and have strangers spell it. This kills bad names fastest.
- Check trademarks first, then the company register, then domains, then handles.
- Distinctiveness is measured against your competitors, not against the dictionary.
- Test for recall and comprehension, not for whether people like it.
- Register everything on the same day you decide, then write down how the name is used.
Frequently Asked Questions
How long should a business name be?
One to three words, and ideally under about fifteen characters. Short names are easier to say, easier to type, easier to fit on a sign or an app icon, and far less likely to be mistyped. Length matters less than syllable count: a two syllable name is repeated correctly more often than a four syllable one, regardless of how it is spelled.
Should my business name describe what I do?
Descriptive names help early, when nobody knows you, because the name does part of the explaining. They hurt later, when you expand beyond what the name describes. A useful compromise is a distinctive name paired with a descriptive tagline, so the name stays flexible while the tagline carries the explanation and can be rewritten at any time.
Do I need the exact .com domain?
It is worth real money but it is not a requirement. Plenty of successful companies run on a country domain, a .co, or a two word variant. What matters far more is that people can find you when they type your name into a search engine. If the .com is taken by a parked page, consider a variant. If it is taken by an active company in a similar field, choose a different name entirely.
How do I know if a business name is already taken?
Check four places in order: your national company register, a trademark database for your country, domain availability, and the handle on the two or three social platforms you actually intend to use. A name that clears all four is genuinely available. Clearing only the domain is the most common and most expensive mistake, because a domain says nothing about trademark rights.
Can I change my business name later?
You can, and businesses do, but it costs more than people expect. Registrations, domains, email addresses, printed material, signage, packaging, app store listings, invoices, and every backlink pointing at your old site all have to be dealt with, and search rankings usually dip for several months. Spending an extra week on the decision up front is far cheaper than a rename in year three.
Is it a bad sign if I cannot find any name I love?
No, and expecting to fall in love is what stalls most naming projects. Almost every famous company name sounded ordinary or slightly odd on the day it was chosen. Names acquire meaning through use, not before it. Pick the strongest option that is clear, available, and easy to say, then go and build something that makes the name mean something.
Putting It Together
The reason naming feels impossible is that people try to do it in one step, judging each idea the moment it appears and hoping one of them arrives fully formed and obviously correct. That almost never happens. What does work is separating the stages: decide what the name must do, generate far more options than feels reasonable, filter them with cheap tests before expensive ones, check availability in the order that eliminates fastest, and test the survivors on real people in realistic situations.
Done that way, naming takes about a week of intermittent work rather than a month of circling. And the result is usually a name that is clear, sayable, available, and durable, which is all a name has to be. The meaning comes later, from what you do with it. Every well known company name was an ordinary word until the company made it mean something, and yours will be too.
← Back to all articles
