Naming a startup when the perfect .com is taken
7 min read
Two problems arrive together at the start of a company. The name you want has no .com, and you do not yet know exactly what you are building — which means you are choosing a permanent name for something still moving.
Both have the same answer, and it is not a compromise. Separate the company from the address, work out which properties of the name you actually need, and generate a family rather than a Plan B.
This guide is about naming under both constraints, because in practice you get both at once.
Separate the startup from the domain
Your startup is the company, the product, the reputation and the experience. The domain is an address for that, and an important one. It is not the thing itself, and it should not be allowed to choose the company name.
So the first question when the exact .com is gone is not 'what else is free'. It is whether the brand is strong enough to be worth keeping, because that determines whether you are changing the name or changing the address — and those are entirely different projects.
The trade in general terms: modifying the brand buys you an exact-match address and one identity everywhere, at the cost of the original word. Modifying the address keeps the word, at the cost of some direct traffic and some spoken clarity, permanently. The full comparison is here, including how to work out which situation you are actually in.
Find the part of the name worth keeping
Before replacing anything, write down what you liked. The first syllable. The meaning. A metaphor. Simply the way it sounds in your mouth.
That is your anchor, and once it is on paper you can change every other letter without losing the direction. Founders who skip this step tend to reject twenty perfectly good variations without being able to say why — the reason is usually that the variations dropped the property that mattered, and nobody had named it.
Create a family, not a Plan B
The instinct is to produce one awkward alternative and then argue about whether it is good enough. It never is, because it has nothing to be compared against.
Generate thirty to a hundred variations around the same root instead: change endings, rotate vowels, swap opening sounds, add and remove a syllable. A family gives you a comparison, and a comparison is what stops you accepting the first available name out of relief. Turning one word into hundreds of ideas is the mechanical version.
Expect the result to surprise you. The most common outcome of generating properly is that the third or fourth available variation is a better company name than the word you originally wanted — not a consolation, an upgrade.
When you do not know what the company will be
Early products change features, customers and business models. A naming strategy should assume that rather than pretend version one is permanent — and the failure mode is specific and common: naming the company after the thing it currently does.
Name the enduring idea, not the feature
Look underneath the current product for the thing that will still be true after two pivots: coordination, discovery, trust, speed, access, creativity. That is what the name should be attached to.
Precise feature language belongs in product copy, where it can be rewritten. A company named after one button or one workflow outgrows its own identity in about eighteen months, and then the name is actively misleading.
A broad metaphor beats a broad abstraction
Metaphors give a brand personality without locking it to a category, which is exactly what you want under uncertainty. The condition is that the metaphor connects to the experience rather than being an attractive object you liked — an arbitrary noun is not flexibility, it is absence of meaning.
This is the line to hold: a name can be broad without being arbitrary. Keep a stated reason for the seed, the sound or the association you chose, and 'it leaves room' is not on its own a reason.
Invented words create the most room
A pronounceable new word carries less fixed meaning than a dictionary word, so it stretches as the company changes and lets the product define the term. It still has to pass the sound, spelling and conflict tests — invented is not a shortcut past any of those.
Test each candidate against three futures
This is the test that separates a name with room from a name you are hoping about, and it takes ten minutes.
- Write one paragraph describing the product as it exists today.
- Write one describing the most likely expansion — the next product, the adjacent customer, the second market.
- Write one describing the ambitious version, the company you would be delighted to become.
- Read every finalist against all three and note where it stops fitting.
A name that works for all three is flexible. A name that only works for the first is a rebrand with a delay on it. And a name that works equally well for all three because it says nothing at all is the other failure — check that it still means something in paragraph one.
Test them as companies, not as strings
A candidate that looks clever in a spreadsheet can be weak in a mouth. Put each one into the sentences it will actually live in: 'I work at ___.' 'We built ___.' 'Visit ___ dot com.' 'You should talk to the people at ___.'
Say them to somebody rather than reading them to yourself. This test outranks anything a generator told you, because the sentences are where the name spends its life.
Compare exact match against modifier, deliberately
Two routes, and the right one depends on a single question: how strongly do you expect people to type your brand directly?
A startup whose customers will arrive from links, an app store or a marketplace loses little to 'get' or 'try' in front of the address. A startup expecting word of mouth, podcasts, conference conversations or anything spoken loses something on every mention — and keeps losing it after the company is well known, because that is when people start typing the bare name.
Do not distort the name for a hypothetical future either way. A deliberate misspelling adopted to force an exact match costs you a spelling correction on every call, forever, which is worse than the modifier it replaced.
Research conflicts before the identity, not after
Search the web, the app stores and the trademark registers for your markets before anybody designs anything. An available domain is not legal clearance, and a one-letter difference from an existing mark does not remove confusion.
One point specific to naming under uncertainty: search the categories you might enter, not only the one you are in. A deliberately broad name has more room to collide, and the collision will surface exactly when you expand — how to check whether a name is already in use covers the searches.
Choose from a shortlist, on criteria
Keep three to five finalists and score them all on the same axes. Startup naming becomes much less emotional when every candidate is judged by the same standard rather than by whoever last argued for one.
Score every finalist
- A stranger can say it from the spelling
- A stranger can spell it from the sound
- Remembered after a delay, against distractors
- Distinctive in the category you will appear in
- Works in all three futures you wrote down
- Clean address, or a modifier you would put on a billboard
- Nothing disqualifying in the conflict research
Then choose. No name anticipates every pivot, and no amount of further generation will produce certainty — aim for reasonable room, pass the practical tests, and let the company build meaning around the word.
The taken .com is useful pressure
Worth ending on, because it does not feel true at the time. Scarcity pushes founders away from obvious generic words and toward something distinctive, and distinctive is what you needed anyway.
The generic name you wanted would have competed for its own search results, shared its handles, and been harder to protect. The constraint that stopped you from having it is doing you a favour on all three counts.
In short
Separate the company from the address, write down the property you are attached to, and generate a family of thirty rather than one Plan B. Then test every finalist against today's product, the likely expansion and the ambitious version — a name that only fits the first is a rebrand with a delay on it.
Frequently asked questions
Should I use a modifier like 'get' or 'try' for a startup?
- It is reasonable while your customers arrive by clicking rather than typing. The cost grows as the company becomes known, because that is when people start typing the bare name — and the bare name belongs to somebody else.
How broad should a startup name be if the product might pivot?
- Broad enough to survive the expansion you consider likely, and no broader. A name that fits any business in any industry has bought room it will never use at the price of meaning nothing on day one.
Is an invented name risky for a startup?
- Less risky than the alternatives, in practice. It is more likely to be clear across the domain, the handles and the trademark class, and it lets the product define the term rather than fighting existing associations.
How long should I spend naming a startup?
- Days, not weeks. The generating and filtering is an afternoon and the research is another; what the remaining time usually buys is not a better name but a later launch.
Sources
These support factual background only. The guide itself is original writing, and nothing here should be read as legal advice.
Related guides
Domain Names
The domain you want is taken. Now what?
Taken means six different things, and each one has a different answer. How to find out which you are facing, then choose between buying it, renaming, and walking away.
9 min read
Startups & SaaS
How to find a short .com name for a SaaS product
From a plain description of what your software does to a shortlist of available .coms — including why category words fail, and the two tests only B2B names have to pass.
7 min read
Startups & SaaS
How to name an AI startup without sounding like every other one
Why AI names converge on the same dozen patterns, what putting the technology in the name costs you later, whether to use .ai, and how to name the job instead of the model.
10 min read
Domain Names
How to turn one word into hundreds of domain ideas
You do not need a hundred unrelated suggestions. You need one seed word and eight ways to work around it — then a filter strict enough to throw most of the results away.
8 min read
Domain Names
Every good .com is taken. Here is what actually works.
Roughly every dictionary word is registered. The way out is not a cleverer search — it is changing the shape of what you search for.
12 min read